Looking for RLs that have persistent loot between dungeon runs

[Eric Colossal]

Hello there!

I'm in the process of doing research for a game I'm designing and I'm looking for any RLs that let you keep some or all of your loot between deaths. I'd like to see if anyone has done what I'm toying with and how they accomplished it before I begin earnestly working.

The basic idea is that when you die in an RL you are punished as severely as possible with levels being reset to zero and all weapons and loot disappearing. What if the punishment wasn't so severe? What if you kept all your loot or could choose one item to take with you from the grave?

Some games like Diablo and Torchlight punish you with gold or xp subtractions and I'm told Shiren has a box you can put things in that persist through death but I'm looking for more traditional RogueLikes that play with these ideas.

The only thing I can find people talking about when I search the newsgroup is persistent levels [of course] and inheritance through death by passing traits and equipment through lineage.

I appreciate any help you can give!
Eric!

[Gerry Quinn]

In article <ded0bdd8-c4af-4c54...@googlegroups.com>,
mrcol...@gmail.com says...

Don't some games with player ghosts leave the items as loot when you
kill the ghost? I'm sure I've come across this in some game - maybe in
some version of NetHack, but I'm not sure which game.

- Gerry Quinn

[bcode]

Gerry Quinn <ger...@gmail.com> wrote:
> In article <ded0bdd8-c4af-4c54...@googlegroups.com>,
> mrcol...@gmail.com says...
>> Hello there!
>>
>> I'm in the process of doing research for a game I'm designing and
>> I'm looking for any RLs that let you keep some or all of your loot
>> between deaths. I'd like to see if anyone has done what I'm toying
>> with and how they accomplished it before I begin earnestly working.
>>
>> The basic idea is that when you die in an RL you are punished as
>> severely as possible with levels being reset to zero and all weapons
>> and loot disappearing. What if the punishment wasn't so severe?
>> What if you kept all your loot or could choose one item to take with
>> you from the grave?

[...]

> Don't some games with player ghosts leave the items as loot when you
> kill the ghost? I'm sure I've come across this in some game - maybe in
> some version of NetHack, but I'm not sure which game.

NetHack may leave a "bones file" on death which may be loaded by
another player visiting that dungeon level (or the same player
playing a new character); containing the level mostly unmodified,
including the dead character's items (which may be randomly cursed),
excluding plot-critical items and any artifacts which would otherwise
exist twice.

(The dead player's ghost usually guards the items - however, a
greater danger is usually whatever killed the former character
as ghosts are rather weak.)

The problem here may be actually getting to that dungeon level again;
especially on public servers, someone else will most likely get the
loot.

[Eric Colossal]

Yeah, Corpse Runs aren't really the same. You CAN go get the items you lost but you still have to work for them. You have to level up, descend, collect look at then find your bone pile and then loot it for items that may not be useful anymore.

[Jason Pickering]

There was in object in shiren the wanderer called the storehouse jar. Any item you placed in it was in the storehouse where you started the game. You could hold items there between runs but if you died while holding them they were lost. But there was also a thing where you could "rescue" a fallen friend. They were given the same dungeon and if they could get past the floor you died on you got all your gear back.

You can also use the idea from dark souls and ZombiU. Where you are given one run to go and retrieve your corpse giving you back all your stuff. In ZombiU when you die a special zombie is left in your place. If you kill that special zombie they drop all your original gear.

In the iOS game infinity blade every time you play you are a new generation of character and the story is you are wearing your fathers armor. I think that would make the most sense from a roguelike point of view, plus the " I will avenge my father" works as good motivation.

You could also try something where maybe the player can choose one item to be revived with. It would make the game more exciting but also force the player to get invested in one item. Maybe you always have the same sword and you can upgrade it.

[Gerry Quinn]

In article <ded0bdd8-c4af-4c54...@googlegroups.com>,
mrcol...@gmail.com says...

> Some games like Diablo and Torchlight punish you with gold or xp subtractions

In fact they really don't punish you at all unless you take the ironman
option.

Also, these games do have a 'shared stash' in which you can store loot
your current character doesn't need but a future one might. A roguelike
could probably use this concept. Note that these games require a
particular level to use an item [*] so the shared stash doesn't
trivialise the game (though for me it would take away some of the fun of
random finds).

[*] Torchlight 2 has an interesting system in which you need either a
level or given attributes. So a sword might require Level 28, but if
you had 48 strength and 30 dexterity you could use it anyway. Te
attributes come both from points put in on leveling, and from other
gear. Seems a nice concept that would work in roguelikes.

- Gerry Quinn

[Radomir Dopieralski]

On 2012-12-05, Gerry Quinn <ger...@gmail.com> wrote:
> [*] Torchlight 2 has an interesting system in which you need either a
> level or given attributes. So a sword might require Level 28, but if
> you had 48 strength and 30 dexterity you could use it anyway. Te
> attributes come both from points put in on leveling, and from other
> gear. Seems a nice concept that would work in roguelikes.

Technically, Torchlight 2 is a roguelike, even more than Torchlight was,
as there finally is perma-death. The twitch of the interface and the fact
that everything is real-time is greatly alleviated with the fact that only
the monsters in your immediate vicinity are active, so you can easily take
a break when you kill a batch.

But I didn't want to talk about definitions, but about another two
mechanics that work quite nicely in Torchlight and could work equally well
in other roguelikes.

One of them is the pet, which is relatively weak (which doesn't matter, as
it is immortal), but fast. It is, together with area-of-effect spells, an
excellent solution for all those weak monsters running from you, which are
always a pain to chase and kill. Your pet does it for you, and it will
even collect the gold scattered around (that, unfortunately, requires
extra clicks).

The second mechanic that I really like is the transmutation. You can go to
the transmuter in town, and have the spurious magical items transmuted. In
Torchlight you would get a gem, that you can then out into a socket in
your weapon to add extra effects to it. In Torchlight 2, you can replace
several items with another, random item of the same level. This way,
a unique two-handed sword is still a prize for a wizard character.

--
Radomir Dopieralski, sheep.art.pl

[Gerry Quinn]

In article <slrnkdsas...@test.moinmo.in>, ne...@sheep.art.pl
says...
[Torchlight 2]

> The second mechanic that I really like is the transmutation. You can go to
> the transmuter in town, and have the spurious magical items transmuted. In
> Torchlight you would get a gem, that you can then out into a socket in
> your weapon to add extra effects to it. In Torchlight 2, you can replace
> several items with another, random item of the same level. This way,
> a unique two-handed sword is still a prize for a wizard character.

Yes. This would be easy to add to a RL and wold not hurt baslance too
much (e.g. if 3 items make a new item, and 90% of found items get
recycled this way, the player would ultimately see 43% more items, which
is balanceable). And would give players something to do with valuable
items (no need to even make it work for small items) if they can't sell
things in a shop.

Another improvement in Torchlight 2 is that they did away with the
building up of gems into higher gems (originally cloned from Diablo 2),
removing a psychologically irresistable grinding mechanism.

- Gerry Quinn

[Radomir Dopieralski]

On 2012-12-02, Eric Colossal <mrcol...@gmail.com> wrote:
> The basic idea is that when you die in an RL you are punished as
> severely as possible with levels being reset to zero and all weapons and
> loot disappearing. What if the punishment wasn't so severe? What if you
> kept all your loot or could choose one item to take with you from the
> grave?

I think the real question you should be asking here is what purpose it
serves and how you can achieve this differently.

Death of a character is a severe punishment, and a lot of players claim
that it adds depth to the game by making it more meaningful -- after all,
you are never going to see that particular version of the world again.
This is one aspect, and it is important, but it's not the only one.

Reseting the game world to the initial state and letting the player replay
it from the beginning lets them practice and perfect the initial stages of
the game, when it's not yet as complex, and also try out different
long-time strategical decisions.

Now, if you let them retain their equipment, you are robbing them from the
opportunity to build and try a different set. You can have very varied
weapons with vastly differing and interesting mechanics, but your players
will never see that, because they will be stuck with the relatively good
sword they found early on.

Second, the simple early game, where all you got was a wooden sword and
all you could do was bashing rats to death with it, is no longer that
simple. You have your wand of fireballs, which you could use to get
through the rats faster, but you don't want to waste it so early, you have
your potions of berserk strength, you even have your plate mail armor that
you are afraid to damage in some random acid trap. The game is
considerably more complex now and the decisions grew harder to make. It
doesn't help that most of the decisions now boil down to "do I have the
patience to grind through it and save my equipment for later".

Of course, it's not at all that negative. I sometimes catch myself
thinking "I wonder what would happen if I had XXX at the beginning of the
game". It can serve as a nice source of variety if done right. For
example, you could only allow to retain a single piece of equipment, and
it can't be the same thing every time -- that would add an interesting
strategic decision.

But in the end, I really think that you should start with a problem that
you see in your game mechanics, and then think of ways of solving it, and
not start with a solution and then try to find a problem that it solves.

--
Radomir Dopieralski, sheep.art.pl

[Gerry Quinn]

In article <slrnke07m...@test.moinmo.in>, ne...@sheep.art.pl
says...

> Reseting the game world to the initial state and letting the player replay
> it from the beginning lets them practice and perfect the initial stages of
> the game, when it's not yet as complex, and also try out different
> long-time strategical decisions.
>
> Now, if you let them retain their equipment, you are robbing them from the
> opportunity to build and try a different set. You can have very varied
> weapons with vastly differing and interesting mechanics, but your players
> will never see that, because they will be stuck with the relatively good
> sword they found early on.

I think in the traditional roguelike the two features of permadeath and
procedural dungeon generation hang together. The more random and
surprising the dungeons are (with the proviso that you can in fact build
up a useful knowledge base relating to game strategy) the more
permadeath is indicated.

Conversely, if your game is highly scripted, and dungeons are relatively
balanced and not extraordinarily random, permadeath may be a mistake, as
you are forcing the player to replay boring content. Permadeath only
works if early content is always interesting.

My preference in more scripted games is to set you back a level on
death, with the dungeon that killed you being generated anew. That way
you are never trying to remember the contents of the map that killed
you, but you do not have to go back to level 1 unless you choose to do
so.

Of course this works best if you are always going down with no return as
in Rogue.

The rules must be suited to the game and the implementation. It is as
wrong to fetishise permadeath as it is to reject it out of hand in all
cases. [That is why I get a bad taste from 'losing is fun' type blurbs
on some modern roguelikes like Dredmor - they seem to miss the point.]

- Gerry Quinn

[Radomir Dopieralski]

On 2012-12-31, Gerry Quinn <ger...@gmail.com> wrote:
> In article <slrnke07m...@test.moinmo.in>, ne...@sheep.art.pl
> says...

[...]

>> Now, if you let them retain their equipment, you are robbing them from the
>> opportunity to build and try a different set. You can have very varied
>> weapons with vastly differing and interesting mechanics, but your players
>> will never see that, because they will be stuck with the relatively good
>> sword they found early on.
>
> I think in the traditional roguelike the two features of permadeath and
> procedural dungeon generation hang together. The more random and
> surprising the dungeons are (with the proviso that you can in fact build
> up a useful knowledge base relating to game strategy) the more
> permadeath is indicated.
>
> Conversely, if your game is highly scripted, and dungeons are relatively
> balanced and not extraordinarily random, permadeath may be a mistake, as
> you are forcing the player to replay boring content. Permadeath only
> works if early content is always interesting.
>
> My preference in more scripted games is to set you back a level on
> death, with the dungeon that killed you being generated anew. That way
> you are never trying to remember the contents of the map that killed
> you, but you do not have to go back to level 1 unless you choose to do
> so.
>
> Of course this works best if you are always going down with no return as
> in Rogue.

[...]

On the contrary. If the game is such that you can't return or undo your
earlier choices, and those choice affect the later game, the need to
"reset" the game to the starting position by permanent death is even
greater! You can think about it as about saves in the older games
like Doom -- you had to carefully pick the place to save, so that
reloading would get you back early enough to undo your mistake. If you
save with just one HP left and no medkits around, then you may need to
load an earlier save or start the game from the beginning. Same here, if
you die on this level, but due to a mistake on an earlier level,
regenerating the current level doesn't help -- you need to get the player
back somehow and let them try different options.

Of course, that's not just about mistakes. You also want them to try
different play styles and different choices in the early game -- otherwise
having those choices is a pure waste, if the player are never going to go
back to them. In games with long-term strategic decisions, perma-death is
necessary for experimenting.

--
Radomir Dopieralski, sheep.art.pl

[oot...@hot.ee]

On Wednesday, 2 January 2013 20:12:27 UTC+2, Radomir Dopieralski wrote:
> On 2012-12-31, Gerry Quinn <ger...@gmail.com> wrote:
> > My preference in more scripted games is to set you back a level on
> > death, with the dungeon that killed you being generated anew. That way
> > you are never trying to remember the contents of the map that killed
> > you, but you do not have to go back to level 1 unless you choose to do
> > so.
> >
> > Of course this works best if you are always going down with no return as
> > in Rogue.
>
> [...]
>
> On the contrary. If the game is such that you can't return or undo your
> earlier choices, and those choice affect the later game, the need to
> "reset" the game to the starting position by permanent death is even
> greater!

Yes, just one level might be is too feeble. Some may be baited to
suicide and grind that level again in hope to get something better from
RNG.

Permanent death is at other edge radical. Things too close to edges
are usually never optimal. We just likely love to implement permanent
death since it is clear and pure and simple to implement, not because
it provides optimal game-play.

There may be better balanced punishment between minus one level and
minus all the levels somewhere. It may be more optimal because
characters die also because of coincidence of "gotcha" features and
bosses they had not met before and that they had become too sleepy
and careless.

It is hard to make game such that it is random and dangerous and
non-boring but still does not contain much certain death situations.

> You can think about it as about saves in the older games
> like Doom -- you had to carefully pick the place to save, so that
> reloading would get you back early enough to undo your mistake. If you
> save with just one HP left and no medkits around, then you may need to
> load an earlier save or start the game from the beginning. Same here, if
> you die on this level, but due to a mistake on an earlier level,
> regenerating the current level doesn't help -- you need to get the player
> back somehow and let them try different options.

Yes, the feature to save the character at some fixed point is bad in a
randomly generated game. First, player bores himself to death by
reloading. Then with the OP gear combination that he scummed out the
game itself might be is too simple and boring too.

However lets take hypothetical example that player falls back long
way. For example 5 levels minimum or one third of his levels (if
that is bigger than 5) so permanently dies only under level 5.

That is not some specific "saved" spot. Being level 10 he falls
half game back, if he now dies again at level 8 then back to level
3. That is punishing and educating enough and not too scummable.
The only problem I see is that it may be is not too simple to
implement. :P

> Of course, that's not just about mistakes. You also want them to try
> different play styles and different choices in the early game -- otherwise
> having those choices is a pure waste, if the player are never going to go
> back to them.

Not so sure about that. Regardless of permanent death a player who is
happy with most of his choices plays 100 dwarf warriors in a row and
never tries kobold necromancers or some-such.

> In games with long-term strategic decisions, perma-death is necessary
> for experimenting.

Why? It is not mandatory in most games to be dead or won with one
character to experiment with other character. Also that can be made
as choice on case of death if he wants to fall back half of the way
or to restart.

[Gerry Quinn]

In article <1d943864-5a1e-424e...@googlegroups.com>,
oot...@hot.ee says...

> On Wednesday, 2 January 2013 20:12:27 UTC+2, Radomir Dopieralski wrote:
> > On 2012-12-31, Gerry Quinn <ger...@gmail.com> wrote:

> > > My preference in more scripted games is to set you back a level on
> > > death, with the dungeon that killed you being generated anew. That way
> > > you are never trying to remember the contents of the map that killed
> > > you, but you do not have to go back to level 1 unless you choose to do
> > > so.
> > >
> > > Of course this works best if you are always going down with no return as
> > > in Rogue.
> >

> > On the contrary. If the game is such that you can't return or undo your
> > earlier choices, and those choice affect the later game, the need to
> > "reset" the game to the starting position by permanent death is even
> > greater!
>
> Yes, just one level might be is too feeble. Some may be baited to
> suicide and grind that level again in hope to get something better from
> RNG.

[answering both posters]

I was toying with the idea that if you die once you restart the current
level (newly generated), but die again and you go back somewhat further.
Maybe the clock resets if you progress a few levels without death.

You could also have 'special' levels appearing occasionally in place of
a standard level. If you die during a special level, you are reborn on
a standard version of the same level. Could be a way to make special
levels more special.

> Permanent death is at other edge radical. Things too close to edges
> are usually never optimal. We just likely love to implement permanent
> death since it is clear and pure and simple to implement, not because
> it provides optimal game-play.
>
> There may be better balanced punishment between minus one level and
> minus all the levels somewhere. It may be more optimal because
> characters die also because of coincidence of "gotcha" features and
> bosses they had not met before and that they had become too sleepy
> and careless.
>
> It is hard to make game such that it is random and dangerous and
> non-boring but still does not contain much certain death situations.

I am thinking along the same lines.

> > You can think about it as about saves in the older games
> > like Doom -- you had to carefully pick the place to save, so that
> > reloading would get you back early enough to undo your mistake. If you
> > save with just one HP left and no medkits around, then you may need to
> > load an earlier save or start the game from the beginning. Same here, if
> > you die on this level, but due to a mistake on an earlier level,
> > regenerating the current level doesn't help -- you need to get the player
> > back somehow and let them try different options.

I don't agree. It is bad roguelike design if you are forced to die due
to a mistake on an earlier level. Unless it is good roguelike design
that if you blow all your resources in a tough fight with an OOD
monster, you will inevitably die within a few levels afterwards. Or
that you can choose abilities that are useless later on. There should
always be a way back, if you have some luck.

I know this is an idealisation. It is impossible to have a good game
where using up all your resources will not bring pain. But I think
roguelikes should in general operate on the principle that if there is
life, there is hope.

Anyway, nobody says the character must continue if he is repeatedly
hitting a brick wall. Anyone can start again, whether there is
permadeath or not.

> > Of course, that's not just about mistakes. You also want them to try
> > different play styles and different choices in the early game -- otherwise
> > having those choices is a pure waste, if the player are never going to go
> > back to them.
>
> Not so sure about that. Regardless of permanent death a player who is
> happy with most of his choices plays 100 dwarf warriors in a row and
> never tries kobold necromancers or some-such.

Yes.

- Gerry Quinn

[Radomir Dopieralski]

On 2013-01-03, oot...@hot.ee <oot...@hot.ee> wrote:
> On Wednesday, 2 January 2013 20:12:27 UTC+2, Radomir Dopieralski wrote:
>> On 2012-12-31, Gerry Quinn <ger...@gmail.com> wrote:
>> > My preference in more scripted games is to set you back a level on
>> > death, with the dungeon that killed you being generated anew. That way
>> > you are never trying to remember the contents of the map that killed
>> > you, but you do not have to go back to level 1 unless you choose to do
>> > so.
>> >
>> > Of course this works best if you are always going down with no return as
>> > in Rogue.
>>
>> [...]
>>
>> On the contrary. If the game is such that you can't return or undo your
>> earlier choices, and those choice affect the later game, the need to
>> "reset" the game to the starting position by permanent death is even
>> greater!
>
> Yes, just one level might be is too feeble. Some may be baited to
> suicide and grind that level again in hope to get something better from
> RNG.

That is not the problem that I'm trying to describe, but if you are afraid
of it, you can always have some "punishment" (or, as I prefer to call it,
cost) to dying, apart from generating a new level.

> Permanent death is at other edge radical. Things too close to edges
> are usually never optimal. We just likely love to implement permanent
> death since it is clear and pure and simple to implement, not because
> it provides optimal game-play.
>
> There may be better balanced punishment between minus one level and
> minus all the levels somewhere. It may be more optimal because
> characters die also because of coincidence of "gotcha" features and
> bosses they had not met before and that they had become too sleepy
> and careless.
>
> It is hard to make game such that it is random and dangerous and
> non-boring but still does not contain much certain death situations.

What I am trying to say here is that permadeath is not a punishment, or at
least it shouldn't be used as a punishment. It's more like a reset button
that lets you try solving a puzzle from beginning when it becomes obvious
that you messed it up so bad, that it is hard to recover. Do you think
that your game is so bad that making the player play more of it is
a punishment?

>> You can think about it as about saves in the older games
>> like Doom -- you had to carefully pick the place to save, so that
>> reloading would get you back early enough to undo your mistake. If you
>> save with just one HP left and no medkits around, then you may need to
>> load an earlier save or start the game from the beginning. Same here, if
>> you die on this level, but due to a mistake on an earlier level,
>> regenerating the current level doesn't help -- you need to get the player
>> back somehow and let them try different options.
>
> Yes, the feature to save the character at some fixed point is bad in a
> randomly generated game. First, player bores himself to death by
> reloading. Then with the OP gear combination that he scummed out the
> game itself might be is too simple and boring too.

I don't quite understand this argument. How is restarting a randomly
generated level multiple times more boring than visiting multiple randomly
generated levels in order? Logically, in both cases you are as likely to
find interesting situation and exciting experiences.

> However lets take hypothetical example that player falls back long
> way. For example 5 levels minimum or one third of his levels (if
> that is bigger than 5) so permanently dies only under level 5.
>
> That is not some specific "saved" spot. Being level 10 he falls
> half game back, if he now dies again at level 8 then back to level
> 3. That is punishing and educating enough and not too scummable.
> The only problem I see is that it may be is not too simple to
> implement. :P

The problem that I see is that a character from 3 days ago when I was
level 5 is going to be as unfamiliar to me as the one I start from
scratch. Or even more, as I might remember things that happened later than
the "save point", and be confused by that. "I could have sworn I had that
scroll of cure blindness!"

I really hate it when people refer to character death as punishment. It
is a punishment in those content-driven, plot-based games, where the
gameplay is just a way to prevent you from consuming the content too fast.
But roguelikes are games of skill, for the most part, and the fun in them
comes from playing them. You are not punishing someone by making them play
some more, provided that your game is fun to play, which it should be,
obviously.

>> Of course, that's not just about mistakes. You also want them to try
>> different play styles and different choices in the early game -- otherwise
>> having those choices is a pure waste, if the player are never going to go
>> back to them.
>
> Not so sure about that. Regardless of permanent death a player who is
> happy with most of his choices plays 100 dwarf warriors in a row and
> never tries kobold necromancers or some-such.

Race and class are not the best choices to illustrate this, because they
are supposed to be balanced in terms of difficulty (to a degree, at
least). But consider choices like "Do I equip this spear I found, or keep
my axe?", "Should I identify rings or potions first?", "Eat a food ration
or a slime mold?", "Use this wand of magic missile now or save it for
stronger monsters?", etc. Those kinds of choices usually have a huge
impact on the later game, so you really want to have a chance to
experiment a little. Obviously, you won't come up with the same situation
every time -- but that only makes it more interesting, as you need more
games to come up with a good heuristic for them.

>> In games with long-term strategic decisions, perma-death is necessary
>> for experimenting.
>
> Why? It is not mandatory in most games to be dead or won with one
> character to experiment with other character. Also that can be made
> as choice on case of death if he wants to fall back half of the way
> or to restart.

Why do you want to ask a question that everybody will answer the same?

--
Radomir Dopieralski, sheep.art.pl

[Radomir Dopieralski]

On 2013-01-04, Gerry Quinn <ger...@gmail.com> wrote:
>> > You can think about it as about saves in the older games
>> > like Doom -- you had to carefully pick the place to save, so that
>> > reloading would get you back early enough to undo your mistake. If you
>> > save with just one HP left and no medkits around, then you may need to
>> > load an earlier save or start the game from the beginning. Same here, if
>> > you die on this level, but due to a mistake on an earlier level,
>> > regenerating the current level doesn't help -- you need to get the player
>> > back somehow and let them try different options.
>
> I don't agree. It is bad roguelike design if you are forced to die due
> to a mistake on an earlier level. Unless it is good roguelike design
> that if you blow all your resources in a tough fight with an OOD
> monster, you will inevitably die within a few levels afterwards. Or
> that you can choose abilities that are useless later on. There should
> always be a way back, if you have some luck.
>
> I know this is an idealisation. It is impossible to have a good game
> where using up all your resources will not bring pain. But I think
> roguelikes should in general operate on the principle that if there is
> life, there is hope.

That idealisation has its embodiment! I know, you didn't think somebody
would a perfect game like that, but there it is. Lo and behold: Farmville.
No decision you make in that game will deter you from further play. Ever.

--
Radomir Dopieralski, sheep.art.pl

[Gerry Quinn]

In article <slrnkeega...@test.moinmo.in>, ne...@sheep.art.pl
says...

Actually Zangband might be a better example - except I suppose you could
get your scroll of recall burned. If you still have that scroll, you
can come back and scum low levels indefinitely.

And yet Zangband has permadeath. So permadeath is neither the cause nor
the cure of Farmville-like play.

What do you think of the graduated system I suggested? (Die once,
restart the level. Die again and you go back several levels.) Must we
choose between permadeath and Farmville?

- Gerry Quinn

[Radomir Dopieralski]

Of course permadeath has nothing to do with farmivilling. I was thinking
more of decisions that have no permanent consequences and thus make all
the play sessions pretty much the same. There is a saying where I live,
that say roughly "he, who never made a bad decision, never lived". And
I really love the feeling of empowerment and responsibility that comes
with the knowledge that whatever I do will have consequences.

Permadeath is just one of multiple tools you have to limit the damage that
the player can make to their characters, of course. It's a good tool, but
it is neither good for everything nor irreplaceable. You can make a deep,
strategical, interesting game with meaningful choices without it, if you
really wanted. But why limit yourself? It's there, it has a certain
effect, it solves a certain problem, you can use it. Don't dismiss it just
because other genres treat death as punishment.

As for the gradual system -- I honestly don't know. It sounds a little bit
like how lives work in some platformers -- you restart from the last save
point until you lose all your lives, at which point you restart at the
beginning. In those game the lives are a sort of a resource, so it's not
really the same. I think it's an excellent mechanic to test on a 7DRL.

By the way, remember the "CONTINUE? 9, 8, 7..." prompt in arcade games?
Even those need an optional permadeath!
--
Radomir Dopieralski, sheep.art.pl

[Gerry Quinn]

In article <slrnkeepu...@test.moinmo.in>, ne...@sheep.art.pl

says...
> On 2013-01-04, Gerry Quinn <ger...@gmail.com> wrote:

> > What do you think of the graduated system I suggested? (Die once,
> > restart the level. Die again and you go back several levels.) Must we
> > choose between permadeath and Farmville?
>
> Of course permadeath has nothing to do with farmivilling. I was thinking
> more of decisions that have no permanent consequences and thus make all
> the play sessions pretty much the same. There is a saying where I live,
> that say roughly "he, who never made a bad decision, never lived". And
> I really love the feeling of empowerment and responsibility that comes
> with the knowledge that whatever I do will have consequences.
>
> Permadeath is just one of multiple tools you have to limit the damage that
> the player can make to their characters, of course. It's a good tool, but
> it is neither good for everything nor irreplaceable. You can make a deep,
> strategical, interesting game with meaningful choices without it, if you
> really wanted. But why limit yourself? It's there, it has a certain
> effect, it solves a certain problem, you can use it. Don't dismiss it just
> because other genres treat death as punishment.

If you look back at my original post, you will see that I think
permadeath is more suitable for some kinds of roguelikes than others. I
am arguing that when certain conditions apply, it may not be a great
choice, and unthinkingly inserting it could make for a worse game.

> As for the gradual system -- I honestly don't know. It sounds a little bit
> like how lives work in some platformers -- you restart from the last save
> point until you lose all your lives, at which point you restart at the
> beginning. In those game the lives are a sort of a resource, so it's not
> really the same. I think it's an excellent mechanic to test on a 7DRL.

That is another extremely 'hard-edged' model, which is the opposite of
what I am interested in. If you need amulets of lifesaving and
suchlike, your permadeath model is breaking your game.

I was thinking more along the lines of: die once, restart level. Die
again, go back 3 levels. That sort of thing. Maybe -3 levels from
where you are every time after the first time you die until you have
progressed three levels without death (after that a single new death
will just cause you to restart the current level). Of course it would
need to be balanced with the full game design.

- Gerry Quinn

[Radomir Dopieralski]

On 2013-01-05, Gerry Quinn <ger...@gmail.com> wrote:
> If you look back at my original post, you will see that I think
> permadeath is more suitable for some kinds of roguelikes than others. I
> am arguing that when certain conditions apply, it may not be a great
> choice, and unthinkingly inserting it could make for a worse game.

I'm sorry, I forgot that. You are of course right.

>> As for the gradual system -- I honestly don't know. It sounds a little bit
>> like how lives work in some platformers -- you restart from the last save
>> point until you lose all your lives, at which point you restart at the
>> beginning. In those game the lives are a sort of a resource, so it's not
>> really the same. I think it's an excellent mechanic to test on a 7DRL.
>
> That is another extremely 'hard-edged' model, which is the opposite of
> what I am interested in. If you need amulets of lifesaving and
> suchlike, your permadeath model is breaking your game.
>
> I was thinking more along the lines of: die once, restart level. Die
> again, go back 3 levels. That sort of thing. Maybe -3 levels from
> where you are every time after the first time you die until you have
> progressed three levels without death (after that a single new death
> will just cause you to restart the current level). Of course it would
> need to be balanced with the full game design.

Hmm, Prince of Persia and the Sand of Time?

--
Radomir Dopieralski, sheep.art.pl

[oot...@hot.ee]

On Friday, 4 January 2013 20:28:06 UTC+2, Radomir Dopieralski wrote:
> On 2013-01-03, oot...@hot.ee <oot...@hot.ee> wrote:
> > On Wednesday, 2 January 2013 20:12:27 UTC+2, Radomir Dopieralski wrote:
> >> On 2012-12-31, Gerry Quinn <ger...@gmail.com> wrote:
> >> > My preference in more scripted games is to set you back a level on
> >> > death, with the dungeon that killed you being generated anew. That way
> >> > you are never trying to remember the contents of the map that killed
> >> > you, but you do not have to go back to level 1 unless you choose to do
> >> > so.
> >> >
> >> > Of course this works best if you are always going down with no return as
> >> > in Rogue.
> >>
> >> [...]
> >>
> >> On the contrary. If the game is such that you can't return or undo your
> >> earlier choices, and those choice affect the later game, the need to
> >> "reset" the game to the starting position by permanent death is even
> >> greater!
> >
> > Yes, just one level might be is too feeble. Some may be baited to
> > suicide and grind that level again in hope to get something better from
> > RNG.
>
> That is not the problem that I'm trying to describe, but if you are afraid
> of it, you can always have some "punishment" (or, as I prefer to call it,
> cost) to dying, apart from generating a new level.

It is just terminology ... positive feedback I call bonus, negative feedback
I call punishment. "Cost" I use when there is some exchange of one good
to other.

He tried to mix conflicting combo of race, class, deity and gear. He
trained it with too narrow or too broad set of skills. To compensate it
he may need valuable consumables or (for more optimal build) overpowered
equipment. If it is possible to grind for gear and/or consumables he has
to grind for long time. That is cost. If he does not or it is not possible
to grind then he dies. That is punishment. If he loses just one level then
the dying can be transformed into grind (a cost). That is the problem how
I see it. Now you say it is not the problem? What is it then?

> > Permanent death is at other edge radical. Things too close to edges
> > are usually never optimal. We just likely love to implement permanent
> > death since it is clear and pure and simple to implement, not because
> > it provides optimal game-play.
> >
> > There may be better balanced punishment between minus one level and
> > minus all the levels somewhere. It may be more optimal because
> > characters die also because of coincidence of "gotcha" features and
> > bosses they had not met before and that they had become too sleepy
> > and careless.
> >
> > It is hard to make game such that it is random and dangerous and
> > non-boring but still does not contain much certain death situations.
>
> What I am trying to say here is that permadeath is not a punishment, or at
> least it shouldn't be used as a punishment. It's more like a reset button
> that lets you try solving a puzzle from beginning when it becomes obvious
> that you messed it up so bad, that it is hard to recover. Do you think
> that your game is so bad that making the player play more of it is
> a punishment?

You play with words here indicating like I said that playing the game
itself is punishment. No, but character death should be, if it is not
then I do not get the concept.

I consider game as interesting for player when player feels connected to
the character and cares about its adventures. So he should feel loss
(and being punished) when the character dies. Notice, that we are both
aiming at same goal, to provide more enjoyable gameplay to players, but
I do not understand why you think that permadeath is somehow especially
enjoyable feature and anything less would take something away?

> >> You can think about it as about saves in the older games
> >> like Doom -- you had to carefully pick the place to save, so that
> >> reloading would get you back early enough to undo your mistake. If you
> >> save with just one HP left and no medkits around, then you may need to
> >> load an earlier save or start the game from the beginning. Same here, if
> >> you die on this level, but due to a mistake on an earlier level,
> >> regenerating the current level doesn't help -- you need to get the player
> >> back somehow and let them try different options.
> >
> > Yes, the feature to save the character at some fixed point is bad in a
> > randomly generated game. First, player bores himself to death by
> > reloading. Then with the OP gear combination that he scummed out the
> > game itself might be is too simple and boring too.

> I don't quite understand this argument. How is restarting a randomly
> generated level multiple times more boring than visiting multiple randomly
> generated levels in order? Logically, in both cases you are as likely to
> find interesting situation and exciting experiences.

Typically the situation gets more hard together with game progress.
If it is possible to farm on earlier level to get best out of it so next
level is easy, then farm there etc. then that is what players will do.
Scumming until ideally prepared is clearly optimal for to ensure victory.

How hard such repetitive farming can be? It is not hard so I imagine they
should feel less of accomplishment doing it. If to prevent that farming
then players have to face the dangers half-prepared or at least not
ideally prepared and logically it is more fun and exiting. Unfortunately
there is bigger chance to generate "too exiting" situations that way.

> > However lets take hypothetical example that player falls back long
> > way. For example 5 levels minimum or one third of his levels (if
> > that is bigger than 5) so permanently dies only under level 5.
> >
> > That is not some specific "saved" spot. Being level 10 he falls
> > half game back, if he now dies again at level 8 then back to level
> > 3. That is punishing and educating enough and not too scummable.
> > The only problem I see is that it may be is not too simple to
> > implement. :P
>
> The problem that I see is that a character from 3 days ago when I was
> level 5 is going to be as unfamiliar to me as the one I start from
> scratch. Or even more, as I might remember things that happened later than
> the "save point", and be confused by that. "I could have sworn I had that
> scroll of cure blindness!"

It can not be that mistakes made more than 3 days ago are primary
long-term cause of that death so it somehow nullifies your main
argument for it?

> I really hate it when people refer to character death as punishment. It
> is a punishment in those content-driven, plot-based games, where the
> gameplay is just a way to prevent you from consuming the content too fast.
> But roguelikes are games of skill, for the most part, and the fun in them
> comes from playing them. You are not punishing someone by making them play
> some more, provided that your game is fun to play, which it should be,
> obviously.

In gaming world there are no ultimate concepts of fun, otherwise we all
would try to develop that one, ultimate game. For me finding good gear
is so easy to consider as positive event (so bonus) and dying is so easy
to consider as negative event (so punishment). If your concept is too
far from that then it is indeed hard to discuss.

So they did fall few levels back instead of dying. What prevents them now
from playing it? Some players play sometimes long hours, when character dies
then they often close the game and go to forum to write about their YASD
that no one reads. It is more likely that it was too harsh for them.

> >> Of course, that's not just about mistakes. You also want them to try
> >> different play styles and different choices in the early game -- otherwise
> >> having those choices is a pure waste, if the player are never going to go
> >> back to them.
> >
> > Not so sure about that. Regardless of permanent death a player who is
> > happy with most of his choices plays 100 dwarf warriors in a row and
> > never tries kobold necromancers or some-such.
>
> Race and class are not the best choices to illustrate this, because they
> are supposed to be balanced in terms of difficulty (to a degree, at
> least).

??? Most games even indicate that some combinations are simpler, some
harder and some very hard. That can be considered like a way to configure
the difficulty of game. Permadeath can be also made configurable. One of
the most hard roguelikes, DCSS, has made it configurable that way ... you
can play feline there, that has more than one life.

> But consider choices like "Do I equip this spear I found, or keep
> my axe?", "Should I identify rings or potions first?", "Eat a food ration
> or a slime mold?", "Use this wand of magic missile now or save it for
> stronger monsters?", etc. Those kinds of choices usually have a huge
> impact on the later game, so you really want to have a chance to
> experiment a little. Obviously, you won't come up with the same situation
> every time -- but that only makes it more interesting, as you need more
> games to come up with a good heuristic for them.

Yes, but management of short term stuff like consumables has shorter
term affect than choice of races, classes, skills and religions. The players
who do not know the basics, like when and how to identify consumables or
when to use those will likely die early in balanced game. I do not argue
that it is OK to let them to permadie on their early levels and to retry the
basics.

The players who have gotten farther have apparently learned to survive the
first levels. It may be not so fun for them to restart that often so it may
be more optimal to offer them option to fall back so they can retry
the part where they apparently did something wrong.

If the novice really just did as single luck out coincidence event to get to
level 10 then after falling to level 5 it is very likely that they die again
at level 6 and so your retry mechanics are still in effect, just slightly
smoothened and delayed.

> >> In games with long-term strategic decisions, perma-death is necessary
> >> for experimenting.
> >
>
> > Why? It is not mandatory in most games to be dead or won with one
> > character to experiment with other character. Also that can be made
> > as choice on case of death if he wants to fall back half of the way
> > or to restart.
>
> Why do you want to ask a question that everybody will answer the same?

I do not think that it is such a question at all. Players who master
roguelikes can very well think on their own analytically. I see no need
to solve that puzzle for them. Player gets fed up of the combo that
constantly dies and so gets nowhere, "dump it, its unlucky, I'll try
somethin else now." That is the benefit of choice, that he feels less
punished even if the outcome is same.

[Radomir Dopieralski]

On 2013-01-05, oot...@hot.ee <oot...@hot.ee> wrote:
[...]

> It is just terminology ... positive feedback I call bonus, negative feedback
> I call punishment. "Cost" I use when there is some exchange of one good
> to other.

But death in roguelikes is not a negative feedback. It's inevitable. It
can be very fun (otherwise people wouldn't write all those YASD posts).
There is a direct correlation between boredom and avoiding death.

>> What I am trying to say here is that permadeath is not a punishment, or at
>> least it shouldn't be used as a punishment. It's more like a reset button
>> that lets you try solving a puzzle from beginning when it becomes obvious
>> that you messed it up so bad, that it is hard to recover. Do you think
>> that your game is so bad that making the player play more of it is
>> a punishment?
>
> You play with words here indicating like I said that playing the game
> itself is punishment. No, but character death should be, if it is not
> then I do not get the concept.
>
> I consider game as interesting for player when player feels connected to
> the character and cares about its adventures. So he should feel loss
> (and being punished) when the character dies. Notice, that we are both
> aiming at same goal, to provide more enjoyable gameplay to players, but
> I do not understand why you think that permadeath is somehow especially
> enjoyable feature and anything less would take something away?

I indeed think that in roguelike games character death should not be
a punishment. Why should I care about randomly generated adventures that
I can't affect in any way? Why should I feel loss when one set of random
future events is replaced with another set of random future events,
exactly as rich in exciting (randomly generated) content and fun? Is it
because this particular combination was one chance in a million, and any
other combination generated later will not be as enjoyable? But it can as
well be much more enjoyable!

[...]

>> > Yes, the feature to save the character at some fixed point is bad in a
>> > randomly generated game. First, player bores himself to death by
>> > reloading. Then with the OP gear combination that he scummed out the
>> > game itself might be is too simple and boring too.
>
>> I don't quite understand this argument. How is restarting a randomly
>> generated level multiple times more boring than visiting multiple randomly
>> generated levels in order? Logically, in both cases you are as likely to
>> find interesting situation and exciting experiences.
>
> Typically the situation gets more hard together with game progress.
> If it is possible to farm on earlier level to get best out of it so next
> level is easy, then farm there etc. then that is what players will do.
> Scumming until ideally prepared is clearly optimal for to ensure victory.
>
> How hard such repetitive farming can be? It is not hard so I imagine they
> should feel less of accomplishment doing it. If to prevent that farming
> then players have to face the dangers half-prepared or at least not
> ideally prepared and logically it is more fun and exiting. Unfortunately
> there is bigger chance to generate "too exiting" situations that way.

Be careful, you are conflating two different kinds of "harder" here, and
that leads to a paradox. You have to distinguish between "harder to master
because of complexity" and "harder to beat because of lower probability of
success". A game should indeed start relatively simple, and become more
complicated as you progress, introducing new factors to consider in your
strategy. But farming doesn't help with this kind of complexity -- you
still have to get out and face it at some point. The other kind of
difficulty, the one to which farming is an answer, is not good to have.
It is connected with gambling and leads to all sorts of addictive
mechanics.

I don't understand what you mean by "exiting" and "too exiting" here. Do
you mean that there is an exit from the situation?

>> > However lets take hypothetical example that player falls back long
>> > way. For example 5 levels minimum or one third of his levels (if
>> > that is bigger than 5) so permanently dies only under level 5.
>> >
>> > That is not some specific "saved" spot. Being level 10 he falls
>> > half game back, if he now dies again at level 8 then back to level
>> > 3. That is punishing and educating enough and not too scummable.
>> > The only problem I see is that it may be is not too simple to
>> > implement. :P
>>
>> The problem that I see is that a character from 3 days ago when I was
>> level 5 is going to be as unfamiliar to me as the one I start from
>> scratch. Or even more, as I might remember things that happened later than
>> the "save point", and be confused by that. "I could have sworn I had that
>> scroll of cure blindness!"
>
> It can not be that mistakes made more than 3 days ago are primary
> long-term cause of that death so it somehow nullifies your main
> argument for it?

If my decisions from 3 days ago have no effect on my character is today,
then those decisions could have been completely removed from the game, as
they are meaningless and only obscure the real, meaningful decisions that
I need to make to master the game.

>> I really hate it when people refer to character death as punishment. It
>> is a punishment in those content-driven, plot-based games, where the
>> gameplay is just a way to prevent you from consuming the content too fast.
>> But roguelikes are games of skill, for the most part, and the fun in them
>> comes from playing them. You are not punishing someone by making them play
>> some more, provided that your game is fun to play, which it should be,
>> obviously.
>
> In gaming world there are no ultimate concepts of fun, otherwise we all
> would try to develop that one, ultimate game. For me finding good gear
> is so easy to consider as positive event (so bonus) and dying is so easy
> to consider as negative event (so punishment). If your concept is too
> far from that then it is indeed hard to discuss.
>
> So they did fall few levels back instead of dying. What prevents them now
> from playing it? Some players play sometimes long hours, when character dies
> then they often close the game and go to forum to write about their YASD
> that no one reads. It is more likely that it was too harsh for them.

There is an extensive body of research about fun in general and fun in
games in particular. As you say there are many ways to have fun (four, to
be exact), but it is not completely arbitrary. In particular, some sources
of fun and pleasure are sustainable, while others give diminishing results
and lead to addiction. Finding good gear falls in there, unfortunately,
together with "leveling up". That's why grinding is boring and unfun.

On the other hand, there is nothing unfun in dying itself -- provided that
it doesn't lead to more grinding. If it's a permanent death, then you are
free -- you can go and share your exciting experience with others on the
forum, or contemplate it quietly yourself and use it to improve your play
next time. On the other hand, if you are merely sent a few level back,
then the message is clear: "you are not ready yet, go grind some and come
back then". You have learned that from other games -- where dying leads to
mandatory grinding which is the real punishment. But in roguelikes
permanent death actually removes that problem -- you *can't* grind, even
if you wanted, because the character is gone, dead, and all you can do is
not make the same mistake with the next character.

>> >> Of course, that's not just about mistakes. You also want them to try
>> >> different play styles and different choices in the early game -- otherwise
>> >> having those choices is a pure waste, if the player are never going to go
>> >> back to them.
>> >
>> > Not so sure about that. Regardless of permanent death a player who is
>> > happy with most of his choices plays 100 dwarf warriors in a row and
>> > never tries kobold necromancers or some-such.
>>
>> Race and class are not the best choices to illustrate this, because they
>> are supposed to be balanced in terms of difficulty (to a degree, at
>> least).
>
> ??? Most games even indicate that some combinations are simpler, some
> harder and some very hard. That can be considered like a way to configure
> the difficulty of game. Permadeath can be also made configurable. One of
> the most hard roguelikes, DCSS, has made it configurable that way ... you
> can play feline there, that has more than one life.

Well, if I told you "you can't win unless you choose the right race at the
beginning", you would tell me that the game is broken, wouldn't you?
But if I told you "you have to equip the good gear that you find,
otherwise you can't win", you wouldn't object.

>> But consider choices like "Do I equip this spear I found, or keep
>> my axe?", "Should I identify rings or potions first?", "Eat a food ration
>> or a slime mold?", "Use this wand of magic missile now or save it for
>> stronger monsters?", etc. Those kinds of choices usually have a huge
>> impact on the later game, so you really want to have a chance to
>> experiment a little. Obviously, you won't come up with the same situation
>> every time -- but that only makes it more interesting, as you need more
>> games to come up with a good heuristic for them.
>
> Yes, but management of short term stuff like consumables has shorter
> term affect than choice of races, classes, skills and religions. The players
> who do not know the basics, like when and how to identify consumables or
> when to use those will likely die early in balanced game. I do not argue
> that it is OK to let them to permadie on their early levels and to retry the
> basics.
>
> The players who have gotten farther have apparently learned to survive the
> first levels. It may be not so fun for them to restart that often so it may
> be more optimal to offer them option to fall back so they can retry
> the part where they apparently did something wrong.
>
> If the novice really just did as single luck out coincidence event to get to
> level 10 then after falling to level 5 it is very likely that they die again
> at level 6 and so your retry mechanics are still in effect, just slightly
> smoothened and delayed.

This is not a game of Sokoban, or Portal, or any similar puzzle game,
where you have one tactical skill per level to learn, and you will not
finish that level until you master that skill. If it was so, there would
be no point in re-playing any of the earlier levels ever. After all, the
player has mastered lesson hidden in them.

If you tried to look for analogy to Portal's levels in a roguelike, they
would be whole playing sessions. They are lessons that get longer and
longer each time, because they teach progressively more subtle and
delicate things.

Sure, the first few sessions will be just the first level -- until you
learn to move around efficiently, fight monsters in right way, etc.

Then, as your game sessions become longer, you learn to equip better gear,
to plan when to drink potions and which items to best identify, you learn
how to deal with certain categories of trouble, like blindness or out of
depth monsters, etc.

As your playing sessions become even longer, you learn to optimize for the
long term. You learn what resistances you need at the specific level and
how to get them, you try out different playing styles and find the best
ones for particular race/class/gear combinations. You minimize the effect
of luck on your games.

And just like in Portal you have those disintegration fields between
levels, preventing you from using things from the previous level, you have
permanent death in roguelikes, preventing the previous session from
affecting the new game. All you can take with you is what you have
learned.

>> >> In games with long-term strategic decisions, perma-death is necessary
>> >> for experimenting.
>> >
>>
>> > Why? It is not mandatory in most games to be dead or won with one
>> > character to experiment with other character. Also that can be made
>> > as choice on case of death if he wants to fall back half of the way
>> > or to restart.
>>
>> Why do you want to ask a question that everybody will answer the same?
>
> I do not think that it is such a question at all. Players who master
> roguelikes can very well think on their own analytically. I see no need
> to solve that puzzle for them. Player gets fed up of the combo that
> constantly dies and so gets nowhere, "dump it, its unlucky, I'll try
> somethin else now." That is the benefit of choice, that he feels less
> punished even if the outcome is same.

This is one additional decision that you are requiring of your players,
and this decision is very hard. Actually, it's so hard, that the players
are unable to make the correct choice with the limited information they
have at the point where they are still learning (so while they are having
fun). Only a player who has already mastered the particular problem can
with confidence say "ok, this way doesn't lead to a solution, try again".
But a player who has already mastered it doesn't need to repeat it.

Not only you make your players struggle in boring grinding, but now you
are also blaming it on them, while they can't possibly make the correct
decision themselves. Very unfun.

--
Radomir Dopieralski, sheep.art.pl

[oot...@hot.ee]

On Friday, 4 January 2013 19:24:29 UTC+2, Gerry Quinn wrote:
> In article <1d943864-5a1e-424e...@googlegroups.com>,

> I was toying with the idea that if you die once you restart the current
> level (newly generated), but die again and you go back somewhat further.
> Maybe the clock resets if you progress a few levels without death.

That is sort of extra life that quite lot of other (not rogue-like) games have.
It is often for games with more linear and less random contents, but
certainly worth trying with random content roguelike too.

Dangers that I see are that it may suggest players to suicide on case of
getting some more permanent disability like injury, hard disease or
bad mutation instead of trying the challenge to live with it.

Also it may provide way to bypass that early identification minigame
(try all that stuff, take notes, suicide). Features that player will ignore
or bypass are not worth implementing then.

> You could also have 'special' levels appearing occasionally in place of
> a standard level. If you die during a special level, you are reborn on
> a standard version of the same level. Could be a way to make special
> levels more special.

If 'special' means more dangerous then yes, you can try to make a game that
adapts to player's skill. However permanent death (if early game is simpler)
serves that purpose too. If early game has just different challenges (that the
player may have already mastered), then permadeath might waste his time.

Even without permanent death we still should aim to make the game
winnable without dying once and such victory should be is rewarded
somehow. Bigger score certainly but I think about unlocking some
content, can be some harder side-quests or challenging races/classes
or just cosmetics like special awards/titles or things like that.

> It is bad roguelike design if you are forced to die due
> to a mistake on an earlier level. Unless it is good roguelike design
> that if you blow all your resources in a tough fight with an OOD
> monster, you will inevitably die within a few levels afterwards. Or
> that you can choose abilities that are useless later on. There should
> always be a way back, if you have some luck.

Depends. I imagine exiting rogue-like as such that there should be majority
of encounters where it is possible to decide if to avoid or if to escape or if to
fight a battle. The tactical situation should be also possible to engineer, to
improve the odds and to optimize costs.

All decisions should have some downside or cost to make it harder and more
meaningful to decide. If player uses up all of his resources because of sub-
optimal decisions then he deserves to be in difficult situation.

> I know this is an idealisation. It is impossible to have a good game
> where using up all your resources will not bring pain. But I think
> roguelikes should in general operate on the principle that if there is
> life, there is hope.

It should certainly bring danger, the odds should be against him. It would
be nicer if his odds to survive were always noticeably bigger than total
0, but in random game that might be is not so simple to implement.

> Anyway, nobody says the character must continue if he is repeatedly
> hitting a brick wall. Anyone can start again, whether there is
> permadeath or not.

Yes. Especially if player did pick some challenging build then learning if
he can get somewhere or not with it takes time.

I still think that tossing him somewhat farther back from that "brick wall"
might be better. Totally to tile 0 (permadeath) feels fair and clear, but it
might be is not optimal if it is too far from that wall. Only a single step
back may be is insufficient since it may encourage him to reroll the
dice of that head-hit instead of stepping aside and rethinking and
preparing better and earlier. Your idea of providing him one retry before
throwing him farther back serves that purpose too.

[oot...@hot.ee]

On Sunday, 6 January 2013 02:30:32 UTC+2, Radomir Dopieralski wrote:
> On 2013-01-05, oot...@hot.ee <oot...@hot.ee> wrote:
> > It is just terminology ... positive feedback I call bonus, negative feedback
> > I call punishment. "Cost" I use when there is some exchange of one good
> > to other.
>
> But death in roguelikes is not a negative feedback. It's inevitable. It
> can be very fun (otherwise people wouldn't write all those YASD posts).
> There is a direct correlation between boredom and avoiding death.

Inevitable death means that it is not winnable with your wit and without
spoilers or studying the source code or playing loong time. Sure it is fun
for some. I somewhat tolerate Nethack, but not like some who seem to
love it. Tastes vary.

I like hard that takes strategic skills most, I like hard that takes long hours
invested into repetitive situations least. The hard that comes from gotcha
situations that I did not know about is somewhere between.

> If my decisions from 3 days ago have no effect on my character is today,
> then those decisions could have been completely removed from the game, as
> they are meaningless and only obscure the real, meaningful decisions that
> I need to make to master the game.

Yes, that I implied. If advancing one character 5 levels takes 3 days then playing
20 characters to level 20 takes 240 days. It can not fit with taste of lot of
players so the game could benefit if to remove some more repetitive content
from it. That is matter of taste perhaps. I can not imagine how your decision
in early game if to identify scrolls or rings first could affect that battle with
late game dragons anyhow.

> On the other hand, there is nothing unfun in dying itself -- provided that
> it doesn't lead to more grinding.

Getting punished for your wrong behavior is part of game. That is fun. It is
still negative feedback (a punishment). I have not advocated removal of such
feedback. I have only suggested to make it less harsh to improve enjoyability
of game. The early game can not be removed or made lot harder,
otherwise the game becomes specialist-only and stops attracting novices.

> If it's a permanent death, then you are
> free -- you can go and share your exciting experience with others on the
> forum, or contemplate it quietly yourself and use it to improve your play
> next time. On the other hand, if you are merely sent a few level back,
> then the message is clear: "you are not ready yet, go grind some and come
> back then". You have learned that from other games -- where dying leads to
> mandatory grinding which is the real punishment.

No, I learned it years ago when I tried to play Angband. It is major roguelike
game that provides exactly the unfun what you describe. Endless grind.
Numerous players love it, i could not understand why. Even the humorous
nature of Nethack trap gotcha you are dead is missing.

> Well, if I told you "you can't win unless you choose the right race at the
> beginning", you would tell me that the game is broken, wouldn't you?
> But if I told you "you have to equip the good gear that you find,
> otherwise you can't win", you wouldn't object.

Why impossible? I merely say that it is suboptimal to play combinations
that have less odds to win unless player can win too regularly with easier
builds.

I certainly understand a player who feels offended when "you have to
equip the good gear that you find, otherwise you can't win" is said to
him. Saying to him simply that "I think that you are idiot" would feel
less insulting to majority of people.

> Sure, the first few sessions will be just the first level -- until you
> learn to move around efficiently, fight monsters in right way, etc.

Redoing that lesson is wasted once mastered.

> Then, as your game sessions become longer, you learn to equip better gear,
> to plan when to drink potions and which items to best identify, you learn
> how to deal with certain categories of trouble, like blindness or out of
> depth monsters, etc.

Identifying minigame is perhaps the only thing that keeps early levels
different because on later levels the character will have most types of
consumables and jewelry identified. Once that minigame is clear then it
provides no much challenge nor additional value to next game.

> As your playing sessions become even longer, you learn to optimize for the
> long term. You learn what resistances you need at the specific level and
> how to get them, you try out different playing styles and find the best
> ones for particular race/class/gear combinations. You minimize the effect
> of luck on your games.

Player has to play long game of less dangerous low levels even to get
somewhere where resistance might be available and useful or where he
has more strategic options than just to cast magic missiles or to hack with
his dagger. That could be improved. I fully understand that you do not
feel that to be wasting of player's time. I think it might be.

> > I do not think that it is such a question at all. Players who master
> > roguelikes can very well think on their own analytically. I see no need
> > to solve that puzzle for them. Player gets fed up of the combo that
> > constantly dies and so gets nowhere, "dump it, its unlucky, I'll try
> > somethin else now." That is the benefit of choice, that he feels less
> > punished even if the outcome is same.
>
> This is one additional decision that you are requiring of your players,
> and this decision is very hard. Actually, it's so hard, that the players
> are unable to make the correct choice with the limited information they
> have at the point where they are still learning (so while they are having
> fun). Only a player who has already mastered the particular problem can
> with confidence say "ok, this way doesn't lead to a solution, try again".
> But a player who has already mastered it doesn't need to repeat it.

He has to. If there is a nuance he has not mastered in deepest depth of
hell with particular character build then in order to get there and to try
something else than last time he has to play all the game down to it.
I am not arguing that some players certainly are fit to do it over and
over but I am not sure why you think that it is enjoyable for everybody.
Also i do not see how dying is so extremely fun but picking between
death or retrying last third of life of the character is so hard.

> Not only you make your players struggle in boring grinding, but now you
> are also blaming it on them, while they can't possibly make the correct
> decision themselves. Very unfun.

You must somehow find balance between your accusations. In one
place you say that the player has even forgot what he did 5 levels ago
so that he looks at the character as totally unknown. In here you say
that replaying these 5 levels is grinding. Make up your mind,
playing a character that you have forgot even what its plans were can
not be less fun than to play totally new character. There is just
the benefit of opportunity to start with somewhat more advanced
character, not too huge benefit if restarting is so fun.

[The Magnificent Zamdan]

WARNING: Long. Set aside some time, or skip. TL;DR: a whole lot of ideas
and theory for tweaking difficulty, giving the player some sort of finite
get out of jail free card resource or other free pass against a certain
amount of bad luck, and ways to discourage or prevent grinding without
*completely* preventing the player getting at least a *bit* more powerful
for an obstacle they're having trouble passing. Some of the ideas are ones
I've never seen implemented anywhere. Much of it's tangential to the
death-issue or assumes permadeath, but some, especially near the start and
end, tries to balance setback-death.

If there's any capsule summary, it's that if resources are finite, bad luck
can cause "screwed by the RNG", but adding one more finite resource that's
unbalancingly strong in any single encounter but unreplenishable gives the
player a way out of such situations that they'd otherwise prefer to reserve
until later in the game than squander. Adding more "width" to the dungeon
allows grinding to overlevel or overequip, but this can be limited or even
made a finite resource (e.g. through consumable, rare reset-level scrolls
the player starts with a few of).

Alternatively, if a specific resource being squandered over the course of a
game tends to be how one gets screwed (especially, screwed in a way that
setback-death would leave you still screwed and forced to restart from
scratch anyway), it can be made impossible to take from level to level and
the game balanced so there's enough on each level, normally, and it can be
managed on a per-level basis. That fixes setback-death for that sort of
screwage.

On Sat, 5 Jan 2013 16:37:00 -0800 (PST), oot...@hot.ee wrote:
> Dangers that I see are that it may suggest players to suicide on case of
> getting some more permanent disability like injury, hard disease or
> bad mutation instead of trying the challenge to live with it.

It may be worth considering whether such permanent bad statuses should even
exist to begin with. (As opposed to tradeoffs, where the player's choices
can result in two or more mutually exclusive outcomes, each with a good
side and a bad side, say. Choosing a god, in games with such a mechanic,
comes to mind.)

> Also it may provide way to bypass that early identification minigame
> (try all that stuff, take notes, suicide). Features that player will ignore
> or bypass are not worth implementing then.

Ah, but that one's simple: lose knowledge of everything learned in the way
of item flavors etc. since the last checkpoint, and then reshuffle all the
non-learned flavors. So if you learn green potion is fire resistance at
level 2, purple potion is haste at level 5, and yellow potion is nasty
poison at level 6, then kill yourself, you may wind up at level 4 with
green still fire resistance but purple not necessarily haste and yellow not
necessarily nasty poison. As if it was a quantum thing and you'd
un-observed it by going back in time to before you'd seen it.

Alternatively, the setback in the earliest part of the game can be "start
over" and later on you'll have invested too much in the character to want
to be set back a large number of levels.

An under-the-hood way to implement something like this would be to have the
savefile contain next-level autosave states for the last few levels, plus
the manual save. Dying reloads the appropriate autosave and discards
everything later, then reshuffles unknown flavors. You lose items and
equipment acquired since the point you're sent back to, as well as
experience gained, etc.

> If 'special' means more dangerous then yes, you can try to make a game that
> adapts to player's skill. However permanent death (if early game is simpler)
> serves that purpose too. If early game has just different challenges (that the
> player may have already mastered), then permadeath might waste his time.

I'd say that the condition for permadeath to be appropriate is that either

a) The whole game is fairly short, if the player doesn't do a lot of
grinding, and there isn't much mileage in doing a lot of grinding
early; or

b) The game has so *much* variability that the early game is fresh
and interesting even 100 times in rapid succession. That will
mean a lot more than just random mazes and a chance of out-of-depth
items and monsters, clearly.

One possibility that comes to mind is to categorize items into
various groups and monsters into various groups, and also chunk them
farther by levels (e.g. 1-3, 4-8, 9-12...), and in each
category-level chunk make some of the items (resp. monsters)
unavailable, rare, ten levels (or more) deeper, not generatable near
the starting town (if there's geography) so "foreign and exotic", or
some such, and not available in (starting town) shops (if
applicable).

If this is balanced, having, say, only any 3 of the 5 lowest-level
healing-type items easily available in a particular game would not
make the game too hard but might sometimes require different
strategies. Each playthrough encounters somewhat different mixtures
of monsters, too, and perhaps even randomly-selectable dungeon
features (particular sorts of rooms, vaults, etc.) which may result
in good and bad sides (no early poisoners! Yay! But no early dwarf
mages either, which have glass jaws and often drap wands. Damn!) but
not game-breakingly or show-stoppingly so. Ideally, the game would
truly be quite different every time.

Probably, if the player travels far enough or gets deep enough they
should still eventually be able to find everything. Unique items and
monsters might particularly be exempted (either altogether, or else
from being possibly *never* available in that game even if other
things are, moving them deeper or making them rarer instead).
Obviously, any plot-necessary items should be present (unless
there's several alternative plot resolutions, where one must only
avoid foreclosing on *all* possible paths to player victory, or else
there is no plot, just "get to the bottom" or "live as long as
possible and get as much loot as possible"; even a bare-bones plot
of "kill specific monster X", e.g. Angband's Morgoth, requires
monster X not be suppressed from being able to be generated).

Something that can stir the above up even more would be to pick a
small number of random items, monsters, and anything else generated
similarly to make "commonly encountered out of depth": make it
shallower and/or less rare. Caution would be needed here that it's
not going to result in a game-breaker (ammo of one-shot-murderize-
everything abundant at level 1) or a show-stopper (monster of
instantly-sense-teleport-to-and-cinderize-level-1-player abundant at
level 1). Generally that means either very strong weapons, or
monsters requiring either a large HP pool or specific defenses to
survive encountering and that aren't easy to just dodge or get the
defenses to at low levels. Not making them able to be *too* out
of depth (so a very limited number of things from up to 4 levels
deeper might be moved up to a given level, but not from 10, let alone
40, perhaps relaxing this later, or using a proportionality rule,
such as no more than min(4, level/3) deeper) would be indicated, as
might disabling this possibility for some specific things
(absolutely no dragons moved shallower than level 15, because fire
protection gear doesn't become abundant until level 13; no weapons
with the quintuple damage vs. dragons flag moved shallower than
level 20, because dragons should pose a challenge for a *few*
levels; getting either early should always require genuine luck,
good or bad, from the RNG at the time, rather than being in an
RNG-blessed (or cursed) entire playthrough).

Actually, the only alternative to the above that I can see working
to make replayed levels sufficiently interesting in a longish game
is if the game naturally generates such a sheer variety of *tactical*
situations that some very interesting and novel ones are bound to
occur in every game. That's probably possible but probably a *lot*
harder to design in intentionally, especially if you want it to
emerge from a small, simple, and orthogonal set of game mechanic
rules that would be friendly for humans to learn quickly, rather
than a giant, special-case-ridden, difficult-to-memorize, principle-
of-least-astonishment-violating, baroque mess of rules amounting to
the gameplay equivalent of spaghetti code, also known as Nethack. ;)

> Even without permanent death we still should aim to make the game
> winnable without dying once and such victory should be is rewarded
> somehow. Bigger score certainly but I think about unlocking some
> content, can be some harder side-quests or challenging races/classes
> or just cosmetics like special awards/titles or things like that.

"New Game +" and "achievements" are the most obvious things to spring to
mind. More meaningful if the game has a social-networking component (so, is
hosted on a server, with everyone's achievements and high scores public,
and of course, because it's on a server, short of black-hat hacking or
being the sysadmin you can't cheat).

> Depends. I imagine exiting rogue-like as such that there should be majority
> of encounters where it is possible to decide if to avoid or if to escape or if to
> fight a battle. The tactical situation should be also possible to engineer, to
> improve the odds and to optimize costs.

The ones I've seen seem to divide starkly into two categories in that
regard. Some don't make it easy to pick your battles and terrain until the
late game (Nethack, Adom, Dungeon Crawl) and some may arguably make that
*too* easy after level 1 or 2 or so (pretty much all the Moria/Angband
descendants).

Controlling the circumstances (and fact) of battle should probably be hard
if:

* The game is short enough that restarting after bad luck is fine.
* The game is varied enough that same, and has no plot, just get as
deep/as rich/etc. as you can
* The game lacks permadeath, so you won't *be* restarting entirely
* The specific battle in question should be survivable, if not
winnable, absent *extraordinary* bad luck or pretty shoddy
preparedness. So, unless the character is very under-leveled or
underequipped or lacks several fairly commonplace consumables
entirely it should be survivable, perhaps at the cost of some
hit points and dignity.

The fourth option can be made generic if:

* Most monsters can't do a large fraction of a normally-leveled-
and-equipped encountering player's max HP in one shot.
* The few that can are mostly easily spotted and evaded or fought
under player-favorable conditions, as the player sees fit.
* The few *remaining* are unique "boss" monsters or otherwise rare,
endgame-deep, or so shallow that starting over isn't much of a
bother when you're still not over-leveled enough to curbstomp
them without breaking a sweat. (To sum these three up: MMAS,
Most Monsters Are Survivable.)
* Curative and status-restorative items are readily available
enough that it takes bad luck to run out of them and die by
attrition across multiple encounters, unless you're out of
your depth. (To sum up: ETH, Easy To Heal.)
* Retreating after finding yourself out of your depth is generally
possible and not too difficult, and you can get better equipped
or leveled before retrying that area. (To sum up: ETLG, Easy To
Level Grind.)

Oddly, I find that the roguelikes that violate MMAA (Most Monsters Are
Avoidable) violate ETH and/or ETLG as well, resulting in a lot of deaths by
attrition (Crawl, especially), and the ones that don't, derive all their
difficulty from violating MMAS (the Angband variants: MMAA, ETH, and ETLG,
but nearly everything later in the game can kill you in one or two hits if
you lack proper defenses, can kill you in a few hits if you don't kill it
first or heal occasionally in battle, and a few just shouldn't be tangled
with at all -- you should never fight a deep Angband monster without
getting it alone, having full HP at the start, and resisting everything
resistable that it can throw at you, including having 100% protection from
every crippling status ailment).

MMAA makes for a highly tactical game (which the Angband family tree is
particularly known for). The alternative combination of MMAS, ETH, and ETLG
deserves a superior roguelike than the ones I've seen thus far, IMO.

MMAS: That's not the one that's a problem, for the most part. Just have a
normally-leveled-and-equipped player need 10-12 hits or very bad luck with
monster criticals or status ailments to be killed by them, or to have been
mobbed without the chance to retreat into a bottleneck. Make cleared areas
of a level stay cleared or become at worst sparsely populated by weak
enemies, so the line of retreat generally won't be blocked, either to a
bottleneck or right off the level and back upstairs as the case may be.
Make being slaughtered with ranged damage or by faster opponents while
trying to flee on foot also rare or otherwise generally avoidable -- e.g.
most in-depth monsters aren't faster than the player, ranged attacks mostly
have damage and/or hit-rate falloff with distance and ranged attackers have
to stop moving to shoot so a player with a shield is soon five or six
squares from a ranged attacker and taking minimal damage, ranged attacks
that are deadly throughout their theoretical range (lightning bolt, say)
are short-range, can't be cast frequently, or etc. (give monsters limited
MP for example), and of course twisty turny passages are common to once the
player gets a sufficient lead from the caster stopping to cast spells now
and again the terrain obstructs the attacks anyway.

ETH: Regen and low monster replacement rate suffice, but with tedium. The
up stairs might be in a room with a healing spring and no monsters, so the
player can never arrive on a new level into an unwinnable situation, and
can retreat to heal (or flee the level) at nearly any time unless
outflanked and cut off, but is encouraged to explore and fight on. Deaths
should result from a player mistake (fighting past when one should have
gone back and healed) or ignorance (next time, have fire resistance before
going into the Ruby Mines -- there be dragons). Alternatively, healing
potions are commonplace.

ETLG: Persistent levels are the usual way that non-Angband-family
roguelikes violate this, resulting in little level-grinding and next-to-no
equipment-grinding possible if you reach a level underleveled or
underequipped for what you encounter there. Worse, permalevels means that
what you encounter there, if out of depth, can't be reset by leaving and
reentering, either, so you can be screwed easily by the RNG. (Dungeon Crawl
is particularly notorious for this,
but even somewhat violates MMAS even without taking OOD baddies plus
permalevels into consideration.)

Grinding deserves a few paragraphs all by itself. If the game is at all
long AND has a victory-condition rather than being a "get as far as you
can" game, it *should* generally be possible so that you can meet a
challenge. Permalevels, while inhibiting grinding by simply regenerating
earlier levels, actually almost *demand* grinding be possible anyway,
because you can otherwise be blocked by a permalevel generated with an OOD
baddie that's beyond your ability to kill or sidestep if you don't become
over-leveled or over-equipped.

* Nonpermalevels without ironman makes grinding automatically easy:
revisit older levels and repeatedly fight and loot on them. This can
produce tedium, but if the game is well-balanced and either MMAA or
MMAS it also makes grinding unnecessary for the most part, as you can
just regenerate the *current* level if it has monsters on it you
can't handle without overleveling. The problem then becomes the
temptation to grind anyway and cheese by being overleveled for the
whole game. Angband is prone to this, and it's probably why there are
actually successful AI players of Angband. Notably, the AI players are
notorious for grinding so much that a human player would be taking
*years* at a single game, playing 24/7, if doing the same thing.
Removing grinding temptation can be done by balancing the game,
keeping the ability to regenerate the current level, and forcing
a partial ironman: you can only be at either the deepest level you've
been or the second deepest; go up and you get a level with no up
stairs and go down and you get a level with up and down stairs. Now
you can regenerate the current level or even grind a little, but not
grind at a much easier level than you've been to. This can be
augmented by a growing chance to not generate up stairs on successive
visits to the same deepest-level-you've-been-so-far. Another option
is downstairs and "sidestairs" only, with the latter giving a new
level at the current depth. You can't get shallower (difficultywise)
but each level is in some sense infinite so you can avoid being
blocked by OOD baddies.

Another antigrinding method could be to limit total monster and
equipment generation at a given depth. Revisiting a given level
enough times results in emptier and emptier levels after a while. The
game should be balanced for visiting each level once but exploring it
fairly thoroughly, allowing a bit of grinding, but not rewarding too
much grinding. A variant: each level has a depletable pool of "monster
points" and "item points". Generating a level produces monsters and
items based on the values for that level, *without* reducing those
values. *Killing* a (non-summoned-or-etc.) monster on that level
depletes it one monster point; *identifying* a floor-generated
item on that level depletes it one item point. If you're forced to
flee a nearly-unexplored level (incl. via "sidestairs") by OOD
monster, breeder explosion, etc. you haven't depleted these much and
can still get your game-balance's worth out of that depth, before
descending farther. If you try to grind on that level, though, you
quickly run out of new monsters and items and your XP and equips
plateau and you must descend to get more.

* Permalevels.

Permalevels present their own challenge in this area. The main issues
are:

ULE -- under leveled or equipped, due to poor RNG luck with items and
monsters in earlier levels.

OOD -- out-of-depth baddie makes the next level impassable without
grinding.


Possible solution 1: Infinite, tree-like dungeon, with one upstair and
at least two downstairs on most or all levels.

Grinding is possible by backtracking and exploring a new branch.
Grinding is discouraged past a certain point as unexplored parts of
the tree that aren't deeper than you've been get to requiring more
and more travel through empty, cleared level to reach.
ULE can thus be avoided, but overleveling is somewhat discouraged,
and can be prevented past a certain point by additionally using the
"monster points" and "item points" from above.
OOD can be avoided by simply routing around the obstacle. Go back up
and take the *other* downstair on that level. If blocked there, too,
go up *twice* and take the unexplored downstair on *that* level, and
so forth.

An implementation downside is that the player can force the savefile
to get very large without going very deep. On a game server this
might even qualify as a denial of service security vulnerability,
allowing a black hat to register an account, start a game, and
exhaust the server's disk space exploring every path systematically
in a breadth-first search for all down stairs, possibly abetted by a
bot client. Server-hosted games will probably want finite and/or
non-persistent dungeons, and a limited number of simultaneous
savefiles per user or else the need to solve a captcha to create a
new savefile.


Possible solution 2: Linear dungeon plus finite, short side-branches.

Side-branches don't let you progress, so OOD on the main line is a
threat. But side branches let you level-grind and equipment-grind
until you can deal with the beast. Side branches can be made themed,
so that they're more interesting than just more-of-the-same in case
grinding proves necessary. Side-branches can even be side-quests with
some guaranteed useful item if you beat a mini-boss at the end. So,
for example, a few levels before dragons are likely, there could be
four elemental-themed side branches, in each of which you can get by
without that element's protections but at the end of which you can
acquire protective gear against that element. If you hit an ice
dragon early enough to block you along the main-line, backtrack to
the ice side-branch and get the guaranteed ice shield, then go back
and thwack the dragon.

Side-branches can be indicated by abnormally-colored down stairs
(e.g. blue for the afore-mentioned ice side-branch vs. grey for
the mainline). Their placement within a certain range of levels
can be guaranteed, while remaining random as to exactly where.

Many games mentioned earlier have such side-branches, while still
having ETLG/ETH problems, usually from long mainline stretches without
side-branches combined with too high a likelihood of an OOD
obstruction on every alternative path.

Side branches should be easier than the main dungeon (low or no OOD
generation at least) or else short-but-abundant, and some should
be available within the first few dungeon levels. The deeper the
player gets before being obstructed, the more side branches will be
accessible, so side branches can get shorter and less abundant with
depth, and even be absent past a certain point, giving the dungeon
the geometry of an inverted Christmas tree.


Possible solution 3: Highly multiply-connected dungeon with finite
levels.

In this case, each depth has finite content but it's fairly large,
and there are abundant cross-connections and many down stairs and up
stairs. This makes OOD nearly impossible as one can almost always
circumvent an obstacle, still get many kills and items at the depth,
and get to a down stair. Unhandlable monsters near every down stair
or near every up stair would be needed to cause the OOD blockage
problem. Game should be balanced for exploring part of each level
before descending, so you'll be mildly overleveled and overequipped
on average, but not strongly so on average, if you clear each level
completely before descending. Player is encouraged by repetitiveness
to descend earlier, but has limited ability to overlevel or play it
slow, without it reaching Angband's ridiculous extremes in that
department.


Possible solution 4: Ironman-in-place.

Despite the terrain persisting and floor-generated items depleting
permanently, each level has ongoing monster generation that's
substantial. The monsters can drop fresh equipment. However, over
time the monsters generated at level N start to get deeper. You can
grind by using monster generation before re-trying a level with OOD
obstruction or to correct ULE, but the more monsters you kill on a
level, the harder the new ones get, eventually forcing you to descend.
Either new monsters are generated just by hanging around on the level,
or it's repopulated every time you descend and ascend. Monster
difficulty increases the more you *kill* on that level. If it's
repopulating, wiping them out and repopulating a couple of times
should produce the next repopulation generated as if about one more
level deeper. Of course, OOD could result, forcing you up yet another
level or trapping you, if you're not careful. If you leave a
partly-depopulated level and return, new monsters only appear in
sparsely-populated areas and only enough to raise the level's
population to "normal".

I don't favor this solution as it could too easily turn against the
player that's already been hurt by ULE.


Possible solution 5: Scrolls of alter reality.

When used, regenerates the level you're on: new terrain, new
monsters, etc. (up stair and down stair placement may be fixed, if
related to placement on the adjacent, persisting levels).

OOD is fixed by using a scroll to remove it, if you have one.
ULE is fixed by using a scroll to refresh the current level.
Anti-grinding is done by limiting availability of the scrolls.
Too much such limiting will cause unfixable OOD or ULE, so the player
should start with one or two and find them occasionally, perhaps every
couple of generated levels on average. Having the scrolls be less
common than once per generated level prevents using the scrolls on an
easy level to acquire infinitely many more of the scrolls so you can
then grind unlimitedly on deeper levels.


Possible solution 6: Wands of banish monster.

This is a solution to OOD only: this wand can simply delete a monster
from existence. You don't get its experience or drop. If unlimitedly
available, results in MMAA, but if in finite supply, does not while
still proving you a few get out of jail free cards.

ULE becomes less burdensome in that you can compensate by blowing
banish monster charges on monsters rendered too troublesome by your
case of ULE, while continuing to explore and kill less troublesome
opponents until the law of large numbers steps in and corrects ULE.

As with alter reality, the player should start with a wand with a few
charges, and more should be infrequently found. It shouldn't be
carried, dropped, or used by monsters, rechargeable, or able to
remove any plot-critical monster (and it goes without saying that it
should be obvious what monsters will resist it). It must work,
however, on all other monsters. Making resisting it a common
difficulty-enhancement on strong monsters entirely defeats its
purpose.


Possible solution 7: Scrolls of acquirement.

This partly fixes ULE and may help OOD if you get a great item to
boost you vs. that opponent. Start the player with a few and make
them otherwise rare; items are generated based on the player's depth
and possibly even situation (e.g. if it rolls the elemental-gear
category, it may then favor items that plug resistance holes to items
that give redundant resistances; or it may roll ten items and keep
the one that gives the player the biggest improvement, weighted in
some manner among gained resists, attack damage, armor, evasion,
stats, etc. if equipped, for rings maximized over both slots it could
be equipped in replacing existing rings).

Underequipped: Should be directly fixed.
Underleveled: May be compensated for by being overequipped.
OOD obstruction: May be compensated for by being overequipped.

Anti-abuse: so long as they aren't common, the player is encouraged to
reserve them for difficult situations, as they produce more powerful
stuff (and may save you from repeating a longer preceding part of the
game) the deeper you get before you use them.


Note that all of the above items, scrolls of acquirement and reality
altering and wands of banish monster, should be insusceptible to
item-destruction attacks that hit the player. Uncollected ones on the
dungeon floor may be susceptible to being destroyed there or not, but
ones the player has collected should be reliable short of the player
consuming or dropping them. Otherwise, one can be ULE-screwed by being
underequipped *with these items*.

They could also be combined -- particularly, the latter two could
each cover some situations where the other was less helpful, and
acquirement could even have a chance to produce banish monster if the
game detected that you were decently equipped but a fairly strong OOD
opponent existed on a mostly unexplored level near the up stairs on
the main line and you were near the corresponding down stair one
level up. (If acquirement detects for that, though, it can also
obviously try to give you equipment that is tuned to make you more
powerful, either in offense or defense, against that specific
opponent, too.)


Note that taking ANY of the above to too much of an extreme makes a game
lose its difficulty. However, the difficulty should be of the type that is
overcome with skill, preferable to with grinding or with luck. Adding
limited grinding-enabling mechanisms, as most of the above do, allows
converting luck into grinding, in effect -- more chances for good items or
an easy OOD kill worth mucho XP, say, and circumvention of various forms of
bad luck. In a game with randomness, some way to convert luck into mild
grinding seems needed.

Returning to the sources of difficulty, there are six ways to get
difficulty:

* Monsters that violate both MMAA and MMAS. The problem is this makes
the game very luck-dependent if it's generic. If the game is short or
lacks a plot this can be acceptable. Otherwise, should be restricted
to bosses found at predictable places, whence MMAA isn't really
violated as you can just not go there yet.

YASDs: would come from tackling a boss you were unprepared for, out
of ignorance or desperation, or from fighting "wrong" out of
ignorance. Mostly genuine YASDs -- from avoidable player
ignorance/error rather than screwed-by-the-RNG.

* Violate ETH. Violating it too strongly makes the game too hard, but
having the player have to somewhat ration consumables is wise. It
should be hard to get completely screwed without gross mismanagement,
but easy with same, and tactics and strategy should have to change if
consumables are low, but preferably not to "grind an early level for
hours for lots of cure potions, then come back".

Ideally, it should be easy to run low on one of several alternative
ways of coping with a particular thing but hard to run low on all of
them at once, forcing tactical changes when one runs low. For example,
having an unresistable teleport-monster-to-far-side-of-level wand and
cure potions, so you can deal with a monster that's beating you down
or that shows up when you're at low HP by healing or by sending it
harmlessly far away, and if you're low on one you can consume the
other as needed to deal with such situations. In that case the item
production should be balanced so your acquisition of both combined
should tend to a bit more than suffice for what you'll need but your
acquisition of just one should tend to be somewhat less than enough.
A player that sticks to one for a while *will* run out, statistically
speaking, but should then have plenty of the other.

The game could also provide advancement mechanisms unrelated to
combat, if sufficiently interesting in their own right. Then
"avoid combat for a while" becomes viable as another alternative when
low on curatives, given MMAA or monster-free areas where the other
advancement mechanisms can be employed.

Combat could also have an "easy" flavor and a "hard" flavor, the
latter requiring more player skill but allowing avoiding more damage.
A realtime combat against a single opponent involving reflexes instead
of the usual turn-based could be the "hard" flavor, with a skilled
player able to dodge a lot of attacks.

YASDs: mainly from squandering a resource one should have rationed.
If that resource could have been stretched, then avoidable.

* Violate ETLG. I proposed many mechanisms for allowing, but limiting,
grinding for levels (and equipment) above. Most of these have the
degree of limitation tunable, so the amount permitted can be made
mildly constricting, so mild ULE or OOD problems become possible
while still making severe ones avoidable. Generally, where it would
be mild without grinding it goes away with the limited grinding
permitted, and where it would be severe without grinding it becomes
mild. Mild to the point the player can survive and progress, with
skill or a little luck or at the cost of some consumables.

YASDs: may be partly bad luck in this instance. Non-grinding
compensatory mechanisms are suggested both above and below, which
make these YASDs result from squandering a resource (scrolls of alter
reality, a nonreplenishable MP pool, or etc.) instead, so from player
ignorance or mismanagement of resources.

* A ticking clock, combined with balance such that difficulty is more
or less linear in grinding. Bad luck with OOD obstructions or being
ULE can be commuted into losing time off the clock via grinding.
The clock can either kill you if it runs out, or be a slow but steady
increase in difficulty from some source, such as harder monsters, or
a constant low-grade XP drain so you'll become *underleveled* by
grinding at too-shallow levels, or something.

YASDs: squandering a finite resource (the clock time).

* A nonreplenishable consumable resource. If that's food, see "a ticking
clock", above. Otherwise, the supply can be stretched with rationing,
trading off against getting less of whatever benefit using the
consumable brings. If HP restoration is large but finite in a game,
for example, then the fruits of combat (XP, items) are traded off. If
MP restoration, spellcasting is bounded. Spells can then be made
powerful even fairly early, balanced by the player wanting to reserve
those nonreplenishable MPs for serious opponents and to have as many
left as possible for the end boss (if existent). But a single
20-MP Greater Fireball spell can wipe out that weak-to-fire OOD Swamp
Beast blocking your progress to level 7 in one shot, from full
health, and you learn it at level 3.

Finite MP restoration strikes me as potentially particularly
interesting, especially if the game is given a final boss that is
tough but beatable without magic and can be cheesed with enough magic.
Early use of magic will then be a good idea, up to a point, past which
it costs you by making the boss fight painful, but using it up won't
generally make the game unwinnable unless you splurge it all on the
first few levels' weak opponents and then run into that Swamp Beast.
With cautious use against only tough opponents, it shouldn't tend to
run out before the player is near endgame-powerful and OOD monsters
aren't going to be showstoppingly powerful anymore even faced with
0MP. (This can be assured if, say, by the time even a player unlucky
with OOD monsters has run out of MP he's strong enough to take out
the end boss without MP with a long and hard slog and careful play.)

Variation: resource is finite per level and you can't take a surplus
with you to the next. I heard of some RL with no HP regen or cure
items but full heal taking downstairs. Finite MP on each level, from
a depletable ambient magical energy or whatnot, is another
possibility. Another is a bag of spilling backpack: your worn
equipment is safe, but every use of stairs empties your inventory.
Consumable management becomes per-level instead of whole-game, which
may be good if the game is balanced for it. Combined with permalevels
this pretty much kills grinding, too.

YASDs: squandering a finite resource.

* Increase variability.

+ Increase probability of OOD monster encounters, while keeping at
least one of MMAA or MMAS with respect to those monsters. Avoidance
might consume a finite resource (e.g. Wand of Banish Monster
charges you can't get more of). Survival might consume a finite
resource (e.g. curatives, finite-MP, grindability of finitely-
grindable earlier levels).

+ The suggestions very early in this post for making the whole game a
bit more random as to what items and monsters tend to be available
when. These could provide some tactical challenges, while still
being balanceable for not screwing the player completely. Dungeons
with limited

> All decisions should have some downside or cost to make it harder and more
> meaningful to decide. If player uses up all of his resources because of sub-
> optimal decisions then he deserves to be in difficult situation.

See above. :)

>> I know this is an idealisation. It is impossible to have a good game
>> where using up all your resources will not bring pain. But I think
>> roguelikes should in general operate on the principle that if there is
>> life, there is hope.
>
> It should certainly bring danger, the odds should be against him. It would
> be nicer if his odds to survive were always noticeably bigger than total
> 0, but in random game that might be is not so simple to implement.

Easy. Give some easy-to-get weapon that grants a 1 in 20 chance at a
one-hit KO (granting drop and full XP) against all opponents except final
boss. If you can always land the first blow by tactical movement (rest one
turn if monster separated from you by one space, step toward it if farther,
little or no ranged attacks worthy of fear; or the player's weapon with
this ability is ranged), this gives you a noticeable chance against any
single encounter. Possibilities include:

Knife of Cleaving. 1 in 20 chance to 1HKO non-bosses and weak damage
otherwise.

Wand of Wonder with a 1 in 20 chance to cast a high-damage nonresistable
nonblockable spell on its target, sufficient to kill most early game
monsters in one hit if it comes up. Later on you've got more uncollected
resources at your back in partially-cleared levels and more ways of dealing
with an OOD or otherwise difficult baddie. Unless you've really gathered
and squandered *everything* you could have to that point, including WoW
charges, in which case, yeah, you're screwed.

> I still think that tossing him somewhat farther back from that "brick wall"
> might be better. Totally to tile 0 (permadeath) feels fair and clear, but it
> might be is not optimal if it is too far from that wall. Only a single step
> back may be is insufficient since it may encourage him to reroll the
> dice of that head-hit instead of stepping aside and rethinking and
> preparing better and earlier. Your idea of providing him one retry before
> throwing him farther back serves that purpose too.

Each death has a more severe penalty is an option, yes. Another, odder
notion, though, would be if each death made the player *stronger*, though
not in a manner amenable to grinding. Bashing your head against an obstacle
then eventually results in a head hard enough to bash the obstacle to bits
with. But it needs to not be cheesable. One possibility is that dying to an
OOD monster resurrects you on the previous level with boosted stats (and
the option to collect loot from that level all over again) and no loss of
stuff, or even you gain one XP level (die and learn!). Die to non-OOD
monster gives no bonus and you lose some stuff. So, bad luck gives you the
means to overcome it, but poor play isn't rewarded. Of course, abusable if
the player can deliberately summon up an OOD monster to die to whenever
they feel they could use an extra level or stat boost.

A mixed-blessing is another option: all deaths boot you back some levels
and lose you all progress made since then (as per the very beginning of my
post), but you can re-get (and better manage) the resources past that
point, and deaths to OOD monsters result in a levelup or stat boost of some
sort.

[Eric Colossal]

On Sunday, December 2, 2012 6:21:31 PM UTC-5, Eric Colossal wrote:
> Hello there!
>
> <SNIP!>

>
> I appreciate any help you can give!
>
> Eric!

Hello there again!

Thank you so much for all your replies! And I apologize to Zamdam, I haven't finished reading all of your post yet!

In most RLs the only thing you carry with you after death is player knowledge of the world and skill at navigating it. I don't generally starve to death anymore in games but that's not because I banked a bunch of food between lives and can feast whenever I want. It's because I banked a bunch of personal skill at playing the game. I did this by grinding. I would play, learn, die over and over again until the 'play, learn' portion of the game got longer before the 'die' end cap.

A large amount of the work that permadeath does is raise the stakes of every fight, every decision, every quaff and sometimes every step if traps are a danger. This is a genius aspect of RLs. So many other games I play do not ask hard decisions of me, let alone ones with dire consequences.

What I'm trying to discover by asking this question and brainstorm game ideas is if there is anything in game that can be transferred after death. As an example: Comparing Canabalt to Jetpack Joyride. ~Similar~ games but Jetpack Joyride offers money as a constant after death. With this money you can buy things that change the game. Not necessarily make it easier, just change it. Offer new decisions and new interactions.

For instance: what if I could keep all gold I found in a dungeon in an RL and use it to purchase something for my next run or unlock passive upgrades for all future runs? What if it cost X amount of gold to start the game with a purchased item from a pregame store? What if it cost X amount of gold to enable a passive upgrade that makes hunger deplete slower? What if it cost XXX amount of gold to bring my last corpse back to life at the exact point of death?

I'm going to grind in a RL no matter what. Unless I install it and defeat it in one life I'm going to grind. I'm going to die and learn the world, I'm going to experiment and succeed and fail in my experiments, I'm going to play and play and play. I'm thinking there is definitely something to the idea of while I'm grinding I also have a larger secondary goal or a larger secondary desire being filled. I am gathering a resource that is more than personal player experience, something in-game and spendable to further my game experience.

I shy away from this next suggestion because the discoverability of RL games is so precious to me but what if there was a list of things that once completing one of them I was rewarded by the game. Get to depth 15 and I've enabled the ability to start with better armor. Kill 4 monsters with a single thrown potion of fire and enable something else. Fall to my death X times. Polymorph something X amount of times. [Recently I was playing FTL and getting halfway through the map you are awarded a new ship design for your next play through. It didn't make the game easier, it made it different]

Anyway, I want to thank you all again for the interesting discussion!
Eric

[oot...@hot.ee]

On Monday, 7 January 2013 04:50:49 UTC+2, Eric Colossal wrote:
> A large amount of the work that permadeath does is raise the stakes of
> every fight, every decision, every quaff and sometimes every step if
> traps are a danger. This is a genius aspect of RLs.

If to lower the stakes then there should be still some "permadeath"
mode where it is like that. The mode itself can be made the "item" that
you request, "reward content" for good performance in previous games.

> I shy away from this next suggestion because the discoverability of RL
> games is so precious to me but what if there was a list of things that
> once completing one of them I was rewarded by the game.

Reward content can be randomized and announcement of it can be delayed.
Player will be told that he has reached requirements of "achievement" and
that it "unlocked" something. What it is (some unique monster, branch,
spellbook or sidequest) will be told only when he meets it in game.

> Get to depth 15 and I've enabled the ability to start with better armor.

Advanced players often complain that early game is too easy. Better armor
does not make it harder.

They may like a gift to start from depth 5 with average character
that they have had down there. Lets say reach depth 15 with 3 different
gnoll hunters and it opens "fast start" from depth 5 with gnoll hunters.

> Kill 4 monsters with a single thrown potion of fire and enable something else.
> Fall to my death X times.
> Polymorph something X amount of times.
> [Recently I was playing FTL and getting halfway through the map you are
> awarded a new ship design for your next play through. It didn't make the
> game easier, it made it different]

The key seems to be to open up new content that is actually adding
complexity to choices (new race, new class, new deity). What solves some
problems should also add some other concerns.

[The Magnificent Zamdan]

On Mon, 7 Jan 2013 08:52:18 -0800 (PST), oot...@hot.ee wrote:

> On Monday, 7 January 2013 04:50:49 UTC+2, Eric Colossal wrote:
>> Get to depth 15 and I've enabled the ability to start with better armor.

Interesting idea.

> Advanced players often complain that early game is too easy. Better armor
> does not make it harder.

It could make it *faster*, though, to get to where the advanced player
finds things more interesting.

On the other hand, see my other post for some ideas on how the early game
could be kept fresh and interesting, for quite a while, even for advanced
players.

> The key seems to be to open up new content that is actually adding
> complexity to choices (new race, new class, new deity). What solves some
> problems should also add some other concerns.

That's an option as well, I suppose.

[oot...@hot.ee]

On Tuesday, 8 January 2013 23:54:24 UTC+2, The Magnificent Zamdan wrote:
> On the other hand, see my other post for some ideas on how the early game
> could be kept fresh and interesting, for quite a while, even for advanced
> players.

Often the firepower and damage resistance of character do progress by orders
of magnitude during few hours. The situations that are too easy for it are
not too far (in time or levels) from the situations that are impossible for
it. That makes it perhaps tricky to balance the game more variable but keep
it still not too hard nor easy.

Removal of that steep progress may diminish one source of fun. Players
enjoy effortlessly bulldozing a whole group of mobs that were feared bullies
in not so distant past.

[Radomir Dopieralski]

On 2013-01-07, Eric Colossal <mrcol...@gmail.com> wrote:
> In most RLs the only thing you carry with you after death is player
> knowledge of the world and skill at navigating it. I don't generally
> starve to death anymore in games but that's not because I banked a bunch
> of food between lives and can feast whenever I want. It's because
> I banked a bunch of personal skill at playing the game. I did this by
> grinding. I would play, learn, die over and over again until the 'play,
> learn' portion of the game got longer before the 'die' end cap.

I think that your meaning for the word "grind" is fundamentally different
from what I expected. Grinding, th way that I meant it, is repeating some
boring, relatively safe and at least a little profitable routine in order
to make the game "easier" -- that is, to remove or marginalise some of the
gameplay mechanics that the player would otherwise have to learn to
overcome. Consequently, as long as you are in danger, as long as you have
to think and, most importantly, as long as you are learning -- it's not
grinding. The way you use the word "grind", it seems to be synonymous with
"play the game".

[...]

> For instance: what if I could keep all gold I found in a dungeon in an
> RL and use it to purchase something for my next run or unlock passive
> upgrades for all future runs? What if it cost X amount of gold to start
> the game with a purchased item from a pregame store? What if it cost
> X amount of gold to enable a passive upgrade that makes hunger deplete
> slower? What if it cost XXX amount of gold to bring my last corpse back
> to life at the exact point of death?

This is an interesting idea. Except you don't want the hunger to be
slower. You don't want to simplify and dumb down the game as the player
becomes more advanced -- that would only lead to boredom. Instead, you
want to give them items and powers that give them power, but complicate
the early levels considerably. You want to give them an option of enabling
additional game mechanics.

How to do that?

There are two simple techniques, they have been mentioned here before, but
I will summarise them now quickly.

The first way is to do what you should be doing with powerful weapons and
other equipment anyways -- make them gradually more complex. Doom (the
first-person-perspective shooter game) is a nice example of that with its
weapons. You start with a pistol, which is a relatively accurate and fast
weapon, with lots of ammunition available and no side effects. It's the
second simplest weapon in the game, the only simpler one being fists,
which don't even need ammo. Once you clear the first few rooms, you find
a second weapon -- the shotgun. The shotgun is considerably more powerful
-- enough to make you want to use it from now on. But it is also
considerably more complicated. The reload time is slow, the accuracy is
low and the ammunition is harder to get. You have to use it more
carefully, and you have to switch back to the pistol sometimes, when the
shotgun is simply unsuitable for the task. At some point you find the
rocket launcher, which is even more complex than the shotgun. Not only it
is even slower, but it also pushes you back. Not only it is even less
accurate, but you can hurt yourself if the rocket explodes too close to
yourself. And the ammunition is even harder to get. But you don't really
have much choice, because the monsters have become strong enough to
effectively force you to use the strongest weapons you have -- with all
their drawbacks.

So, back to our roguelike, if you let me start the game with a rocket
launcher (or wand of fireballs), then sure, the monsters will be
relatively weaker (at least until I run out of the rockets), but I will
also need to consider all the extra things that I need to consider
whenever I use the rocket launcher, so the early game becomes less risky,
but mor complicated at the same time.

The second way you can do it are optional challenges. For example, you can
have inaccessible treasure vaults generated on the early levels. Then let
the players start the game with equipment that lets them access the vaults
-- wand of digging, scroll of teleport, lockpick, etc. But the vaults are
guarded, and you really need to use all your wits and early game skills to
grow powerful enough so early. On the other hand, the vaults would be
mostly gold, or some other "score" items -- not really affecting the later
game.

[Angband spoiler alert]
One extreme example of an optional challenge on the early levels is the
Greater Hell Beast monster in Angband. Sure, it's a great joke on the
player (and I will periodically forget about it and get caught on it
again), but it is also a challenge for the players near the end of the
game. And the prize doesn't really affect the game, while it is still
a fun thing.
[/spoiler]

[...]

--
Radomir Dopieralski, sheep.art.pl

[Radomir Dopieralski]

On 2013-01-06, The Magnificent Zamdan <t...@gmai1.invalid> wrote:
> I'd say that the condition for permadeath to be appropriate is that either
>
> a) The whole game is fairly short, if the player doesn't do a lot of
> grinding, and there isn't much mileage in doing a lot of grinding
> early; or

Actually, a well-written game will have little grinding even if it is
quite long.

> b) The game has so *much* variability that the early game is fresh
> and interesting even 100 times in rapid succession. That will
> mean a lot more than just random mazes and a chance of out-of-depth
> items and monsters, clearly.
>
> One possibility that comes to mind is to categorize items into
> various groups and monsters into various groups, and also chunk them
> farther by levels (e.g. 1-3, 4-8, 9-12...), and in each
> category-level chunk make some of the items (resp. monsters)
> unavailable, rare, ten levels (or more) deeper, not generatable near
> the starting town (if there's geography) so "foreign and exotic", or
> some such, and not available in (starting town) shops (if
> applicable).

My favorite roguelike game, Rogue, does that pretty well, in my opinion.
While the monsters and room types depend on the depth (with an occasional
out-of-depth monster or zoo), the items are generated in a completely
arbitrary way. You have the same chance of finding a plate armor on the
first level, as on the 20th level. This works very well variety-wise, as
you have to adapt your whole playing style to utilize the equipment that
you find the best. This makes every game interesting, *especially* in the
early levels. In the later levels there is a good chance that you have
already found whatever item your strategy is based on, so it's less about
adapting and more about general efficiency. A similar effect can be
observed in most major roguelike games, although there it's more about
the artifacts and out-of-depth items, and not "everyday" loot.

--
Radomir Dopieralski, sheep.art.pl

[Radomir Dopieralski]

On 2013-01-06, The Magnificent Zamdan <t...@gmai1.invalid> wrote:

> If there's any capsule summary, it's that if resources are finite, bad
> luck can cause "screwed by the RNG", but adding one more finite resource
> that's unbalancingly strong in any single encounter but unreplenishable
> gives the player a way out of such situations that they'd otherwise
> prefer to reserve until later in the game than squander. Adding more
> "width" to the dungeon allows grinding to overlevel or overequip, but
> this can be limited or even made a finite resource (e.g. through
> consumable, rare reset-level scrolls the player starts with a few of).

Having read it all I can't help but notice that there is a quaint
underlying assumption here: that it should be possible, or even
preferable, to win the game in one go, in the first attempt. That it is
somehow "unfair" if you have to die to discover a certain mechanic or
strategy, or when you can paint yourself into a corner and end up with
unwinnable situation.

I don't think that this assumption works in roguelikes. In the best case,
balancing your game becomes an order of magnitude harder with that rule in
place. In the worst case, you give up on other constraints in order to
meet that rule. In both cases you pay a cost for enforcing a rule that
makes your game "fair" but doesn't make it "fun".

In particular, one of the dimensions of "fun" is learning. Humans
generally find learning new things pleasant and also get a rush of
pleasure from suddenly understanding things. But if your game can be
finished in a single play-trough, then you are not making your player
learn. They are merely using what they have learned already and luck.
That's bad, because you are removing one useful way of making your game
better.

Another dimension of "fun" is the challenge. It has to be carefully
balanced. Make the game too easy, and the players will be bored and fell
that they are mindlessly grinding. Make it too hard, and they will grow
frustrated and finally abandon it. But in practice, it's not possible
to keep the challenge exactly right -- you will have it too easy or too
hard for some time spans -- and it's fine as long as it's *about* right.
The problem is, that assumption makes it very hard to make the game too
hard -- it will be consistently too easy, and that makes it too easy and
boring on average.

My advice is that you stop caring whether the game is fair or not, but
care whether it is fun to play. It's fine for the players to have their
characters killed by "bad luck" -- they will play more defensively and
leave a bigger margin next time, until they find a play style that is
mostly independent from luck.

--
Radomir Dopieralski, sheep.art.pl

[The Magnificent Zamdan]

On Fri, 11 Jan 2013 14:56:48 +0000 (UTC), Radomir Dopieralski wrote:

> On 2013-01-06, The Magnificent Zamdan <t...@gmai1.invalid> wrote:
>> If there's any capsule summary, it's that if resources are finite, bad
>> luck can cause "screwed by the RNG", but adding one more finite resource
>> that's unbalancingly strong in any single encounter but unreplenishable
>> gives the player a way out of such situations that they'd otherwise
>> prefer to reserve until later in the game than squander. Adding more
>> "width" to the dungeon allows grinding to overlevel or overequip, but
>> this can be limited or even made a finite resource (e.g. through
>> consumable, rare reset-level scrolls the player starts with a few of).
>
> Having read it all I can't help but notice that there is a quaint
> underlying assumption here: that it should be possible, or even
> preferable, to win the game in one go, in the first attempt. That it is
> somehow "unfair" if you have to die to discover a certain mechanic or
> strategy, or when you can paint yourself into a corner and end up with
> unwinnable situation.

Ah. I was not suggesting that -- only to make deaths from sheer bad luck
much less common than deaths from ignorance or pilot error.

Getting stuck from squandering resources counts as the latter; getting
stuck from the game having given you too few resources before obstructing
you with a too-strong obstacle counts as the former.

I thought my post made that clear in the middle section; perhaps not.

> My advice is that you stop caring whether the game is fair or not, but
> care whether it is fun to play. It's fine for the players to have their
> characters killed by "bad luck" -- they will play more defensively and
> leave a bigger margin next time, until they find a play style that is
> mostly independent from luck.

My point being that such a play style should exist and be discoverable.
Having certain resources be finite, but fairly powerful and available
early, seems like it could especially help there, as using the resource is
defensive in the short term, but may leave you a smaller margin later in
the game, and conserving it is defensive in the long term, but increases
your risk at the moment.

[Gerry Quinn]

In article <1fp66kq942o8r$.1mo296y3...@40tude.net>,
t...@gmai1.invalid says...

> WARNING: Long. Set aside some time, or skip. TL;DR: a whole lot of ideas
> and theory for tweaking difficulty, giving the player some sort of finite
> get out of jail free card resource or other free pass against a certain
> amount of bad luck

Some nice ideas there.

I read it as a bag of ideas rather than a systemisation of the
possibilities. Every game works differently, and no design fully
survives contact with players.

- Gerry Quinn

[Gerry Quinn]

In article <slrnkf0a6...@test.moinmo.in>, ne...@sheep.art.pl
says...

> On 2013-01-06, The Magnificent Zamdan <t...@gmai1.invalid> wrote:
> > If there's any capsule summary, it's that if resources are finite, bad
> > luck can cause "screwed by the RNG", but adding one more finite resource
> > that's unbalancingly strong in any single encounter but unreplenishable
> > gives the player a way out of such situations that they'd otherwise
> > prefer to reserve until later in the game than squander. Adding more
> > "width" to the dungeon allows grinding to overlevel or overequip, but
> > this can be limited or even made a finite resource (e.g. through
> > consumable, rare reset-level scrolls the player starts with a few of).
>
> Having read it all I can't help but notice that there is a quaint
> underlying assumption here: that it should be possible, or even
> preferable, to win the game in one go, in the first attempt. That it is
> somehow "unfair" if you have to die to discover a certain mechanic or
> strategy, or when you can paint yourself into a corner and end up with
> unwinnable situation.
>
> I don't think that this assumption works in roguelikes. In the best case,
> balancing your game becomes an order of magnitude harder with that rule in
> place. In the worst case, you give up on other constraints in order to
> meet that rule. In both cases you pay a cost for enforcing a rule that
> makes your game "fair" but doesn't make it "fun".

I would call it an ideal rather than a rule. Certainly it would be hard
for a roguelike to achieve it (certain CRPGs come close IMO), but I do
think in the ideal game you should feel that someone smart enough who
had decent luck might conceivably win the game on the first playthrough.

Perhaps that's part of what's behind my disdain for the ID game. I
don't believe in killing players through lack of information. I want
them to be always making an informed choice (if they care to).

- Gerry Quinn

[Gerry Quinn]

In article <1p96m8qo1yhqv$.4t18l46d...@40tude.net>,
t...@gmai1.invalid says...

> My point being that such a play style should exist and be discoverable.
> Having certain resources be finite, but fairly powerful and available
> early, seems like it could especially help there, as using the resource is
> defensive in the short term, but may leave you a smaller margin later in
> the game, and conserving it is defensive in the long term, but increases
> your risk at the moment.

I don't like this kind of system much; it leads to hoarding. Rogues
should travel light and keep on finding better stuff. They should die
often when they take risks to preserve resources.

I like limited inventories and having to throw useful stuff away: it
encourages the right frame of mind.

- Gerry Quinn

[paul-d...@sbcglobal.net]

Gerry Quinn <ger...@gmail.com> writes:

> ... but I do think in the ideal game you should feel that someone

> smart enough who had decent luck might conceivably win the game on the
> first playthrough.

I don't think so. For one thing, this seems to give a negative
impression; sometimes people do win roguelikes on the first try, but not
win so easily in later games, and as I recall discussions indicate it
makes people think a game is too luck-based, whether or not that's
actually the case.

Secondly, I guess the argument could be made that an ideal game should
teach the player so well that they are prepared for every situation they
encounter (or the most likely situations), but doesn't that assume that
spending a minimum amount of time in game is ideal? I would say that an
ideal game is one that players want to play extensively, meaning the
time taken to learn to win is at least irrelevant, because after winning
a player will just start another game anyway, and possibly implies a
better game when longer, since a first win is a bigger carrot than a
second, or tenth.

Thirdly, and this may be just me, as a player I feel a sense of pressure
to play a game "correctly" when I have a chance of winning. I'd prefer
to keep the pressure off until I've gotten to know the game better.

> Perhaps that's part of what's behind my disdain for the ID game. I
> don't believe in killing players through lack of information. I want
> them to be always making an informed choice (if they care to).

How does the ID game (in theory) prevent players from always making
informed choices? Assuming by "the ID game" you mean ID games in
general, surely it's possible for the ID game not to require uninformed
choices. Having unidentified items doesn't mean you have to use them,
after all!

[Gerry Quinn]

In article <87fw234...@sbcglobal.net>, paul-d...@sbcglobal.net
says...

> Gerry Quinn <ger...@gmail.com> writes:
>
> > ... but I do think in the ideal game you should feel that someone
> > smart enough who had decent luck might conceivably win the game on the
> > first playthrough.
>
> I don't think so. For one thing, this seems to give a negative
> impression; sometimes people do win roguelikes on the first try, but not
> win so easily in later games, and as I recall discussions indicate it
> makes people think a game is too luck-based, whether or not that's
> actually the case.
>
> Secondly, I guess the argument could be made that an ideal game should
> teach the player so well that they are prepared for every situation they
> encounter (or the most likely situations), but doesn't that assume that
> spending a minimum amount of time in game is ideal? I would say that an
> ideal game is one that players want to play extensively, meaning the
> time taken to learn to win is at least irrelevant, because after winning
> a player will just start another game anyway, and possibly implies a
> better game when longer, since a first win is a bigger carrot than a
> second, or tenth.

I think in practice it won't be a problem, because nobody is the perfect
player, and they will make mistakes and fail.

> Thirdly, and this may be just me, as a player I feel a sense of pressure
> to play a game "correctly" when I have a chance of winning. I'd prefer
> to keep the pressure off until I've gotten to know the game better.

I don't think trying silly random things should ever be a useful
strategy.

> > Perhaps that's part of what's behind my disdain for the ID game. I
> > don't believe in killing players through lack of information. I want
> > them to be always making an informed choice (if they care to).
>
> How does the ID game (in theory) prevent players from always making
> informed choices? Assuming by "the ID game" you mean ID games in
> general, surely it's possible for the ID game not to require uninformed
> choices. Having unidentified items doesn't mean you have to use them,
> after all!

ID games in general can only be learned by repeated failure. And
anyway, informed choiced mean you have information. An inventory full
of unidentified objects, without you even knowing the spectrum of what
objects are possible and how to identify them safely, is the antithesis
of informed choice.

I prefer it if the potion the player found is called 'Protection from
Fire' rather than 'Glutinous Plaid' and the dragon is described as
having wisps of smoke coming from its nostrils. It seems a waste to
have to start over because the game author chose not to make those
relevant facts clear.

- Gerry Quinn

[The Magnificent Zamdan]

My suggestion wouldn't apply to all items. Just to the "get out of jail
free card" item or ability -- it would be finite, early gotten (likely
you'd start with all of what you'd ever have), and you'd have to decide in
any serious situation whether to use one or try to get past that situation
without doing so.

Note that it's similar to, but somewhat less powerful and more
choice-requiring than, starting with a finite supply of extra lives,
something that was also discussed in this thread.

[The Magnificent Zamdan]

Death by ignorance takes three forms.

Example one: you die to a slightly out of depth orc warrior at level 2.
Your pack turns out to have contained an unidentified potion of heal wounds
that would have likely saved you, as the orc warrior was down to one block
on its health bar when it nailed you. If you'd known the potion's identity
you'd have used it a round or so earlier, then hit it a couple more times
for victory.

In this instance, the ignorance is not alleviated, since the same color
potion is probably something different next game. You can't "die and learn"
in this case -- at least, not and learn the specific thing whose knowledge
would have saved you.

Example two: you kill a couple of dwarven sorcerors, but the third casts a
blinding spell and kills you as you flail about in the dark.

In this instance, you learn that a dwarven sorceror has a chance to cast a
crippling status ailment on you, so you can't just plow through them or
sooner or later you'll get unlucky. Have a way to cure blindness before
tackling those guys.

Example three: there are monsters called "dreadfuls" that occasionally cast
a spell, "zap", that does moderate damage. One game, a dreadful casts "zap"
and does ten times the usual damage, enough to kill you in one shot. Over
many future games, this happens a grand total of 2 more times; one time you
even have lightning resistance and enough hitpoints to survive the
overly-powerful "zap". Every other time -- dozens -- that "zap" is cast it
does only moderate damage. You chalk it up to a low-probability 1 in 1000
or so chance for massive extra damage and either start avoiding dreadfuls
or just kill them quickly and expect that they will only rarely get you
with an extra-strong "zap". In actual fact, some "clever" developer thought
it would be cute if "zap" did 10x extra damage if the target's level was
equal to the last two digits of your gold, and short of reading the source
code this isn't documented anywhere.

In this instance, the infrequent deaths could be avoided by knowing an
obscure game mechanic, but it's unlikely you'd guess the rules for "zap"'s
extra damage. Eventually it will be in a spoiler file that lots of players
will assiduously avoid, though maybe someone will make a FAQ for just a few
specific niggling things including that one.

I would argue that example two is "preferable" to example one and example
three as the sort of death-preventable-with-more-information a game tends
to generate.

Note that if example 3 is documented, with some info popping up the first
time you see the extra damage come up, it is transmuted into a case like
example 2, since next time you'll know when the extra damage is risked and
when it's not.

Note also that, lest you think example 3 is ludicrously unlikely, I've seen
a game, though it was a CRPG, that had *multiple* gimmicky spells that
cared about levels equaling gold and similar things -- though a) it had
save checkpoints, as CRPGs are wont to do, and b) it did provide
information about MOST of those spells' oddball mechanics when you first
saw them in action.

[The Magnificent Zamdan]

On Mon, 14 Jan 2013 11:37:39 -0600, paul-d...@sbcglobal.net wrote:

> Gerry Quinn <ger...@gmail.com> writes:
>
>> ... but I do think in the ideal game you should feel that someone
>> smart enough who had decent luck might conceivably win the game on the
>> first playthrough.
>
> I don't think so. For one thing, this seems to give a negative
> impression; sometimes people do win roguelikes on the first try, but not
> win so easily in later games, and as I recall discussions indicate it
> makes people think a game is too luck-based, whether or not that's
> actually the case.
>
> Secondly, I guess the argument could be made that an ideal game should
> teach the player so well that they are prepared for every situation they
> encounter (or the most likely situations), but doesn't that assume that
> spending a minimum amount of time in game is ideal? I would say that an
> ideal game is one that players want to play extensively, meaning the
> time taken to learn to win is at least irrelevant, because after winning
> a player will just start another game anyway, and possibly implies a
> better game when longer, since a first win is a bigger carrot than a
> second, or tenth.
>
> Thirdly, and this may be just me, as a player I feel a sense of pressure
> to play a game "correctly" when I have a chance of winning. I'd prefer
> to keep the pressure off until I've gotten to know the game better.

That last one suggests that perhaps the best sort of roguelike should lack
a victory condition entirely, instead being like some of the older arcade
games: the game always ends the same way (you die), and the challenge is to
see how far you can get; the fun is from beating (or at least coming close
to) yours (or someone else's) record level or score, seeing new stuff, and
the like. With a roguelike, of course, there's lots of possibilities for
seeing new stuff and other possible fun, like getting to be overpowered for
a bit one game due to a nice rare find.

>> Perhaps that's part of what's behind my disdain for the ID game. I
>> don't believe in killing players through lack of information. I want
>> them to be always making an informed choice (if they care to).
>
> How does the ID game (in theory) prevent players from always making
> informed choices? Assuming by "the ID game" you mean ID games in
> general, surely it's possible for the ID game not to require uninformed
> choices. Having unidentified items doesn't mean you have to use them,
> after all!

The choice not to use the item is still a choice, and uninformed by any
knowledge of what the item does.

[The Magnificent Zamdan]

On Sun, 13 Jan 2013 16:29:21 -0000, Gerry Quinn wrote:

> In article <1fp66kq942o8r$.1mo296y3...@40tude.net>,
> t...@gmai1.invalid says...
>> WARNING: Long. Set aside some time, or skip. TL;DR: a whole lot of ideas
>> and theory for tweaking difficulty, giving the player some sort of finite
>> get out of jail free card resource or other free pass against a certain
>> amount of bad luck
>
> Some nice ideas there.
>
> I read it as a bag of ideas rather than a systemisation of the
> possibilities.

Yes, that was the idea. I'm sure there's lots more possibilities.

> Every game works differently, and no design fully survives contact with
> players.

The latter is true of software and its users in general, and is why testing
and occasional updates are the norm. :)

[Gerry Quinn]

In article <186t4ik26dh7h$.1sj76nic69c03$.d...@40tude.net>,
t...@gmai1.invalid says...

> That last one suggests that perhaps the best sort of roguelike should lack
> a victory condition entirely, instead being like some of the older arcade
> games: the game always ends the same way (you die), and the challenge is to
> see how far you can get; the fun is from beating (or at least coming close
> to) yours (or someone else's) record level or score, seeing new stuff, and
> the like.

This has been done in some games. Torchlight is notable for it; when
you finish you can enter an infinite dungeon. Torchlight II modifies
this slightly in that the dungeon is infinite but your character level
is capped.

Of course you can play a home-made version of this in many roguelikes.

- Gerry Quinn

[Radomir Dopieralski]

I actually find it a very nice option, but that's an old Moria player
speaking.

--
Radomir Dopieralski, sheep.art.pl

[Radomir Dopieralski]

This is not a very constructive approach. If every game is completely
different and there are no hard or soft rules and guidelines, then we can
as well pack our bags and go home, and just start throwing stuff randomly
into our games and hoping that it somehow improves them.

I work on the assumption that there actually are some rules in there, even
if they are not obvious and a little fuzzy, and I think that it is the
responsibility of a game designer to try and learn those rules, either by
examining the existing games, experimenting with novel approaches or
exchanging ideas and experiences with other designers.

I agree with Noam Chomsky in that the "everything is relative" and "it's
just your opinion" or "it's just a theory" approach is very harmful for
progress in any field.

Sure, every game works differently, but it's possible to analyze it for
specific games and draw generally useful conclusions. No design survives
contact with players, but it's very insightful to see how it fails and
very challenging to find ways to improve it.

--
Radomir Dopieralski, sheep.art.pl

[The Magnificent Zamdan]

On Tue, 15 Jan 2013 13:57:01 +0000 (UTC), Radomir Dopieralski wrote:

> On 2013-01-15, The Magnificent Zamdan <t...@gmai1.invalid> wrote:
>> On Sun, 13 Jan 2013 16:29:21 -0000, Gerry Quinn wrote:
>
>>> Every game works differently, and no design fully survives contact with
>>> players.
>
>> The latter is true of software and its users in general, and is why
>> testing and occasional updates are the norm. :)
>
> This is not a very constructive approach. If every game is completely
> different and there are no hard or soft rules and guidelines, then we can
> as well pack our bags and go home, and just start throwing stuff randomly
> into our games and hoping that it somehow improves them.

???

I never said there were no rules or guidelines. Only that there's always a
need to test and adjust before it "fits well"; theory only takes you so
far, in practice.

> I work on the assumption that there actually are some rules in there, even
> if they are not obvious and a little fuzzy, and I think that it is the
> responsibility of a game designer to try and learn those rules, either by
> examining the existing games, experimenting with novel approaches or
> exchanging ideas and experiences with other designers.
>
> I agree with Noam Chomsky in that the "everything is relative" and "it's
> just your opinion" or "it's just a theory" approach is very harmful for
> progress in any field.

So do I.

> Sure, every game works differently, but it's possible to analyze it for
> specific games and draw generally useful conclusions. No design survives
> contact with players, but it's very insightful to see how it fails and
> very challenging to find ways to improve it.

I don't think we actually disagree.

[paul-d...@sbcglobal.net]

Gerry Quinn <ger...@gmail.com> writes:

> In article <87fw234...@sbcglobal.net>, paul-d...@sbcglobal.net
> says...
>> Gerry Quinn <ger...@gmail.com> writes:
>>
>> > ... but I do think in the ideal game you should feel that someone
>> > smart enough who had decent luck might conceivably win the game on the
>> > first playthrough.
>>
>> I don't think so. For one thing, this seems to give a negative
>> impression; sometimes people do win roguelikes on the first try, but not
>> win so easily in later games, and as I recall discussions indicate it
>> makes people think a game is too luck-based, whether or not that's
>> actually the case.
>>
>> Secondly, I guess the argument could be made that an ideal game should
>> teach the player so well that they are prepared for every situation they
>> encounter (or the most likely situations), but doesn't that assume that
>> spending a minimum amount of time in game is ideal? I would say that an
>> ideal game is one that players want to play extensively, meaning the
>> time taken to learn to win is at least irrelevant, because after winning
>> a player will just start another game anyway, and possibly implies a
>> better game when longer, since a first win is a bigger carrot than a
>> second, or tenth.
>
> I think in practice it won't be a problem, because nobody is the perfect
> player, and they will make mistakes and fail.

Then what's the value in a game being theoretically winnable on the
first attempt? If a real person will take, say, five at minimum, why not
spend four games preparing them to win?

>> Thirdly, and this may be just me, as a player I feel a sense of pressure
>> to play a game "correctly" when I have a chance of winning. I'd prefer
>> to keep the pressure off until I've gotten to know the game better.
>
> I don't think trying silly random things should ever be a useful
> strategy.

I'm not sure what you mean.

>> > Perhaps that's part of what's behind my disdain for the ID game. I
>> > don't believe in killing players through lack of information. I want
>> > them to be always making an informed choice (if they care to).
>>
>> How does the ID game (in theory) prevent players from always making
>> informed choices? Assuming by "the ID game" you mean ID games in
>> general, surely it's possible for the ID game not to require uninformed
>> choices. Having unidentified items doesn't mean you have to use them,
>> after all!
>
> ID games in general can only be learned by repeated failure.

There's no reason an ID game can't have safe techniques to identify
items, and that the game can't teach them as they are needed. It could
provide a wealth of ID information right on the item description screen,
if it chose to, or even step-by-step instructions, since most ID games
are not complex and rely heavily on process of elimination (which I
don't think is ideal, but that's another discussion).

> And anyway, informed choiced mean you have information. An inventory
> full of unidentified objects, without you even knowing the spectrum of
> what objects are possible and how to identify them safely, is the
> antithesis of informed choice.

Why is it assumed you don't know what objects are possible and how to
identify them safely?

> I prefer it if the potion the player found is called 'Protection from
> Fire' rather than 'Glutinous Plaid' and the dragon is described as
> having wisps of smoke coming from its nostrils. It seems a waste to
> have to start over because the game author chose not to make those
> relevant facts clear.

I think there's something to be said for a two-phase process of
discovery, at least for things I can investigate at my leisure. It can
make for a more emotionally involved game, teasing the player with the
chance that the item they hold is the one they want, giving time for the
suspense to build until the big reveal. It makes the process of
collecting items less mechanical. In the early game, it shifts the focus
to core gameplay, since a wise player won't mess with items until they
have the means to identify them (and I know players aren't always helped
to be wise effectively, but that's not inherent). It teaches beneficial
behaviors like patience, and teaching good behaviors is one of the
highest goals a game can aspire to.

> - Gerry Quinn

[paul-d...@sbcglobal.net]

I prefer a game with an end. Of course some people might like a
neverending high-scoring roguelike. Still, I wouldn't call it "best"
(which is relevant since it's me you're replying to :P).

>>> Perhaps that's part of what's behind my disdain for the ID game. I
>>> don't believe in killing players through lack of information. I want
>>> them to be always making an informed choice (if they care to).
>>
>> How does the ID game (in theory) prevent players from always making
>> informed choices? Assuming by "the ID game" you mean ID games in
>> general, surely it's possible for the ID game not to require uninformed
>> choices. Having unidentified items doesn't mean you have to use them,
>> after all!
>
> The choice not to use the item is still a choice, and uninformed by any
> knowledge of what the item does.

But informed by the knowledge that it has a non-negligible chance of
being bad. In any case I meant that it's always possible to be patient
and play the ID game "properly" assuming you know how, so as not to take
risks. Not that one could just abstain from using unidentified items the
whole game. And why couldn't a game teach you how to play the ID game as
well as it teaches you how to play any part of itself?

[Gerry Quinn]

In article <87r4lkj...@sbcglobal.net>, paul-d...@sbcglobal.net

says...
> Gerry Quinn <ger...@gmail.com> writes:

> > I think in practice it won't be a problem, because nobody is the perfect
> > player, and they will make mistakes and fail.
>
> Then what's the value in a game being theoretically winnable on the
> first attempt? If a real person will take, say, five at minimum, why not
> spend four games preparing them to win?

It makes it feel fair and balanced.

To a certain extent I am probably prejudiced by CRPGs, which have tended
to fall into a cycle of necessary reloads instead of being properly
balanced. Maybe it is more annoying in these games. But 'unfair'
deaths in CRPGs that require you to learn meta-game information in order
to win are actually the same issue as 'unfair' deaths in roguelikes.
The only difference is the quasi-solution to the design problem -
restart or reload. I say: fix the design problem and get rid of both.

>
> >> Thirdly, and this may be just me, as a player I feel a sense of pressure
> >> to play a game "correctly" when I have a chance of winning. I'd prefer
> >> to keep the pressure off until I've gotten to know the game better.
> >
> > I don't think trying silly random things should ever be a useful
> > strategy.
>
> I'm not sure what you mean.

In one word, I mean NetHack. In general, I don't think that trying
'crazy' things should be helpful to winning.

> > ID games in general can only be learned by repeated failure.
>
> There's no reason an ID game can't have safe techniques to identify
> items, and that the game can't teach them as they are needed. It could
> provide a wealth of ID information right on the item description screen,
> if it chose to, or even step-by-step instructions, since most ID games
> are not complex and rely heavily on process of elimination (which I
> don't think is ideal, but that's another discussion).

There's no real reason for an ID game at all, IMO. Sure, it's no harm
putting them in the odd game for a change of pace, but I don't think
they should be the default. (In commercial games the tendency has been
to completely trivialise them while still leaving in ID scrolls in case
somebody complains about 'dumbing down'.)

I think in most games a potion of healing should be labelled 'potion of
healing'. Was there a flood in the dungeon or something, and all the
labels washed off? That could actually explain a lot of roguelike
tropes!

> > And anyway, informed choiced mean you have information. An inventory
> > full of unidentified objects, without you even knowing the spectrum of
> > what objects are possible and how to identify them safely, is the
> > antithesis of informed choice.
>
> Why is it assumed you don't know what objects are possible and how to
> identify them safely?

Because to assume otherwise is generally to assume you died repeatedly
to learn it.

>
> > I prefer it if the potion the player found is called 'Protection from
> > Fire' rather than 'Glutinous Plaid' and the dragon is described as
> > having wisps of smoke coming from its nostrils. It seems a waste to
> > have to start over because the game author chose not to make those
> > relevant facts clear.
>
> I think there's something to be said for a two-phase process of
> discovery, at least for things I can investigate at my leisure. It can
> make for a more emotionally involved game, teasing the player with the
> chance that the item they hold is the one they want, giving time for the
> suspense to build until the big reveal. It makes the process of
> collecting items less mechanical. In the early game, it shifts the focus
> to core gameplay, since a wise player won't mess with items until they
> have the means to identify them (and I know players aren't always helped
> to be wise effectively, but that's not inherent). It teaches beneficial
> behaviors like patience, and teaching good behaviors is one of the
> highest goals a game can aspire to.

I'd sooner teach them how to optimise their chances based on known
probabilities. What is the chance of this monster killing me, and if it
is non-trivial what resources should I burn to reduce it?

- Gerry Quinn