Exploitation and Evasion
Cuboidz
exploitation and evasion in roguelikes:
"[...] as the player descends in the dungeon, the level of threat from
monsters increases, exponentially so, without a consequent increase in
the rewards for defeating them, so that in the last few levels it is
better for the player to evade encounters with monsters and conserve
the resources he has accumulated earlier in the game, than attempt to
stand his ground and fight."
The author concedes it isn't always this clear-cut, and defines player
skill as knowing when to exploit, and when to evade.
A few remarks:
(1) ascension kit
From the Nethack wiki:
"The ascension kit is the nickname given to a set of items that are
virtually required for a successful ascension."
In terms of exploitation and evasion, the ascension kit determines
when you can stop exploiting. This implies, that sometimes, you need
to continue exploiting, even if you know you should be evading.
(2) grinding
Like Andrew Doull says, "on the face of it, exploiting is grinding
like behaviour, [...] but what distinguishes exploiting from grinding
is two-fold: limitation of exploitable resources and permadeath."
If I understand correctly, exploitation should never be without risk,
lest it degenerate into grinding. This is a slippery slope, however,
so I wonder what can be done to prevent it.
(3) weakness
Interestingly, the dilemma between exploitation and evasion explains
most of my YASD's. I seem to have an uncontrollable urge to explore
the entire level, even if I know it's dangerous to do so. So
apparently, "evasion" is not in my dictionary. =D
[ARTICLE]
http://www.gamasutra.com/php-bin/news_index.php?story=24468
[/ARTICLE]
Ray
> Like Andrew Doull says, "on the face of it, exploiting is grinding
> like behaviour, [...] but what distinguishes exploiting from grinding
> is two-fold: limitation of exploitable resources and permadeath."
>
> If I understand correctly, exploitation should never be without risk,
> lest it degenerate into grinding. This is a slippery slope, however,
> so I wonder what can be done to prevent it.
This was exactly the same conclusion I reached. More than anything
else, grinding requires near-perfect safety to be worthwhile. If
you take away perfect safety, there is no more grinding.
Bear
Fenrir
> Interestingly, the dilemma between exploitation and evasion explains
> most of my YASD's. I seem to have an uncontrollable urge to explore
> the entire level, even if I know it's dangerous to do so. So
> apparently, "evasion" is not in my dictionary. =D
Same here. I always try to explore and purge every level entirely
before moving on.
Paul Donnelly
It seems to me that most games support this behavior, since it's
necessary to fight a lot to level up and to find useful items.
And I wonder how the existence of an automap affects this. Without a
map, would people still feel the urge to fill it in?
David Ploog
> Fenrir <monstr...@gmail.com> writes:
>
>> Cuboidz wrote:
>>> Interestingly, the dilemma between exploitation and evasion explains
>>> most of my YASD's. I seem to have an uncontrollable urge to explore
>>> the entire level, even if I know it's dangerous to do so. So
>>> apparently, "evasion" is not in my dictionary. =D
>>
>> Same here. I always try to explore and purge every level entirely
>> before moving on.
>
> It seems to me that most games support this behavior, since it's
> necessary to fight a lot to level up and to find useful items.
I think that Crawl strongly tries to not support full exploration but most
people need a while to learn this, and some never do.
> And I wonder how the existence of an automap affects this. Without a
> map, would people still feel the urge to fill it in?
I am not sure if it's the completionist in those players or mere greed. I
am definitely prone to this myself.
David
Fenrir
> It seems to me that most games support this behavior, since it's
> necessary to fight a lot to level up and to find useful items.
>
> And I wonder how the existence of an automap affects this. Without a
> map, would people still feel the urge to fill it in?
I don't think so. I do get a certain satisfaction from filling in the
automap entirely.
It's not really logical to bypass a level if the monsters are giving
you trouble, since the next level is usually going to be even more
difficult. Roguelike levels are usually linear in difficulty.
This gives me an idea: All dungeon levels could be equally perilous,
except the final level that contains the player's goal (boss creature,
precious artifact, etc.), which would be exceptionally difficult. All
levels would be populated with evey type of monster, from the lowliest
rat to the mightiest dragon. Winning would depend upon avoiding the
dangerous creatures while killing only those you could handle, until
you were able to take on the bigger ones.
I'm working on my first roguelike and this doesn't sound difficult at
all to implement, so I think I'll try it.
Derek Ray
> On Mon, 10 Aug 2009, Paul Donnelly wrote:
>> It seems to me that most games support this behavior, since it's
>> necessary to fight a lot to level up and to find useful items.
> I think that Crawl strongly tries to not support full exploration but most
> people need a while to learn this, and some never do.
It will be extremely hard to block full exploration of every map before
proceeding to the next (or "level-hoovering", for short).
>> And I wonder how the existence of an automap affects this. Without a
>> map, would people still feel the urge to fill it in?
> I am not sure if it's the completionist in those players or mere greed. I
> am definitely prone to this myself.
I'm not much of a completionist, nor am I greedy -- but I do know that
in a game where randomness is king, failing to check for all possible
random items can lead to passing up something you'll want later.
The corollary of course is that the more stuff you have, the more likely
you are to have something you'll want. You can always drop something
you have and don't want -- you can't pick up something you never had in
the first place. Level-hoovering is generally _always_ worthwhile
unless your food clock is so brutal as to cause starvation -- not
actually "fun".
--
Derek
Game info and change log: http://sporkhack.com
Beta Server: telnet://sporkhack.com
IRC: irc.freenode.net, #sporkhack
Paul Donnelly
> On Aug 10, 3:49 pm, Paul Donnelly <paul-donne...@sbcglobal.net> wrote:
>> It seems to me that most games support this behavior, since it's
>> necessary to fight a lot to level up and to find useful items.
>>
>> And I wonder how the existence of an automap affects this. Without a
>> map, would people still feel the urge to fill it in?
>
> I don't think so. I do get a certain satisfaction from filling in the
> automap entirely.
Yeah, same here. Or at least I'm frequently confused by an automap with
holes along the edges, so I try to fill them in. I'd probably be happier
without having one at all.
> It's not really logical to bypass a level if the monsters are giving
> you trouble, since the next level is usually going to be even more
> difficult. Roguelike levels are usually linear in difficulty.
>
> This gives me an idea: All dungeon levels could be equally perilous,
> except the final level that contains the player's goal (boss creature,
> precious artifact, etc.), which would be exceptionally difficult. All
> levels would be populated with evey type of monster, from the lowliest
> rat to the mightiest dragon. Winning would depend upon avoiding the
> dangerous creatures while killing only those you could handle, until
> you were able to take on the bigger ones.
>
> I'm working on my first roguelike and this doesn't sound difficult at
> all to implement, so I think I'll try it.
It sounds hard to me, but good luck. I like the idea.
Ray
> On 2009-08-10, David Ploog <pl...@mi.fu-berlin.de> wrote:
>> On Mon, 10 Aug 2009, Paul Donnelly wrote:
>>> And I wonder how the existence of an automap affects this. Without a
>>> map, would people still feel the urge to fill it in?
>> I am not sure if it's the completionist in those players or mere greed. I
>> am definitely prone to this myself.
If each "level" were infinite, ie, the map goes off as far as you
want to explore in every direction, would that prompt people to make
more reasoned judgments about when to proceed?
> I'm not much of a completionist, nor am I greedy -- but I do know that
> in a game where randomness is king, failing to check for all possible
> random items can lead to passing up something you'll want later.
Don't persistent levels solve this though? I mean, if you know
that this dragon (and its treasure) will still be here later, and
that no generated items vanish from the game when you take the stairs,
do you still feel compelled to take it on before you take the stair?
> Level-hoovering is generally _always_ worthwhile
> unless your food clock is so brutal as to cause starvation -- not
> actually "fun".
I disagree. This may be correct in games with nonpersistent
levels, but with persistent levels - where items and monsters
once generated stay in the game unless used up or destroyed -
you lose nothing by checking out the next level.
Bear
Cuboidz
> And I wonder how the existence of an automap affects this. Without a
> map, would people still feel the urge to fill it in?
I don't know whether it's about filling the automap. In the Japanese
roguelike Shiren The Wanderer, for example, there was an item that
granted you the ability to telepathically detect all the items on the
map. Once people know 100% sure they've gathered all the items on the
level, they don't care about unexplored territory anymore.
On Aug 10, 10:33 pm, Derek Ray <de...@moot.its.only.a.spamtrap.org>
wrote:
> I'm not much of a completionist, nor am I greedy -- but I do know that
> in a game where randomness is king, failing to check for all possible
> random items can lead to passing up something you'll want later.
Maybe the exploitation versus evasion dilemma is mostly a post-
ascension kit affair?
For example: you encounter a dangerous out of depth monster with more
than half of the level unexplored. You're still starved for items, so
diving down might jeapordize your long term survival. On the other
hand, trying to defeat or otherwise incapacitate the monster could get
you killed right now. What do you do?
The question is, do you really have a choice? By diving down you might
only be postponing the inevitable. Put differently, early evasion is a
strategic error that might leave you scratching your head a few
dungeon levels further, wondering how you've could have handled the
situation differently - while you were just missing a piece of the
puzzle.
Paul Donnelly
> On Aug 10, 9:49 pm, Paul Donnelly <paul-donne...@sbcglobal.net> wrote:
>
>> And I wonder how the existence of an automap affects this. Without a
>> map, would people still feel the urge to fill it in?
>
> I don't know whether it's about filling the automap. In the Japanese
> roguelike Shiren The Wanderer, for example, there was an item that
> granted you the ability to telepathically detect all the items on the
> map. Once people know 100% sure they've gathered all the items on the
> level, they don't care about unexplored territory anymore.
Oh, hey. That's something to write down. I know I'd feel better about
leaving a partially explored level if I could get an idea about what was
still on it.
David Ploog
> On 2009-08-10, David Ploog <pl...@mi.fu-berlin.de> wrote:
>> On Mon, 10 Aug 2009, Paul Donnelly wrote:
>>> And I wonder how the existence of an automap affects this. Without a
>>> map, would people still feel the urge to fill it in?
>>
>> I am not sure if it's the completionist in those players or mere greed. I
>> am definitely prone to this myself.
>
> I'm not much of a completionist, nor am I greedy -- but I do know that
> in a game where randomness is king, failing to check for all possible
> random items can lead to passing up something you'll want later.
Sure. But (part of) a level may be hard -- you might lose more (in
resources or otherwise) than you gain, especially if it's not clear
whether there is any loot at all.
For example, I usually skip exploring Zot:1-4 in Crawl. There is lots of
xp and some loot to be had there, but I find it to be too costly.
David
Derek Ray
> Derek Ray wrote:
>> I'm not much of a completionist, nor am I greedy -- but I do know that
>> in a game where randomness is king, failing to check for all possible
>> random items can lead to passing up something you'll want later.
>
> Don't persistent levels solve this though? I mean, if you know
> that this dragon (and its treasure) will still be here later, and
> that no generated items vanish from the game when you take the stairs,
> do you still feel compelled to take it on before you take the stair?
They both solve it and don't. In the rare instances I encounter OOD
monsters (and yes, even though I've played Crawl, it's still 'rare'),
then of course the correct option is to take the stairs. But this
doesn't stop me from coming back later, or trying to find a way around
to get a look at the unmapped space (in Crawl, multiple stairs, etc.)
>> Level-hoovering is generally _always_ worthwhile
>> unless your food clock is so brutal as to cause starvation -- not
>> actually "fun".
> I disagree. This may be correct in games with nonpersistent
> levels, but with persistent levels - where items and monsters
> once generated stay in the game unless used up or destroyed -
> you lose nothing by checking out the next level.
(The goalposts just moved a bit here, by the way -- level-hoovering isn't
about _never leaving the level_ until you've gotten everything, it's about
cleaning out the levels completely. That said...)
I always go downstairs the minute I see stairs to trigger the "initial
population" of a level based on my current XL, _and_ to mark the
upstairs on my map in case I hit a trapdoor. But I also come right back
up and finish emptying out the level I'm on.
As someone else has noted, difficulty typically increases in roguelikes
as you descend in levels. Why would I increase my difficulty before I
absolutely had to?
Also, there's the "easy to remember" thing; if I know that I'm hoovering
levels in sequence, I don't have to try to remember which levels I
didn't "finish", I just descend stairs until I find the first map that
isn't complete. This leaves my brain free for the business of beating
the game. If I start randomly going up and down stairs, then I have to
keep track of which levels I _have_ cleaned out and which I _haven't_,
which is just wasted mental effort.
Derek Ray
> On Aug 10, 10:33 pm, Derek Ray <de...@moot.its.only.a.spamtrap.org>
> wrote:
>> I'm not much of a completionist, nor am I greedy -- but I do know that
>> in a game where randomness is king, failing to check for all possible
>> random items can lead to passing up something you'll want later.
>
> Maybe the exploitation versus evasion dilemma is mostly a post-
> ascension kit affair?
Possibly. If I'm playing a game that can have an "ascension kit", then
once I have all parts of it, I no longer bother to hoover levels.
In games without a well-defined 'ascension kit' (Crawl), I am always on
the prowl for more single-use toys.
> For example: you encounter a dangerous out of depth monster with more
> than half of the level unexplored. You're still starved for items, so
> diving down might jeapordize your long term survival. On the other
> hand, trying to defeat or otherwise incapacitate the monster could get
> you killed right now. What do you do?
This isn't actually a choice, as mentioned in my response to Ray.
Killed-later is always better than killed-now. But that doesn't mean I
might not try to figure out a way to go around the critter, or come back
to this level a couple XL later.
Derek Ray
> On Mon, 10 Aug 2009, Derek Ray wrote:
>> On 2009-08-10, David Ploog <pl...@mi.fu-berlin.de> wrote:
>>> I am not sure if it's the completionist in those players or mere greed. I
>>> am definitely prone to this myself.
>> I'm not much of a completionist, nor am I greedy -- but I do know that
>> in a game where randomness is king, failing to check for all possible
>> random items can lead to passing up something you'll want later.
>
> Sure. But (part of) a level may be hard -- you might lose more (in
> resources or otherwise) than you gain, especially if it's not clear
> whether there is any loot at all.
No, I won't. Even Crawl does not hold consistently true in this regard;
there just isn't enough of a clock (food, resources, or otherwise) to
force me off of a level passively.
Big nasty OOD monsters will force me off a level, but that's more
"actively" than "passively". Death removes my one non-replaceable
resource, see. :D But even those -- well, Crawl has multiple
stairsets, see. I'll just try to come back up a different way.
> For example, I usually skip exploring Zot:1-4 in Crawl. There is lots of
> xp and some loot to be had there, but I find it to be too costly.
A lot of this depends on your play style, as far as Crawl is concerned.
There's a point where no matter what you find, you're not going to find
anything "better enough" -- and also, Crawl has the potential for a
"dash-n-grab" approach to the Orb, which would logically then be
extended to Zot:1-4 as well as Zot:5.
David Ploog
> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
>> ...but (part of) a level may be hard -- you might lose more (in
>> resources or otherwise) than you gain, especially if it's not clear
>> whether there is any loot at all.
>
> No, I won't. Even Crawl does not hold consistently true in this regard;
> there just isn't enough of a clock (food, resources, or otherwise) to
> force me off of a level passively.
I largely agree. However, there are threats that make me skip a level
completely: in my current DDHe game on CAO, D:16 has the vampire tower.
This means lots of draining and no pacifying. I may be able to clear this
out later, but I probably won't: the draining is there unless I manage to
get rNeg+++. In particular, I may never know which loot the vampires guard
(they have some, I've seen the vault map :).
My Vault:3 has the old secret door vault (I think it's 4x4 very nasty
monsters surrounded by three layers of secret doors). Unfortunately,
they're on the loose by now, which means very many shapeshifters. This is
another level I may skip entirely. (In this case I know that the vault
contains no loot but the rest of the level will contain some items.)
With regard to your "dip into next level, then clear up this one": I think
that's more important for Nethack for monster generation purposes. In
Crawl, we have added small, timed portal vaults that start their timer
once you set foot in the level. The safe approach of checking all
staircases once likely means you'll skip on those vaults.
David
Derek Ray
> On Tue, 11 Aug 2009, Derek Ray wrote:
>> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
>>> ...but (part of) a level may be hard -- you might lose more (in
>>> resources or otherwise) than you gain, especially if it's not clear
>>> whether there is any loot at all.
>> No, I won't. Even Crawl does not hold consistently true in this regard;
>> there just isn't enough of a clock (food, resources, or otherwise) to
>> force me off of a level passively.
> I largely agree. However, there are threats that make me skip a level
> completely: in my current DDHe game on CAO, D:16 has the vampire tower.
> This means lots of draining and no pacifying. I may be able to clear this
> out later, but I probably won't: the draining is there unless I manage to
> get rNeg+++. In particular, I may never know which loot the vampires guard
> (they have some, I've seen the vault map :).
> My Vault:3 has the old secret door vault (I think it's 4x4 very nasty
> monsters surrounded by three layers of secret doors). Unfortunately,
> they're on the loose by now, which means very many shapeshifters. This is
> another level I may skip entirely. (In this case I know that the vault
> contains no loot but the rest of the level will contain some items.)
I have to point out that both of your examples are using not necessarily
objectively-OOD monsters, but certainly subjectively-OOD, ie. critters that
your character can't handle. They're locked behind the doors, so
they're technically passive, but the problem in this case isn't that
there's a clock of some kind forcing you off the level, it's that the
critters are literally in your way and you _can't_ get to that part. It
might as well be a locked door, and as such it doesn't fall into
"voluntary bypass" but rather "forbidden zone".
> With regard to your "dip into next level, then clear up this one": I think
> that's more important for Nethack for monster generation purposes. In
> Crawl, we have added small, timed portal vaults that start their timer
> once you set foot in the level. The safe approach of checking all
> staircases once likely means you'll skip on those vaults.
I call those "traps for the foolish". Having run into enough of those
levels in the past, I know better than to even consider trying to rush
around the map trying to find them (leads to quick death) -- so as far as
I'm concerned, they might as well not exist in the first place. If I
trip over one in normal exploration, that's a bonus, but the risk of
death is too high to be worth the effort. (And magic mapping is never
bad, of course.)
I certainly would never consider not using my "dip to next level to mark
the staircase" strategy simply because of the possibility of missing a
timed portal vault; knowing how to get back to a safe location quickly
is worth far more than any item that doesn't include that as part of its
description. It's all about arranging not to be involved in
emergencies, -- it's just that in Crawl, you have to learn to anticipate
them a lot better than, say, Nethack.
Ray
> On 2009-08-10, Ray <be...@sonic.net> wrote:
>>> Level-hoovering is generally _always_ worthwhile
>>> unless your food clock is so brutal as to cause starvation -- not
>>> actually "fun".
>> I disagree. This may be correct in games with nonpersistent
>> levels, but with persistent levels - where items and monsters
>> once generated stay in the game unless used up or destroyed -
>> you lose nothing by checking out the next level.
> (The goalposts just moved a bit here, by the way -- level-hoovering isn't
> about _never leaving the level_ until you've gotten everything, it's about
> cleaning out the levels completely. That said...)
Ah. I thought you were talking about a different behavior then.
In that case level-hoovering as you call it is rewarding except
in cases where
a) more valuable resources will be consumed by screwing with it
than you can reasonably hope to recover.
b) (a) remains true even throughout the late game. ("I will come
back and kick it's butt for free when I am level 60, moving at
triple speed, and armed with a +15 axe of Wretched Overkill"
is not an option).
c) You have no reason to suspect otherwise.
A monastery filled with vampires who used to be monks and who
still live a monastic life (no weapons, no jewelry, vows of
poverty, etc) for example, might be a place you just never
want to mess with. Risking level-draining and for what?
> I always go downstairs the minute I see stairs to trigger the "initial
> population" of a level based on my current XL, _and_ to mark the
> upstairs on my map in case I hit a trapdoor. But I also come right back
> up and finish emptying out the level I'm on.
That is ... interesting. I do not believe that the hero's level
should be used as an input into monster generation. You have just
demonstrated yet another reason why that is true, which I had not
yet considered. I prefer a model where the same creatures are
living there, in possession of the same stuff, no matter who the
adventurer is. In fact, if it weren't for the fact that savefile
abuse would become too rewarding, the entire map, population, and
treasures would be decided from the moment you started the game.
> As someone else has noted, difficulty typically increases in roguelikes
> as you descend in levels. Why would I increase my difficulty before I
> absolutely had to?
> Also, there's the "easy to remember" thing; if I know that I'm hoovering
> levels in sequence, I don't have to try to remember which levels I
> didn't "finish", I just descend stairs until I find the first map that
> isn't complete. This leaves my brain free for the business of beating
> the game. If I start randomly going up and down stairs, then I have to
> keep track of which levels I _have_ cleaned out and which I _haven't_,
> which is just wasted mental effort.
This is all worthwhile input from a game design POV; but mostly it
challenges me to identify the underlying assumptions and think of
ways to subvert or obstruct your reasoning.
Many of these dynamics are subvertable somewhat by a long but
quite shallow power curve combined with treasure and monster
distribution that's highly random and much less dependent on
dungeon level than we've seen in most games. If someone has
a chance of winning, at a fairly low level and with equipment
they might have acquired early, provided they can get the
character to a *VERY* difficult-to-reach spot on the last
floor of the dungeon to do whatever it is that wins the game,
then physical progress (in location as opposed to kit and
level) becomes more important to winning the game.
Bear
David Ploog
> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
>> On Tue, 11 Aug 2009, Derek Ray wrote:
>>> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
>> With regard to your "dip into next level, then clear up this one": I think
>> that's more important for Nethack for monster generation purposes. In
>> Crawl, we have added small, timed portal vaults that start their timer
>> once you set foot in the level. The safe approach of checking all
>> staircases once likely means you'll skip on those vaults.
>
> I call those "traps for the foolish". Having run into enough of those
> levels in the past, I know better than to even consider trying to rush
> around the map trying to find them (leads to quick death) -- so as far as
> I'm concerned, they might as well not exist in the first place. If I
> trip over one in normal exploration, that's a bonus, but the risk of
> death is too high to be worth the effort. (And magic mapping is never
> bad, of course.)
The small ones are not announced -- so you wouldn't know they exist from
stair dipping!
> I certainly would never consider not using my "dip to next level to mark
> the staircase" strategy simply because of the possibility of missing a
> timed portal vault
From my point of view as a designer, I am most happy if the answer to some
question (e.g. "Should I stair dip?") is conditional, i.e. depends on the
circumstances. A good indicator for that if there are vocal factions for
both answers. You represent the "dip at all costs" faction, which is fine
by me :)
(By the way, it would be trivial to make stair dipping the bad choice:
simply assume every other level had a timed, unannounced portal vault with
great loot.)
David
Derek Ray
>> population" of a level based on my current XL, _and_ to mark the
>> upstairs on my map in case I hit a trapdoor. But I also come right back
>> up and finish emptying out the level I'm on.
>
> That is ... interesting. I do not believe that the hero's level
> should be used as an input into monster generation.
I'm not sure I'm inclined to agree based on only this example, simply
because the problem in this case is not that someone used the hero's XL,
it's that someone pre-populated a level with monsters presuming that the
hero's first entrance to the level coincides with his beginning to
explore the level.
Different methods of monster generation are a huge derail to this issue,
however. :)
> I prefer a model where the same creatures are
> living there, in possession of the same stuff, no matter who the
> adventurer is. In fact, if it weren't for the fact that savefile
> abuse would become too rewarding, the entire map, population, and
> treasures would be decided from the moment you started the game.
Correct. Generation "on the fly" almost has to be mandatory to a
certain degree, since savescumming is not practically preventable (and
thus not worth bothering about).
>> As someone else has noted, difficulty typically increases in roguelikes
>> as you descend in levels. Why would I increase my difficulty before I
>> absolutely had to?
> This is all worthwhile input from a game design POV; but mostly it
> challenges me to identify the underlying assumptions and think of
> ways to subvert or obstruct your reasoning.
Of course. :D It's no accident that some of the changes in Spork only
really hose people who are familiar with the shortcuts in Nethack --
people who are playing it "straight up" may not ever notice a
difference. In many cases, I went directly from "How do I bypass this
challenge" to "and behold, a solution and/or patch."
> Many of these dynamics are subvertable somewhat by a long but
> quite shallow power curve combined with treasure and monster
> distribution that's highly random and much less dependent on
> dungeon level than we've seen in most games. If someone has
> a chance of winning, at a fairly low level and with equipment
> they might have acquired early, provided they can get the
> character to a *VERY* difficult-to-reach spot on the last
> floor of the dungeon to do whatever it is that wins the game,
> then physical progress (in location as opposed to kit and
> level) becomes more important to winning the game.
I have some ideas on this myself, but I'm experimenting with them first;
if they turn out to be even plausible in jabtesting, I'll probably drag
them through here for a collective look-see.
Derek Ray
> On Tue, 11 Aug 2009, Derek Ray wrote:
>> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
>>> With regard to your "dip into next level, then clear up this one": I think
>>> that's more important for Nethack for monster generation purposes. In
>>> Crawl, we have added small, timed portal vaults that start their timer
>>> once you set foot in the level. The safe approach of checking all
>>> staircases once likely means you'll skip on those vaults.
>> I call those "traps for the foolish". Having run into enough of those
>> levels in the past, I know better than to even consider trying to rush
>> around the map trying to find them (leads to quick death) -- so as far as
>> I'm concerned, they might as well not exist in the first place. If I
>> trip over one in normal exploration, that's a bonus, but the risk of
>> death is too high to be worth the effort. (And magic mapping is never
>> bad, of course.)
>
> The small ones are not announced -- so you wouldn't know they exist from
> stair dipping!
Then that makes them even less meaningful to me. Which I'm OK with,
because I've been tempted to use lua to block the "OMG PORTAL CLOSING"
timed announcements; either I get to it or I don't. :)
>> I certainly would never consider not using my "dip to next level to mark
>> the staircase" strategy simply because of the possibility of missing a
>> timed portal vault
> From my point of view as a designer, I am most happy if the answer to some
> question (e.g. "Should I stair dip?") is conditional, i.e. depends on the
> circumstances. A good indicator for that if there are vocal factions for
> both answers. You represent the "dip at all costs" faction, which is fine
> by me :)
Be careful here, though. There are highly vocal, very popular factions for
some clearly second-rate strategies in Nethack; they're so popular
because they seem _so_ clever that people just can't believe it would be
the wrong choice. But the raw results indicate otherwise...
I haven't played Crawl long enough to know for sure whether what I'm
doing is right. But I can say that I know your design style well enough
to be confident that my assessment of these as "traps; don't turn them
down but don't break discipline to chase them" is accurate. (You're
still too cautious to put incredible loot in 'em. :)
> (By the way, it would be trivial to make stair dipping the bad choice:
> simply assume every other level had a timed, unannounced portal vault with
> great loot.)
Sure, but then you swing the predictability factor the other way; if
portal vaults are _that_ common, I avoid touching stairs except in dire
emergency, and I do what I can to prepare for the rush-to-loot. :D (I
know a happy medium isn't easy to obtain, but it's still vaguely
amusing to watch the pendulum swing.)
Ray
> On 2009-08-11, Ray <be...@sonic.net> wrote:
>> Many of these dynamics are subvertable somewhat by a long but
>> quite shallow power curve combined with treasure and monster
>> distribution that's highly random and much less dependent on
>> dungeon level than we've seen in most games.
> I have some ideas on this myself, but I'm experimenting with them first;
> if they turn out to be even plausible in jabtesting, I'll probably drag
> them through here for a collective look-see.
I have some ideas along these lines myself. It came up while
I was thinking of the reasons people "grind" for experience.
The character level mechanic makes XP *THE* premium resource
you have to harvest on each and every map. Most roguelikes
reenforce this in dozens of different ways, which I could
enumerate if you want. But XP makes them very much games of
linear progress, and virtually requires you to kill almost
everything you meet.
While not prepared to abandon the character-level mechanic
entirely, I had planned to de-emphasize it somewhat by making
smaller the differences between lower-level and higher-level
characters, relying on specific buffs/debuffs/equipment
rather than character level for most "powers" (which also ties
into a crusade against Thousanduplets), and broadening
dramatically the range of dungeon levels and character levels
at which everything is available.
The chat about "level-hoovering" as you call it naturally prompted
me to think about it in terms of the above plan and I was delighted
to find that the shallow power curve sort of breaks the reasoning
behind it.
Bear
Cuboidz
> I have some ideas along these lines myself. It came up while
> I was thinking of the reasons people "grind" for experience.
Character levels only make sense when you need to decouple character
skill from player skill. This is important in Dungeons & Dragons,
because it's a role playing game, and MMORPG's, because they need to
appeal to a wide audience of less than skilled players.
Roguelikes, on the other hand, are obviously skill games, so character
levels are non-essential. In fact, unless they are designed to require
skill, i.e. meaningful choices, character levels are a nuisance. They
make your game harder to balance, for one thing.
Cuboidz
> I have some ideas along these lines myself. It came up while
> I was thinking of the reasons people "grind" for experience.
Character levels only make sense when you need to decouple character
Timofei Shatrov
to confuse everyone with this message:
>On Aug 10, 3:49=A0pm, Paul Donnelly <paul-donne...@sbcglobal.net> wrote:
>> It seems to me that most games support this behavior, since it's
>> necessary to fight a lot to level up and to find useful items.
>>
>> And I wonder how the existence of an automap affects this. Without a
>> map, would people still feel the urge to fill it in?
>
>I don't think so. I do get a certain satisfaction from filling in the
>automap entirely.
>
>It's not really logical to bypass a level if the monsters are giving
>you trouble, since the next level is usually going to be even more
>difficult. Roguelike levels are usually linear in difficulty.
There are many situations where Ironman-ing (going down as soon as you find the
down staircase) is preferable to full exploration. For example:
- limited time (Puppy quest in ADOM)
- dangerous terrain (ToEF in ADOM, Hells in Crawl)
- annoying monsters (Zot levels in Crawl)
- lower levels have better rewards and exploring the current level is not
worth it because you're going to die anyway (The Sewer Massacre)
In all of these situations, fighting monsters is not the player's goal. In fact,
it almost never is, it's just that games often don't push the player hard enough
to go straight for their goal and this allows some leisurely monster-whacking.
--
|Don't believe this - you're not worthless ,gr---------.ru
|It's us against millions and we can't take them all... | ue il |
|But we can take them on! | @ma |
| (A Wilhelm Scream - The Rip) |______________|
David Damerell
>That is ... interesting. I do not believe that the hero's level
>should be used as an input into monster generation. You have just
>demonstrated yet another reason why that is true, which I had not
>yet considered. I prefer a model where the same creatures are
>living there, in possession of the same stuff, no matter who the
>adventurer is.
This is a bit realism over gameplay, though; while, of course, stair
dipping is pleasantly absurd, you're throwing away information you can use
to ensure the challenge is appropriate to the player.
It can work, too. NetHack opening game strategy is a bit subtler than
Angband's "sit on DL n until you are completely confident about DL n+1"
for just this reason, the food clock (sadly) being a bit of a joke.
--
David Damerell <dame...@chiark.greenend.org.uk> flcl?
Yesterday was Brieday, July.
Today is Gouday, July.
Tomorrow will be Chedday, July.
Pender
wrote:
> This is a bit realism over gameplay, though; while, of course, stair
> dipping is pleasantly absurd, you're throwing away information you can use
> to ensure the challenge is appropriate to the player.
Disagree. A sufficient reason to object to rewarding this kind of
player behavior is simply that it's a dumb ritual that shouldn't be
rewarded -- no appeal to realism is necessary. I doubt this objection
will go anywhere, though, since it seems to me that primary currency
in trade in Nethack is memorizing and exploiting as many dumb,
interlocking rituals as possible.
To address the Nethack model of adjusting monster difficulty to player
level more directly, the whole business of "ensuring the challenge is
appropriate to the player" is a fundamentally tawdry affair. The point
of gaining levels is to become more powerful. The game leads the
player to think that he will have a marginal advantage of one level
over the monster he faces for every marginal level he gains. In fact,
in Nethack, that's a lie; despite the representations of the
interface, he receives a marginal advantage of only half a level over
the monsters he faces for each level he gains. It's cheating.
Why do it? Why not ignore the player level in choosing monsters but
have each player level confer half as much stat/ability gain? Then the
player would receive the same marginal advantage of what is now half a
level over the monsters per level gained, but this fact would be in
plain sight instead of hidden by a dishonest interface.
David Damerell
>On Aug 12, 10:47=A0am, David Damerell <damer...@chiark.greenend.org.uk>
>>This is a bit realism over gameplay, though; while, of course, stair
>>dipping is pleasantly absurd, you're throwing away information you can use
>>to ensure the challenge is appropriate to the player.
>Disagree. A sufficient reason to object to rewarding this kind of
>player behavior is simply that it's a dumb ritual that shouldn't be
>rewarded -- no appeal to realism is necessary.
Perhaps you missed the bit where I don't like stair-dipping either.
>To address the Nethack model of adjusting monster difficulty to player
>level more directly, the whole business of "ensuring the challenge is
>appropriate to the player" is a fundamentally tawdry affair.
Call it what you like; it works, where the Angband method encourages
endless lingering.
>The point of gaining levels is to become more powerful.
Well, no. We're playing roguelikes, where we expect a stiff challenge at
every turn, not CRPGS where sheer munchkinism is rampant. If the player
can gain levels indefinitely for no penalty, the challenge just went out
of the window.
>Why do it? Why not ignore the player level in choosing monsters but
>have each player level confer half as much stat/ability gain?
Because that cannot produce the effect you have now where you gain a bunch
of levels but because you didn't get any equipment to go with it, when you
go downstairs, something eats your face. This isn't just an interface
issue; if you don't stiffen the challenge as the player levels up (or
wastes time, or _something_) you promote Angband-style grind.
Pender
wrote:
> Well, no. We're playing roguelikes, where we expect a stiff challenge at
> every turn, not CRPGS where sheer munchkinism is rampant. If the player
> can gain levels indefinitely for no penalty, the challenge just went out
> of the window.
Doesn't seem to be a problem in Rogue.
> Because that cannot produce the effect you have now where you gain a bunch
> of levels but because you didn't get any equipment to go with it, when you
> go downstairs, something eats your face.
It's fair, I think, to make it HARD to gain levels. It's fair to
restrict the experience points you can get from a single dungeon
level, or to implement such a punishing exponential curve in level
requirements that it's essentially impossible to progress too fast.
It's NOT fair to have a situation where gaining a level by itself
makes you LESS powerful, especially since the interface advertises the
opposite.
> This isn't just an interface
> issue; if you don't stiffen the challenge as the player levels up (or
> wastes time, or _something_) you promote Angband-style grind.
There are lots of ways to avoid Angband-style grind that don't involve
lying to the player.
Billy Bissette
news:hbf*hw...@news.chiark.greenend.org.uk:
> Quoting Ray <be...@sonic.net>:
>>That is ... interesting. I do not believe that the hero's level
>>should be used as an input into monster generation. You have just
>>demonstrated yet another reason why that is true, which I had not
>>yet considered. I prefer a model where the same creatures are
>>living there, in possession of the same stuff, no matter who the
>>adventurer is.
>
> This is a bit realism over gameplay, though; while, of course, stair
> dipping is pleasantly absurd, you're throwing away information you can
> use to ensure the challenge is appropriate to the player.
If you really want to base monster levels on the hero's level at
entry, then you could somewhat counter dipping by re-evaluating
monster levels when the player re-enters after some duration has
passed. (The player can spend that time getting better, so why
can't the monsters?) It doesn't deal with every issue, but setting
level difficulty to player level has some innate flaws anyway.
> It can work, too. NetHack opening game strategy is a bit subtler than
> Angband's "sit on DL n until you are completely confident about DL
> n+1" for just this reason, the food clock (sadly) being a bit of a
> joke.
Some would argue that "sit on DL n until you are completely
confident about DL n+1" isn't the proper way to play Angband. It
just is the easy way to play until you learn what you are doing.
Just like "sit at stat gain until you max your stats" and "sit at
[X] until you get [Y] resistance" are disputed.
Diving is much more productive, costs you less time spent when
you *do* eventually die, and has been argued to actually be safer
than playing cautious. At certain levels with certain characters,
it quite simply *is* safer to dive and perform intelligent strikes
than to take your time. (This was certainly the case for starting
mages for many years. Mages were so fragile at the start that
playing it safe was almost guaranteed to get you killed by a
simple trap or a single inconvenient enemy, while diving several
dungeon levels meant a single kill would immediately advance you
to competitive durability. I have read complaints that Angband
has been getting progressively easier over the years though.)
Food clock is a non-factor in Angband though, outside of
Ironman. It is something that maybe kills a beginner, and from
that point onwards is nothing but an annoyance. It won't be
removed though due to player inertia. On the other hand,
attempting to make it a factor would require reshaping the
game.
Though one might debate the food clock in other Roguelikes
in general. If food isn't so scarce that it randomly leads to
the death of even skilled players, then it arguably is only a
factor for beginners. Depending on the game, once you learn to
"play" the game, it stops being a driving factor. It doesn't
even necessarily prevent "abusive" behaviors.
Billy Bissette
news:941d4c61-0f19-45a3...@e34g2000vbm.googlegroups.com:
> On Aug 12, 2:36�pm, David Damerell <damer...@chiark.greenend.org.uk>
> wrote:
>> Well, no. We're playing roguelikes, where we expect a stiff challenge
>> at every turn, not CRPGS where sheer munchkinism is rampant. If the
>> player can gain levels indefinitely for no penalty, the challenge
>> just went out of the window.
>
> Doesn't seem to be a problem in Rogue.
Enemies got powerful faster than the player, from what I vaguely
recall?
You might not even need other timers (like food) if progress itself
is your timer.
Pender
> Though one might debate the food clock in other Roguelikes
> in general. If food isn't so scarce that it randomly leads to
> the death of even skilled players, then it arguably is only a
> factor for beginners. Depending on the game, once you learn to
> "play" the game, it stops being a driving factor. It doesn't
> even necessarily prevent "abusive" behaviors.
Even though it's vanishingly rare that I die of starvation in Rogue,
it still constrains my behavior in that certain kinds of monster
farming and too much healing through resting aren't viable strategies.
It's like how stop signs affect our driving patterns even if we've
never gotten a ticket for blowing through a stop sign. The threat of a
consequence is enough to shape even the experts' behavior.
Andrew Doull
> >> It seems to me that most games support this behavior, since it's
> >> necessary to fight a lot to level up and to find useful items.
> > people need a while to learn this, and some never do.
>
> It will be extremely hard to block full exploration of every map before
> proceeding to the next (or "level-hoovering", for short).
>
> >> And I wonder how the existence of an automap affects this. Without a
> >> map, would people still feel the urge to fill it in?
> > am definitely prone to this myself.
>
> I'm not much of a completionist, nor am I greedy -- but I do know that
> in a game where randomness is king, failing to check for all possible
> random items can lead to passing up something you'll want later.
>
> you are to have something you'll want. You can always drop something
> you have and don't want -- you can't pick up something you never had in
> the first place. Level-hoovering is generally _always_ worthwhile
> unless your food clock is so brutal as to cause starvation -- not
> actually "fun".
I want to jump in here.
Level completionism should be strongly discouraged by the game design.
You should always design your game so the amount of time you spend
deeper in the dungeon always generates higher rewards than earlier
in the dungeon. So the incentive is to dive, leaving levels
incomplete,
as much as possible.
Level completionism is grinding by another name...
Andrew
Andrew Doull
> >> Many of these dynamics are subvertable somewhat by a long but
> >> quite shallow power curve combined with treasure and monster
> >> distribution that's highly random and much less dependent on
> >> dungeon level than we've seen in most games.
> > I have some ideas on this myself, but I'm experimenting with them first;
> > if they turn out to be even plausible in jabtesting, I'll probably drag
> > them through here for a collective look-see.
>
> I have some ideas along these lines myself. It came up while
> I was thinking of the reasons people "grind" for experience.
>
> The character level mechanic makes XP *THE* premium resource
> you have to harvest on each and every map. Most roguelikes
> reenforce this in dozens of different ways, which I could
> enumerate if you want. But XP makes them very much games of
> linear progress, and virtually requires you to kill almost
> everything you meet.
It is almost never worth trying to get XP in Angband - you should
always
try to get equipment. A naked level 50 character in Angband will be a
grease stain on the dungeon floor in about the same time as a naked
level 1 character.
This is a good thing IMHO. Having a resource that you can infinitely
accumulate at any level is another way of encouraging grinding...
Andrew
Cuboidz
> You should always design your game so the amount of time you spend
> deeper in the dungeon always generates higher rewards than earlier
> in the dungeon. So the incentive is to dive, leaving levels
> incomplete, as much as possible.
I agree, but this is easier said than done. Higher rewards usually
imply higher risks, as well. As long as exploring the early levels
completely gives players a strategical advantage, "level-hoovering"
will not stop.
> Level completionism is grinding by another name...
Interesting. This would explain why a roguelike like Crawl, that is
designed to minimize grinding, still feels grindy at times. Of course,
given Crawl's huge levels, it shouldn't be a surprise that level
completionism can be tedious.
The battle against grinding is an uphill one, though, since,
subconsciously, many of us _like_ grinding - no matter how much we
argue against it. Discussions about grinding often bleed to death
because nobody is prepared to fully explore the implications of a
grindless roguelike.
Derek Ray
> On Aug 11, 6:33 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
>> in a game where randomness is king, failing to check for all possible
>> random items can lead to passing up something you'll want later.
>>
>> The corollary of course is that the more stuff you have, the more likely
>> you are to have something you'll want. You can always drop something
>> you have and don't want -- you can't pick up something you never had in
>> the first place. Level-hoovering is generally _always_ worthwhile
>> unless your food clock is so brutal as to cause starvation -- not
>> actually "fun".
> I want to jump in here.
>
> Level completionism should be strongly discouraged by the game design.
That's easy to say, much less easy to do.
> You should always design your game so the amount of time you spend
> deeper in the dungeon always generates higher rewards than earlier
> in the dungeon. So the incentive is to dive, leaving levels
> incomplete, as much as possible.
Why should that be an "always"? Also, you appear to be handwaving
vigorously over the process of _actually making that happen_.
> Level completionism is grinding by another name...
No. As long as someone is exploring new territory, it's not "grinding".
Ray
> Though one might debate the food clock in other Roguelikes
> in general. If food isn't so scarce that it randomly leads to
> the death of even skilled players, then it arguably is only a
> factor for beginners. Depending on the game, once you learn to
> "play" the game, it stops being a driving factor. It doesn't
> even necessarily prevent "abusive" behaviors.
I dunno... To me the food game isn't a "clock" per se, it's just a
means of restricting some combinations of player choices. I think
that whether limited food sources are a factor should depend on how
you play and what kind of character you've got. F'r example if
you have a troll, you get fearsome combat power even without any
equipment, but you have to find a lot of food. If the food game
is challenging-but-possible for a troll, then someone playing a
gnome or similar race will have no food problems at all. But gnomes
have to be alert, agile and sneaky to avoid being squished.
In some games spellcasting or the use of some magic items can cause
you to become hungry. Again, you've got a food game that is no
problem for a "normal" character but restricts some types of
behavior. Your troll warrior, for example, will probably starve
to death fast if he wears speed boots (which is an intentional
balancing point because speed boots would multiply his already-
fearsome combat power).
And, yes, the player can (and *should* be able to) accumulate an
unsustainable kit of powerful abilities, items and spells that they
can use for a few rounds now and then to handle major opponents etc,
but need to take off again fairly soon.
Bear
David Ploog
> On 13 aug, 06:51, Andrew Doull <andrewdo...@gmail.com> wrote:
>> Level completionism is grinding by another name...
It's the wrong mail, but here I'd like to mildly protest: grinding comes
in various degrees (compare excessive Elbereth spamming in Nethack with
polypiling your stash). Also, completing levels is different in a game
with persistent levels than from a game without. And among games with
persistent levels, it is a rather minor form of grinding, because you do
nothing repeatedly *on* the level. How boring it is now essentially
depends on the number of levels (not so much on the size).
> Interesting. This would explain why a roguelike like Crawl, that is
> designed to minimize grinding, still feels grindy at times. Of course,
> given Crawl's huge levels, it shouldn't be a surprise that level
> completionism can be tedious.
Agreed. But I believe that the tactical value of large levels is worth it.
We are reacting to the problem in another way, but cutting levels (see
above that the number is more important than the size in regard to the
completing levels grind).
> The battle against grinding is an uphill one, though, since,
> subconsciously, many of us _like_ grinding - no matter how much we
> argue against it. Discussions about grinding often bleed to death
> because nobody is prepared to fully explore the implications of a
> grindless roguelike.
Full-heartedly seconding this. As another example, the recent introduction
of a species without innate healing into Crawl convinced me that innate
healing is such a case: proper, modern, grindfree roguelikes should do
without. (And some do, which is where we got the inspiration to add that
species.)
David
David Ploog
> Though one might debate the food clock in other Roguelikes
> in general. If food isn't so scarce that it randomly leads to
> the death of even skilled players, then it arguably is only a
> factor for beginners. Depending on the game, once you learn to
> "play" the game, it stops being a driving factor. It doesn't
> even necessarily prevent "abusive" behaviors.
I think I can safely say that this is not true within Crawl. In Crawl,
spellcasting costs hunger, the amount depening on the spell levels and
they can be non-negligible. The food clock then means that a caster
(a) has to adapt spell selection/use to the food situation (tactical
effect) and
(b) has to take total food amount into account (strategical effect). It is
not uncommon to see games finished with only marginal food left. (In
general, not enough corpses are created to keep going.)
For non-casters, the food clock is less severe although some species have
higher hunger and there are hunger mutations for everyone to get.
The idea of a food clock is not to starve players to death but to keep
them going: only a novice would probably just starve. A more experienced
player would see the food shortage in advance and react: by skipping full
exploration in favour of going to new levels, taking risks rather than
avoiding fights (in order to get corpses).
David
Ray
> On 13 aug, 06:51, Andrew Doull <andrewdo...@gmail.com> wrote:
>
>> You should always design your game so the amount of time you spend
>> deeper in the dungeon always generates higher rewards than earlier
>> in the dungeon. So the incentive is to dive, leaving levels
>> incomplete, as much as possible.
>
> I agree, but this is easier said than done. Higher rewards usually
> imply higher risks, as well. As long as exploring the early levels
> completely gives players a strategical advantage, "level-hoovering"
> will not stop.
I think the mistake was in giving the player an option which
is substantially safe. A character who is "safe" on level 1
has no incentive to go deeper because deeper represents an
*infinite* increase in the amount of danger the character
is facing.
Most games, players facing the "go down stairs or not" question,
are looking at a cost-benefit equation that goes something like
this: "going deeper will get me in reach of rewards 5% greater
than those I can get here. It will also put me in 5000% greater
danger per round of death." So they continue until the next level
is also substantially safe, and the ratio of increased rewards
to increased danger is therefore roughly equal.
If a game incentivizes going deeper, it means that most of the time
the marginal danger increases when you go down stairs by a *smaller*
percentage than does the marginal reward.
And that means that in such a game no character can be completely
safe, ever, anywhere.
If the player finds a completely safe way for the character to
reap some reward, then doing anything else - anything which involves
*any* danger - means increasing his marginal risk *infinitely*
(compared to his current level of risk, ie, zero) for some *finite*
(compared to his current level of reward, ie, nonzero) marginal
reward. Consciously or not, he does the cost/benefit analysis, and
poof, you get grinding.
In order to make the cost/benefit analysis favor moving on to the
next level before it's *completely* safe to do so, you have to
prevent the current level from ever being *completely* safe. There
has to be some positive baseline danger to compare with the
increased danger, just as there is a positive baseline reward to
compare to the increased level of reward.
This is why Angband mages, who are so very fragile that they're in
serious danger on level one, usually want to dive to level 3 or 4
in the early game. The marginal increase in danger compared to
their high baseline danger level on DL1 is roughly equivalent
to the increase in rewards compared to what they can get on DL1.
For more durable characters who are safer on DL1, going downstairs
early is crazy by comparison.
Even though they wouldn't put themselves in nearly as much absolute
danger, the percentage by which they'd *increase* their danger level
is out of all proportion to the percentage by which they'd increase
their reward.
> The battle against grinding is an uphill one, though, since,
> subconsciously, many of us _like_ grinding - no matter how much we
> argue against it. Discussions about grinding often bleed to death
> because nobody is prepared to fully explore the implications of a
> grindless roguelike.
I think a grindless roguelike will occasionally murder your
characters for no good reason, due to the analysis I outlined
above. And I think I *am* prepared to fully explore the
implications of that conclusion. What do you think a grindless
roguelike looks like?
Bear
David Damerell
>On Aug 12, 2:36=A0pm, David Damerell <damer...@chiark.greenend.org.uk>
>>Well, no. We're playing roguelikes, where we expect a stiff challenge at
>>every turn, not CRPGS where sheer munchkinism is rampant. If the player
>>can gain levels indefinitely for no penalty, the challenge just went out
>>of the window.
>Doesn't seem to be a problem in Rogue.
But Rogue has a stiff food clock. Sure, there are other ways around this
problem (like a stiff food clock) but they have gameplay consequences too.
>It's fair, I think, to make it HARD to gain levels.
I don't care what's fair. I care what produces interesting gameplay. And
in this particular case, NetHack does better than Angband because the
decision to grind endlessly is not merely about how much tedium you can
stomach.
>>This isn't just an interface
>>issue; if you don't stiffen the challenge as the player levels up (or
>>wastes time, or _something_) you promote Angband-style grind.
>There are lots of ways to avoid Angband-style grind that don't involve
>lying to the player.
I think the characterisation of this as "lying to the player" is as
unhelpful as it is nonsensical. As you get stronger, stronger monsters come
looking for you. That's really not all that remarkable.
--
David Damerell <dame...@chiark.greenend.org.uk> Distortion Field!
Yesterday was Gouday, July.
Today is Chedday, July.
Tomorrow will be Stilday, July - a weekend.
David Damerell
>David Damerell <dame...@chiark.greenend.org.uk> wrote in
>>Angband's "sit on DL n until you are completely confident about DL
>>n+1" for just this reason, the food clock (sadly) being a bit of a
>>joke.
>Some would argue that "sit on DL n until you are completely
>confident about DL n+1" isn't the proper way to play Angband.
I think the current results shown by the Borg show that it is - at the
very least from once you are over the immediate infant mortality problem
with fragile classes you document [1], until you get very deep in the dungeon
and there's really no way to get confident about n+1. The APW Borg does
play this strategy and, last I looked at it, essentially never loses
characters in the middle game.
>Food clock is a non-factor in Angband though, outside of Ironman.
It's slightly less of a joke in NetHack, but only because it occasionally
consume a prayer you might need for something else.
[1] In which case you _can't_ play this strategy because you can't be
confident about DL 1.
David Damerell
>The battle against grinding is an uphill one, though, since,
>subconsciously, many of us _like_ grinding - no matter how much we
>argue against it.
I think it's not so much that we _like_ it as that it's seductive in two
ways.
If the opportunity to grind exists and it's clearly the right strategy,
well, I want to win. If I don't take the opportunity, I'm playing with
one hand tied behind my back.
The second way is the World of Warcraft way - it's sort of moreish even
though it isn't very nice and it's all the same, like a tube of Pringles.
You can be bored stiff but keep pressing the lever for the food pellets
even though you don't really know why.
David Ploog
> ... that means that in such a game no character can be completely
> safe, ever, anywhere.
I agree. We try to balance Crawl such that no player is ever safe. This is
not only about preventing grinding, it has two other purposes:
(1) If a game can be won safely, it is broken in the sense that skill
takes a backseat over endurance and spoiler knowledge. Nethack is almost
there. If a game cannot be won safely, we'll obviously get unfair games --
the unfairness could be the sudden appearance of three lethal bad guys who
snipe the player in two rounds, or also the absolute lack of poison
resistance (say). In either case, good players will be distinguished from
the best: the latter will simply be able to pull more victories from such
games. This comparison of high skills would be lost otherwise.
(2) We, as human designers, are not able to tweak game balance such that
*barely winning* is always possible but sloppy play is not. At least with
the random approach to roguelikes, I don't see a way to do that.
Therefore, we can err in favour of the player or in favour of the game. I
suggest to do the latter.
Put in another way, if we try to make sure the game can be won, it will be
much, much easier than we wanted to originally, because of our inability
to precisely hit the thin line between victory and failure.
> I think a grindless roguelike will occasionally murder your
> characters for no good reason, due to the analysis I outlined
> above. And I think I *am* prepared to fully explore the
> implications of that conclusion. What do you think a grindless
> roguelike looks like?
As I said, your conclusion is spot on, to me. To me, grindless also means
these:
* get rid of innate healing -- it leads to grinding on the miniscule level
(optimal play: rest after every fight)
* no ascension kits -- make sure that winning characters must look
different (this achieves that players cannot work down a laundry list of
things to get/do)
* all resources must be finite -- in particular, persistent levels, no
unlimited death drops
* have at least one clock that drives the game (this need not be food:
I've seen action game [1] where the level would slowly flood with water,
keeping you running up and up)
[1] The Killing Game Show on the Amiga did this in a very good way. It is
a cool jump and run (with lots of shooting) and they avoided
another type of grinding so common in non-random action games: if you
start a new game and do nothing, it would reply your old game. You can
pick up the joystick at any moment and take over control.
David
Krice
wrote:
> If the opportunity to grind exists and it's clearly the
> right strategy, well, I want to win.
Yes. The design of some roguelikes require the player to use
certain kind of tactics, let's call them grinding, in order
to win the game. The problem is not in grinding, it's in the
original design. In particular the concept of experience is
one of the main reasons for grinding.
David Ploog
This might be so in the standard approach. But what if only the first
killed monster of a given type gave experience? I think that experience is
a natural measure of progress -- and we will want characters that are weak
at the start and become more powerful later on. The problem is rather with
the hordes of meaningless monsters than with the experience, in my view.
David
Billy Bissette
news:Pine.LNX.4.62.09...@paste.mi.fu-berlin.de:
> (1) If a game can be won safely, it is broken in the sense that skill
> takes a backseat over endurance and spoiler knowledge.
I'd like to think that this can be argued. Even if everyone can win
eventually, you can distinguish skill by some other measurement.
Pretty much anyone can complete a marathon if they aren't concerned
about time, don't do anything stupid, and don't fall victim to bad
luck. It is still considered an accomplishment to complete a marathon.
But talent is measured in how long it takes to complete the race.
You can coin-feed through many shoot'em-ups, which makes them a
bit more accessible to average players. Skill is measured through
beating such a game on one credit and on achieving high scores.
The scoring systems themselves tend to lay multiple restrictions on
how a player plays a game.
The two tricky parts for Roguelikes would be trying to find
viable alternative measurements of victory (turn counts, conducts
in Nethack, etc) and getting players to actually acknowledge and
challenge these measurements. Which, honestly, might not work
particularly well.
Pender
> As I said, your conclusion is spot on, to me. To me, grindless also means
> these:
> * get rid of innate healing -- it leads to grinding on the miniscule level
> (optimal play: rest after every fight)
> * no ascension kits -- make sure that winning characters must look
> different (this achieves that players cannot work down a laundry list of
> things to get/do)
> * all resources must be finite -- in particular, persistent levels, no
> unlimited death drops
> * have at least one clock that drives the game (this need not be food:
> I've seen action game [1] where the level would slowly flood with water,
> keeping you running up and up)
All of these can be accomplished with a real food clock. None of the
rest are necessary.
1) Innate healing takes time, which consumes food. If there's only so
much food, it's no longer optimal to rest after every fight. A real
food clock means HP is only a semi-renewable resource.
2) I agree that eliminating ascension kits and the shopping lists that
they engender is a good thing for reasons other than eliminating
grind, but a real food clock means you can't just grind your way to a
complete kit. You have to take what you can find.
3) Only food really has to be finite; neither levels nor monster drops
need be limited. Perhaps only X rations of food can be collected per
depth level, after which they stop spawning on that level. I prefer
persistent levels, again for reasons besides eliminating grind, but a
true food clock in Angband with a finite supply of food would totally
change the character of the gameplay, and for the better in my
opinion. Similarly, farming monsters with item drops also takes time,
and therefore food, and therefore is not unlimited either.
Definitely agree with David Damerell that eliminating grind falls to
the game designer, not the player. It's always baffled me that Angband
and Nethack both shamelessly broke the food clock when they split off
from Rogue -- and to the extent that the food clock is the only thing
preventing grind, getting rid of it had predictable effects.
On Aug 13, 12:51 am, Andrew Doull <andrewdo...@gmail.com> wrote:
> Level completionism should be strongly discouraged by the game design.
>
> You should always design your game so the amount of time you spend
> deeper in the dungeon always generates higher rewards than earlier
> in the dungeon. So the incentive is to dive, leaving levels
> incomplete,
> as much as possible.
>
> Level completionism is grinding by another name...
If the gameplay remains interesting, why is vertical exploration
intrinsically preferable to horizontal exploration? And if the levels
are perfectly safe for the player, then vertical exploration is just
as mindless and grindy as horizontal exploration. Level completionism
is only a symptom of the problem in Angband, not the problem itself.
David Ploog
> David Ploog <pl...@mi.fu-berlin.de> wrote
>
>> (1) If a game can be won safely, it is broken in the sense that skill
>> takes a backseat over endurance and spoiler knowledge.
>
> I'd like to think that this can be argued. Even if everyone can win
> eventually, you can distinguish skill by some other measurement.
Oh, that is a misunderstanding due to my sloppy terminology: by "won
safely" I meant that one player can win every game. This is distinct from
any player can win some game which is a fine goal.
David
David Ploog
> On Aug 13, 1:10 pm, David Ploog <pl...@mi.fu-berlin.de> wrote:
>
>> As I said, your conclusion is spot on, to me. To me, grindless also means
>> these:
>> * get rid of innate healing
>> * all resources must be finite
>
> rest are necessary.
[snip reasoning]
This is a good observation but I think it is still worth pointing out that
you can achieve these goals without a food clock. This is useful because
we want roguelikes to be diverse, not all alike.
> Definitely agree with David Damerell that eliminating grind falls to
> the game designer, not the player.
Yes. In Crawl, we are trying to kill scum and grindtraps mercilessly and
the usual argument "but players who don't like it don't have to do it"
(which you'll be familiar with from r.g.r.nethack) won't cut it for us.
David
Kusigrosz
> In order to make the cost/benefit analysis favor moving on to the
> next level before it's *completely* safe to do so, you have to
> prevent the current level from ever being *completely* safe. There
> has to be some positive baseline danger to compare with the
> increased danger, just as there is a positive baseline reward to
> compare to the increased level of reward.
means increasing the chance of winning the game. When the PC arrives on
a level, there has to be an expected reward (why visit the level
otherwise?), but only for the PC of that moment, and the initial level
stuff untouched.
--
Kusi...@AUtorun.itvk.pl To send mail, remove 'AU' from the address
[No net access for the next few days]
You see here a scroll labeled "Q-e0fHUoKD8"
Billy Bissette
news:97da10d3-0d1f-44b7...@z28g2000vbl.googlegroups.com:
> On Aug 13, 1:10 pm, David Ploog <pl...@mi.fu-berlin.de> wrote:
>
>> As I said, your conclusion is spot on, to me. To me, grindless also
>> means these:
>> * get rid of innate healing -- it leads to grinding on the miniscule
>> level
>> (optimal play: rest after every fight)
This isn't so much innate healing as it is healing over time or
healing through effectively self-replinishing resources. Anything
which allows "optimal" play where you spend your game time recovering
rather than advancing the game.
If you remove innate healing but retain innate magic point recovery
and have healing spells, then you've only really changed the specific
thing that you sit around recovering. Same for if you give the player
the ability to pray every thousand turns to heal.
Even things like vampiric weapons/attacks could theoretically lead
to degenerate gameplay where you regularly spend time topping off
your character on weaklings rather than facing serious threats. Or
an extremely limited amount of healing items could cause players to
try to farm weak creatures that can produce them.
At the same time, you could theoretically go grindless by boosting
healing. Give the player full life recovery after every encounter
(if you could figure out some way to measure an encounter,) and you
also avoid grinding to recover.
>> * no ascension kits -- make sure that winning characters must look
>> different (this achieves that players cannot work down a laundry list
>> of things to get/do)
>> * all resources must be finite -- in particular, persistent levels,
>> no unlimited death drops
>> * have at least one clock that drives the game (this need not be
>> food: I've seen action game [1] where the level would slowly flood
>> with water, keeping you running up and up)
>
> All of these can be accomplished with a real food clock. None of the
> rest are necessary.
>
> 1) Innate healing takes time, which consumes food. If there's only so
> much food, it's no longer optimal to rest after every fight. A real
> food clock means HP is only a semi-renewable resource.
> 2) I agree that eliminating ascension kits and the shopping lists that
> they engender is a good thing for reasons other than eliminating
> grind, but a real food clock means you can't just grind your way to a
> complete kit. You have to take what you can find.
This isn't entirely true for ascension kits.
It is possible to end up with a game design where certain equipment
is seen as "mandatory", or at least highly preferable. A tight food
clock doesn't address the existance of such a kit, it just means
fewer players will actually obtain it, and (depending on game
balance) might be forced to make potentially suicidal end game
attempts without it.
Secondly, depending on game design, obtaining these useful items
might not be so much grinding away waiting for blind luck to favor
you. Nethack, for example, has wishing as an alternative means of
getting specific items. Items that form such a kit might even be
guaranteed in a way where a food clock cannot realistically prevent
their gain. For example, as rewards for mandatory (or nearly
mandatory) sections of the game, or expected achievements.
> On Aug 13, 12:51�am, Andrew Doull <andrewdo...@gmail.com> wrote:
>>
>> Level completionism is grinding by another name...
>
> If the gameplay remains interesting, why is vertical exploration
> intrinsically preferable to horizontal exploration? And if the levels
> are perfectly safe for the player, then vertical exploration is just
> as mindless and grindy as horizontal exploration. Level completionism
> is only a symptom of the problem in Angband, not the problem itself.
Angband suffers because, given the choice, many people would rather
play safe than play exciting.
I've repeatedly seen people say that Angband is boring. But when
you tell them to dive, they refuse because it is dangerous. They'd
rather play a game in a way that they find boring than to voluntarily
take the game at a more risky pace. The same happens with stat gain
and certain resists. Some people, as long as they aren't forced to
go faster, will spend as much time as necessary to max themselves
and then will complain about spending that time.
A stiff food clock doesn't entirely address the problem, though.
After a fashion, it sweeps it under the rug. The nature of the
player is still there. You've just disallowed certain styles of
play. At the same time, there are some people who honestly like
being able to set their own pace, which a stiffer food clock may
very well prevent.
Kusigrosz
recently faced challenges? I think it is the accumulation of
non-decaying assets that allows grinding. The decay doesn't have to be
immediately dangerous to the PC, it's just enough if it makes grinding
a suboptimal strategy.
--
Kusi...@AUtorun.itvk.pl To send mail, remove 'AU' from the address
[no net access for the next few days]
Billy Bissette
> On Thu, 13 Aug 2009, Billy Bissette wrote:
I'm not sure how much it actually is a misunderstanding. We do
have different viewpoints, after all. Nothing bad about that.
Having other measurements of skill besides "winning" works for
either terminology. Such other measurements are something of a
cheat though, in that they redefine "winning" into something
beyond just "beating the game", something that gives more of an
accounting of skill.
Krice
> I think that experience is a natural measure of progress
No, it's just a result of a certain kind of design, where
the player meet monsters that require superhuman powers
he can reach only by "experience". But it's nothing like
real experience, it's just a strange counter, a reward from
killing a lot of monsters (=grinding).
Experience (in role-playing games) and grinding go hand
in hand. I think attempts to fix that are limited in
experience-based gameplay, but you can always try.
Xecutor
> Experience (in role-playing games) and grinding go hand
> in hand. I think attempts to fix that are limited in
> experience-based gameplay, but you can always try.
I'd say experience in real life is also often
gained thru some kind of grinding-like exercises.
A lot of things can only be learned thru
numerous repetitions.
For YEARS!
So, compared to this, grinding in RLs is like a childsplay.
Recently I tried to play a korean action rpg on iPhone.
That's grinding!!!
Monsters are respawning constantly.
And you cant leave current area until you
are strong enough to beat monsters in next one.
And you are sure not strong enough even after
cleaning area 2-3 times.
In order to complete drop-gathering quest for 5 items,
you need to kill like 100 monsters.
That's grinding!
Xecutor
> You should always design your game so the amount of time you spend
> deeper in the dungeon always generates higher rewards than earlier
> in the dungeon. So the incentive is to dive, leaving levels
> incomplete,
> as much as possible.
If I *LIKE* to explore levels, why should I stop this?
The games are here to entertain players, not to
fulfill designer's dreams :)
Andrew Doull
wrote:
> On 2009-08-13, Andrew Doull <andrewdo...@gmail.com> wrote:
> > You should always design your game so the amount of time you spend
> > deeper in the dungeon always generates higher rewards than earlier
> > in the dungeon. So the incentive is to dive, leaving levels
> > incomplete, as much as possible.
>
> Why should that be an "always"? Also, you appear to be handwaving
> vigorously over the process of _actually making that happen_.
I'd like to think that developing Unangband was a little more than
handwaving - and I think I've achieved at least some of this
requirement. I'll concede that the only real evidence is in the code,
but I find writing about the design to be a useful process for
thinking about the implementation details.
Andrew
Andrew Doull
> in a game where randomness is king, failing to check for all possible
> random items can lead to passing up something you'll want later.
Assuming that your random items are generated independently of each
other, that's incorrect. If a random item is generated, which you
never find, it's exactly the same as not being generated at all. If
you choose to do something else which results in you encountering the
same number and distribution of random drops, you're no worse off
ignoring the first n items.
An analogous statement to what you're saying is you're less likely to
roll a six on a dice, if the previous dice roll was also a six.
Now, in a game without random drops, you do want to check every
possible location - because a designer will place items deliberately
to reward this activity.
Andrew
Andrew Doull
Which is why I'm really interested in the idea of non-transferable
experience. By this, I mean difference types of experience, each of
which only accumulates to a certain level, and no further. So you
start off with level 1 xp points. When you max those out, you have to
descend to level 2 to progress further, where you start earning level
2 experience points. These also max out, so you move to level 3 etc.
This way, you are forced to progress deeper in the dungeon, and to
higher risk areas, in order to get any benefit from playing further.
Andrew
Derek Ray
> On Aug 13, 10:17 pm, Derek Ray <de...@moot.its.only.a.spamtrap.org>
> wrote:
>> On 2009-08-13, Andrew Doull <andrewdo...@gmail.com> wrote:
>> > You should always design your game so the amount of time you spend
>> > deeper in the dungeon always generates higher rewards than earlier
>> > in the dungeon. So the incentive is to dive, leaving levels
>> > incomplete, as much as possible.
>>
>> Why should that be an "always"? Also, you appear to be handwaving
>> vigorously over the process of _actually making that happen_.
>
> I'd like to think that developing Unangband was a little more than
> handwaving
Handwaving, in this case, is the part where you should perhaps be
explaining how it is possible to do such a thing, rather than just
saying "You should do it like this."
Still didn't mention why it should be an "always", either.
Derek Ray
> wrote:
>> I'm not much of a completionist, nor am I greedy -- but I do know that
>> in a game where randomness is king, failing to check for all possible
>> random items can lead to passing up something you'll want later.
>
> Assuming that your random items are generated independently of each
> other, that's incorrect. If a random item is generated, which you
> never find, it's exactly the same as not being generated at all.
> If you choose to do something else which results in you encountering the
> same number and distribution of random drops, you're no worse off
> ignoring the first n items.
But that's not the case in most roguelikes with persistent levels; it is
certainly not true in Nethack and/or Crawl. The items on the floor are
important; they exist merely for entering the level. Death drops are
also a way to get items, but that occurs at a much lower rate and
(depending on the critter) may not even yield you the items you desire.
Once you start really grinding against monsters for items, you start
running into alternate resource issues you must solve (food clock, etc).
What it comes down to is -- skilled play involves not passing up the easy
items which you get just for walking around, just because you can grind
later and make more items.
> An analogous statement to what you're saying is you're less likely to
> roll a six on a dice, if the previous dice roll was also a six.
Would you like a shoehorn and some WD-40 to help you make that
analogy fit? 'cos it looks like you missed by a mile, there.
Pender
> Angband suffers because, given the choice, many people would rather
> play safe than play exciting.
>
> I've repeatedly seen people say that Angband is boring. But when
> you tell them to dive, they refuse because it is dangerous. They'd
> rather play a game in a way that they find boring than to voluntarily
> take the game at a more risky pace. The same happens with stat gain
> and certain resists. Some people, as long as they aren't forced to
> go faster, will spend as much time as necessary to max themselves
> and then will complain about spending that time.
>
> A stiff food clock doesn't entirely address the problem, though.
> After a fashion, it sweeps it under the rug. The nature of the
> player is still there. You've just disallowed certain styles of
> play. At the same time, there are some people who honestly like
> being able to set their own pace, which a stiffer food clock may
> very well prevent.
See, this is that "blame the player" mentality again. I think it's
pretty much axiomatic that roguelike designers should assume that
players will try to play optimally and therefore design their games so
that optimal play is also fun play. Angband misses that mark by a
mile, since optimal play involves excruciatingly protracted grinding.
A real food clock would be the easiest solution.
Actually maybe that's not quite fair. Some people claim to enjoy
grinding. I always saw Angband as a sort of single-player roguelike
MMORPG. The endless accumulation of ever-so-marginally better items
and stats -- that could literally last for weeks -- struck me as very
similar to the Diablo II or WoW experience. Lots of people like Diablo
II and WoW. But on this score, I'm with David Damerell and that guy
who made Braid: they're not fun so much as addictive, and you find
yourself pulling the lever for food pellets hour after hour even as
you wish you could stop.
Kenneth 'Bessarion' Boyd
> Assuming that your random items are generated independently of each
> other, that's incorrect. If a random item is generated, which you
> never find, it's exactly the same as not being generated at all. If
> you choose to do something else which results in you encountering the
> same number and distribution of random drops, you're no worse off
> ignoring the first n items.
Not true in general: you're minus the utility of those n items (and
gain the corresponding amount of time in not looking for them).
With non-persistent levels, the time cost can win out (what matters is
maximizing utility/time). With persistent levels, you're plain out n
items, and the only way you could come out ahead is if most of the n
items were junk in combination with being personally critically
overpowered.
> An analogous statement to what you're saying is you're less likely to
> roll a six on a dice, if the previous dice roll was also a six.
Only with non-persistent levels.
> Now, in a game without random drops, you do want to check every
> possible location - because a designer will place items deliberately
> to reward this activity.
Finite random drops elicits this as well.
Pender
wrote:
> On Aug 14, 7:06 am, Andrew Doull <andrewdo...@gmail.com> wrote:
>
> > Assuming that your random items are generated independently of each
> > other, that's incorrect. If a random item is generated, which you
> > never find, it's exactly the same as not being generated at all. If
> > you choose to do something else which results in you encountering the
> > same number and distribution of random drops, you're no worse off
> > ignoring the first n items.
>
> Not true in general: you're minus the utility of those n items (and
> gain the corresponding amount of time in not looking for them).
You're minus the utility of those n items either way. Doesn't matter
if they were generated or not if you never find them. That was
Andrew's point.
> With non-persistent levels, the time cost can win out (what matters is
> maximizing utility/time). With persistent levels, you're plain out n
> items, and the only way you could come out ahead is if most of the n
> items were junk in combination with being personally critically
> overpowered.
Exploring can also be risky. It's quite possible to have a high
proportion of good items in your game but still frequently choose not
to pursue them. The endgame in Rogue fits this model. Death lurks
around every corner, so it's not worth the risk to seek out the
(finite, generally useful) items.
> > An analogous statement to what you're saying is you're less likely to
> > roll a six on a dice, if the previous dice roll was also a six.
>
> Only with non-persistent levels.
Nope. If item generations are independent events, failing to find a
given item is precisely equivalent to it never having been generated
at all.
David Ploog
> If item generations are independent events, failing to find a given item
> is precisely equivalent to it never having been generated at all.
Yes, but note that "indepedent" clause. In Nethack, if you happen to not
find a generated artefact, your chances for getting other artefacts later
are reduced.
David
Ray
> On 2009-08-13, Ray <be...@sonic.net> wrote:
>> In order to make the cost/benefit analysis favor moving on to the
>> next level before it's *completely* safe to do so, you have to
>> prevent the current level from ever being *completely* safe. There
>> has to be some positive baseline danger to compare with the
>> increased danger, just as there is a positive baseline reward to
>> compare to the increased level of reward.
> Is the positive baseline reward unavoidable?
No, it isn't. And that's a good observation. Thank you.
If the marginal reward for staying on a given level drops to
zero, then the cost/benefit equation will favor going to the
next level where, presumably, some reward exists. There is
by definition some benefit in going to the next level; it
gets your character physically closer to the game location
where winning is possible.
And it's easy to have the marginal reward drop to zero; just
don't replace the initial stock of monsters and items as they
get killed off. No more stuff, no more experience, no more
reason to hang around.
Bear
Cuboidz
> I think a grindless roguelike will occasionally murder your
> characters for no good reason, due to the analysis I outlined
> above. And I think I *am* prepared to fully explore the
> implications of that conclusion. What do you think a grindless
> roguelike looks like?
Interesting analysis. It does seem that certain kinds of improvised
grinding cannot be eliminated. For example:
You've stumbeled upon a snake pit, without any means of dealing with
their deadly poisons. You decide to make a run for the stairs, but due
to a twist of fate, you find a ring of poison resistance on the way.
At that point, you're inclined to reassess. Your cost/benefit analysis
has tilted in favor of killing the snakes, which are now easy prey.
What's a developer to do? You never know when the random number
generator is going to bless (or curse) players with an chance to
grind. What you _can_ do, is keep the window of opportunity small:
- Don't allow players access to previous levels. Coming back later
almost always means waiting until the risk is low. So make stairs one
way only. This will also get rid of another kind of grinding: stash
management. *shudders*
- If the player doesn't want to come to danger, let danger come to the
player. Slaying a horde of monsters can't possibly go unnoticed. Maybe
after a while critters from the next dungeon level will come take a
look, to see what happened to their comrades.
Derek Ray
This only really applies, in practice, to wishing up artifacts; once two
have been generated, it's no longer a 100% chance. And the chance is
reduced whether or not you find the artifact... but the important part
there is that you _really_ want to know that one's been generated.
Sacrificing for 'em... well, it gets slightly slower, but you're already
in a potentially time-consuming process anyway. It's the wish that you
want to hold onto.
Derek Ray
> - If the player doesn't want to come to danger, let danger come to the
> player. Slaying a horde of monsters can't possibly go unnoticed. Maybe
> after a while critters from the next dungeon level will come take a
> look, to see what happened to their comrades.
...Thereby creating another type of grind, as why would a player bother
to do anything but hang out near the stairs, where he knows there are no
traps, and wait for the free XP to come to him?
:)
Cuboidz
wrote:
> On 2009-08-14, Cuboidz <dieter.be...@gmail.com> wrote:
> ...Thereby creating another type of grind, as why would a player bother
> to do anything but hang out near the stairs, where he knows there are no
> traps, and wait for the free XP to come to him?
>
> :)
ACK! Grinding is like a disease, that keeps mutating, and resurfacing,
with a vengeance.
Darren Grey
> Assuming that your random items are generated independently of each
> other, that's incorrect. If a random item is generated, which you
> never find, it's exactly the same as not being generated at all. If
> you choose to do something else which results in you encountering the
> same number and distribution of random drops, you're no worse off
> ignoring the first n items.
>
> An analogous statement to what you're saying is you're less likely to
> roll a six on a dice, if the previous dice roll was also a six.
You're more likely to roll more sixes if you keep rolling the dice.
Exploring every part of every dungeon is letting the player roll the
dice as often as is possible, thus increasing the number of sixes
they'll get overall. The only way to discourage this is to increase
the number of 1s (making the game harder overall) or have some
specific cost for dice-rolling in general. Food is one way, since the
act of exploring reduces your supply, but it's hard to make a game too
harsh on this point without interfering with the fun. Corruption in
ADOM is fairly effective on this point - at lower levels players will
prefer to find the stairs and descend straight away than spend time
exploring an area that is constantly corrupting them. Of course it's
hard to introduce such a mechanic without it feeling artificial and a
general punishment for the player - ADOM succeeds in this area by
making corruption a theme of the whole game, and even making some of
them beneficial.
Traps are another method of punishing the player for exploring too
much, but again they can be infuriating during the course of normal
gameplay. Equipment destroying traps or enemies could also really put
people off seeking out new items, but at the risk of reducing the
player's experience. I remember Izuna had lots of water and fire
enemies in the later dungeons that could destroy your precious
scrolls, and in the end I found it best to zip through dungeons as
quickly as possible. It was not a fun experience.
> Now, in a game without random drops, you do want to check every
> possible location - because a designer will place items deliberately
> to reward this activity.
And similarly the RNG might have items hidden in every nook and
cranny. The chances may be low, but it's hard for the player to turn
his thoughts away from what lies behind every corner.
There's also the issue of wanting to actually experience the game
content fully. I've played plenty of games where due to spoilers I've
known a certain location has no useful items and no NPCs of interest,
and yet I've still gone ahead and explored them to get the most from
the game. If a roguelike creates interesting content then it's fun to
go out and explore every bit of it. Sometimes it will just be a dead
end, but other time it could be an interesting level arrangement, a
challenging group of monsters, or maybe even a special item.
Personally I'd rather a game rewarded players for exploring more
simply by it being a fun game to play.
--
Darren Grey
Derek Ray
>> ...Thereby creating another type of grind, as why would a player bother
>> to do anything but hang out near the stairs, where he knows there are no
>> traps, and wait for the free XP to come to him?
>> :)
>
> ACK! Grinding is like a disease, that keeps mutating, and resurfacing,
> with a vengeance.
Possibly. I think it's more just an example of how hard it is to really
work around grinding without being willing to have a (virtual) critter
standing directly behind the hero whackin' him with a stick every time
he slows his mad dash for the Macguffin.
Players are crafty. I was able to do a lot with Sporkhack because I was
one of the top-end Nethack players, and thus I knew what abuses to look
for -- but even then I made my designs under the theory that even if the
puzzle or challenge looked hard to ME, that some player was out there
who would say "Oh! Oh, of course, you just ___ and ___ and bob's your
uncle."
So far that has proven consistently true, too; some challenges I created
with no visible solution have been worked 'round or otherwise dealt
with. Clearly, I need to quit worrying about solutions except from the
"don't make an easy one by accident" perspective.
David Ploog
> It does seem that certain kinds of improvised grinding cannot be
> eliminated. For example:
>
> You've stumbeled upon a snake pit, without any means of dealing with
> their deadly poisons. You decide to make a run for the stairs, but due
> to a twist of fate, you find a ring of poison resistance on the way.
> At that point, you're inclined to reassess. Your cost/benefit analysis
> has tilted in favor of killing the snakes, which are now easy prey.
I don't think that is grinding: the snakes are a finite resource (of xp
and perhaps something else).
Part of the problem may be that we all use the term "grinding" to mean
something bad, but everyone of us means something different by it.
> What's a developer to do? You never know when the random number
> generator is going to bless (or curse) players with an chance to
> grind. What you _can_ do, is keep the window of opportunity small:
>
> - Don't allow players access to previous levels. Coming back later
> almost always means waiting until the risk is low. So make stairs one
> way only. This will also get rid of another kind of grinding: stash
> management. *shudders*
Yes, you can do that. But there already is Rogue. And allowing to go back
makes for interesting strategy (as opposed to tactics-only when it comes
to combat on the current level).
I found stashing to be very annoying in Nethack where I (needlessly?)
hauled all the items to some altar. In Crawl, stash management is very
efficient because you can (a) look at all items in the dungeon from
anywhere (there is a Find command, even allowing regexes :) and (b) you
can do interlevel travel by a single command. In other words, the
interface suddenly made a grinding aspect less bad.
> - If the player doesn't want to come to danger, let danger come to the
> player. Slaying a horde of monsters can't possibly go unnoticed. Maybe
> after a while critters from the next dungeon level will come take a
> look, to see what happened to their comrades.
This is a good idea: in my opinion, new monsters should spawn but they
should *quickly* become harder the longer the player stays on a level,
thus driving him eventually deeper.
David
Kenneth 'Bessarion' Boyd
> On Aug 14, 12:14 pm, "Kenneth 'Bessarion' Boyd" <zaim...@zaimoni.com>
> wrote:
>
> > On Aug 14, 7:06 am, Andrew Doull <andrewdo...@gmail.com> wrote:
>
> > > Assuming that your random items are generated independently of each
> > > other, that's incorrect. If a random item is generated, which you
> > > never find, it's exactly the same as not being generated at all. If
> > > you choose to do something else which results in you encountering the
> > > same number and distribution of random drops, you're no worse off
> > > ignoring the first n items.
>
> > Not true in general: you're minus the utility of those n items (and
> > gain the corresponding amount of time in not looking for them).
>
> You're minus the utility of those n items either way. Doesn't matter
> if they were generated or not if you never find them. That was
> Andrew's point.
-n items from an unbounded supply: they will be replaced, assuming
independent generation. It's just n different items than the ones you
missed. You're out zero items (which immediately leads to the Too
Much Junk problem). This is what happens in *bands.
What matters is making the finds happen quickly and have a high
plausibility of being useful. (This is actually in-theme for Moria,
Angband's ancestor.)
-n items from a *finite* supply: you're out n items, period. This is
what practically happens in NetHack, Crawl, and POWDER. (Not true for
Crawl if Pan is survivable, and only somewhat false when the Abyss is
survivable [Abyss doesn't generate food, Pan does])
> > ....
>
> Exploring can also be risky. It's quite possible to have a high
> proportion of good items in your game but still frequently choose not
> to pursue them. The endgame in Rogue fits this model. Death lurks
> around every corner, so it's not worth the risk to seek out the
> (finite, generally useful) items.
Yes. The risk of getting those items, directly reduces their
usefulness when considering acquiring them.
> > > An analogous statement to what you're saying is you're less likely to
> > > roll a six on a dice, if the previous dice roll was also a six.
>
> > Only with non-persistent levels.
>
> Nope. If item generations are independent events, failing to find a
> given item is precisely equivalent to it never having been generated
> at all.
The whole set of items generated is *not* independent of missing/
finding items (even when abusing the probability-theory sense with a
pseudorandom number generator), when the total supply whole-game is
finite.
And that's what actually matters when analyzing game balance. You
can't predict what this forest looks like, just from understanding a
single tree to death.
Numeron
The reason for this is the loose definition of griding and that being
repedative in any way for some reward is a bad thing. Which is
entirely silly anyway, since looking at the broad picture of what
roguelikes, and infact basically all CRPGs, amount to is as follows:
1)Kill X monsters and Explore X area in order to gain XP/Items enough
to get final item/kill Foozle
2)Retrieve Item/kill Foozle
3)(Optional) Kill X more monsters en route to escape
That means that 95% of the game can be defined as a grind. Whether you
are grinding your way to be strong enough for the games completion by
going downward to the final goal, or occasionally sideways because
diving doesnt yield fast enough benefits for its increased difficulty,
is irrelevant.
Robert
hunting. Also, I have once or twice cleared Zot 4 to make it a safer
place to rest when anticipating the need to stair scum heavily when
working Zot 5, such as when attempting it with teleportitis.
Pender
On Aug 15, 1:26 am, "Kenneth 'Bessarion' Boyd" <zaim...@zaimoni.com>
wrote:
> -n items from a *finite* supply: you're out n items, period. This is
> what practically happens in NetHack, Crawl, and POWDER. (Not true for
> Crawl if Pan is survivable, and only somewhat false when the Abyss is
> survivable [Abyss doesn't generate food, Pan does])
If you don't find N items that were generated, you're out N items
relative to the world in which you DID find the N items that were
generated. But you've lost nothing relative to the world in which the
N items were never generated in the first place (and in which,
consequently, the finite supply was N items smaller).
> Yes. The risk of getting those items, directly reduces their
> usefulness when considering acquiring them.
No, the items are just as useful. Their usefulness has not decreased.
The cost of acquiring them has merely increased. It is an
*alternative* to Too Much Junk with divergent gameplay implications,
not another instance of it.
> The whole set of items generated is *not* independent of missing/
> finding items (even when abusing the probability-theory sense with a
> pseudorandom number generator), when the total supply whole-game is
> finite.
This "abuse of the probability-theory sense with a pseudorandum number
generator" is implicit in the assumption of independent item
generation. It's true but unremarkable that if you change the
assumption, you can change the conclusion.
But if you accept the stated assumptions, then I'm not sure what
you're arguing. The whole set of items FOUND is not affected at all by
extra unfound items. To the extent you're claiming that the size of
the finite item pool stays the same, then yes, failing to find more of
those items means fewer total items are found. But again, that
violates the independent generation assumption and isn't, I think, a
terribly interesting observation.
> And that's what actually matters when analyzing game balance. You
> can't predict what this forest looks like, just from understanding a
> single tree to death.
It's admittedly a very narrow debate, but nevertheless Andrew was
correct given his stated assumptions.
Pender
wrote:
> The reason for this is the loose definition of griding and that being
> repedative in any way for some reward is a bad thing. Which is
> entirely silly anyway, since looking at the broad picture of what
> roguelikes, and infact basically all CRPGs, amount to is as follows:
>
> 1)Kill X monsters and Explore X area in order to gain XP/Items enough
> to get final item/kill Foozle
> 2)Retrieve Item/kill Foozle
> 3)(Optional) Kill X more monsters en route to escape
>
> That means that 95% of the game can be defined as a grind. Whether you
> are grinding your way to be strong enough for the games completion by
> going downward to the final goal, or occasionally sideways because
> diving doesnt yield fast enough benefits for its increased difficulty,
> is irrelevant.
I think this is quite right and well said. "Grind" cannot be defined
usefully without relying on some notion of what is fun. Being coerced/
encouraged either by the rules or by the incentives of a game to
engage in boring play is the essence of grind. But every game coerces/
incents you to do SOMETHING. The question is whether it's fun.
That is why I think the essential prescription is as follows:
(1) The game designer should assume that the player will attempt to
play optimally, and
(2) Attempted optimal play should be fun.
Different people have different ideas of what is fun. Some people,
apparently, willingly engage in tasks that I consider boring and
repetitive, which explains the success of games like Angband and WoW.
Some people, apparently, dislike tasks in which the difficulty curve
cannot be adjusted at will to suit the individual player, which
explains the frustration of some with the strictures of a demanding
food clock.
We can debate whether players who willingly engage in certain tasks
are actually enjoying those tasks -- i.e. whether revealed preferences
necessarily reflect a philosophical notion of utility, a debate which
I think is a valid one if we are willing to entertain the concepts of
addiction specifically and irrational behavior generally -- but, at
the end of the day, defining grind will be irreducible past the
question of what kinds of gameplay are fun.
What we can and in my opinion SHOULD conclude, though, is that the
responsibility is on the game designer to make attempted optimal play
be fun, not on the player to play sub-optimally.
Billy Bissette
a7b13d...@d34g2000vbm.googlegroups.com:
> Some people, apparently, dislike tasks in which the difficulty curve
> cannot be adjusted at will to suit the individual player, which
> explains the frustration of some with the strictures of a demanding
> food clock.
Part of the appeal is not being forced to play the game in a way
that you might not like. People have differing ideas of what is fun.
This isn't just a matter of difficulty, either. (Just because someone
doesn't want a food clock doesn't necessarily mean they want the game
easier over all. They might just abhor the mechanic and the specific
impacts that it has upon the game in question. Even if you want the
difficulty, you just might find some mechanics to simply be too
"un-fun" for the challenges that they offer.)
Of course, people have different degrees of fun at different
difficulties as well. Some want a casual experience of achievement,
while others want a battle to the death where you have to scrape
and claw for even a slim chance of victory.
Part is that different people find different things to be of
different difficulties. This is one of the killers of balancing a
game's difficulty curve. Sometimes people might want a bit more
freedom in setting that difficulty curve just so that they don't
get tripped up by some spot too much. And people's opinions
change along with their abilities. Not only might they want
something more lax at first, as they improve they may want
something more restrictive.
I like diving in Angband, to keep a challenge, and don't like
sitting on a dungeon level until I've rendered the next level
harmless. Angband supports both play styles.
I don't compete for extremely low turn count victories, because
I don't like some of the play methods required to achieve them.
Others do, and regularly play in that style. Again, Angband
supports both. If you didn't want another Angband example, then
you can pretend I was talking about Nethack's conducts instead.
> We can debate whether players who willingly engage in certain tasks
> are actually enjoying those tasks -- i.e. whether revealed preferences
> necessarily reflect a philosophical notion of utility, a debate which
> I think is a valid one if we are willing to entertain the concepts of
> addiction specifically and irrational behavior generally -- but, at
> the end of the day, defining grind will be irreducible past the
> question of what kinds of gameplay are fun.
>
> What we can and in my opinion SHOULD conclude, though, is that the
> responsibility is on the game designer to make attempted optimal play
> be fun, not on the player to play sub-optimally.
I'm not entirely sure.
The above is dependent upon the idea that "winning" is the ultimate
"fun" to be found within a game. While many people have that belief,
I'm just not entirely sure that it should be so encouraged by
catering to it.
If a player is not being forced to play a game, then it should be
at least somewhat his responsibility to play it in a manner that he
finds fun. Or at least it shouldn't be entirely the designer's blame
when the player plays it in a manner that they (the player) don't
find fun simply because they believe that manner to be a more
optimal method of play.
There are also two methods of approach to "make attempted optimal
play be fun" to consider. One is to try to make the existing
method more interesting, while the other is to try to remove the
existing method. The latter method sacrifices those who don't
actually find that method to be "un-fun", while the former can
be akin to slapping new coats of paint onto a damaged object.
(For examples of each method, consider a game where the optimal
play is to keep killing the same creature for his first two
levels. An approach in the first method might try to add a few
new enemies of similar experience viability for variety, and to
expand their behaviors. An approach in the second method might
be to make the enemy in question stop giving XP after the first
level is obtained, forcing the player to go after presumably more
dangerous prey to obtain the second level.)
Billy Bissette
news:Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de:
There is a game idea in there, where your dungeon level timer might
be in the form of an unbeatable enemy that chases you. You could
even make it the end boss, with the player continually scrambling
deeper to amass experience and items until he thinks he is ready to
face down his foe.
Berserk had Evil Otto, the unkillable bouncing happy face that
would arrive whenever you spent too long in a room. Clock Tower
had the unkillable Scissorman repeatedly appear regardless of the
player's rate of story progress. Nemesis' appearances in Resident
Evil 3 were story progress-dictated, but the same idea was there of
the end boss chasing the player throughout the game from the
beginning.
Paul Donnelly
> David Ploog <pl...@mi.fu-berlin.de> wrote in
> news:Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de:
>> On Fri, 14 Aug 2009, Cuboidz wrote:
>
>>> - If the player doesn't want to come to danger, let danger come to
>>> the player. Slaying a horde of monsters can't possibly go unnoticed.
>>> Maybe after a while critters from the next dungeon level will come
>>> take a look, to see what happened to their comrades.
>>
>> This is a good idea: in my opinion, new monsters should spawn but they
>> should *quickly* become harder the longer the player stays on a level,
>> thus driving him eventually deeper.
>
> There is a game idea in there, where your dungeon level timer might
> be in the form of an unbeatable enemy that chases you. You could even
> make it the end boss, with the player continually scrambling deeper to
> amass experience and items until he thinks he is ready to face down
> his foe.
We talk about "food clocks" and other ways to force progress so much,
but we never mention plain old clocks. Is that just because they're
obvious? Personally, I think racing to beat a time limit is a lot of
fun, and even derive some satisfaction about taking a playthrough to the
end even though I know I don't have time to win. Plus time limits are
easy to make up a justification for.
> Clock Tower
I don't think I'm sleeping tonight.
Derek Ray
> We talk about "food clocks" and other ways to force progress so much,
> but we never mention plain old clocks. Is that just because they're
> obvious? Personally, I think racing to beat a time limit is a lot of
> fun, and even derive some satisfaction about taking a playthrough to the
> end even though I know I don't have time to win. Plus time limits are
> easy to make up a justification for.
One of the primary appeal factors of any roguelike is that it is _not_ a
twitch game; it is a turn-based game. You will find little support for
enforced real-time time limits here, I suspect.
Paul Donnelly
> On 2009-08-15, Paul Donnelly <paul-d...@sbcglobal.net> wrote:
>> We talk about "food clocks" and other ways to force progress so much,
>> but we never mention plain old clocks. Is that just because they're
>> obvious? Personally, I think racing to beat a time limit is a lot of
>> fun, and even derive some satisfaction about taking a playthrough to the
>> end even though I know I don't have time to win. Plus time limits are
>> easy to make up a justification for.
>
> One of the primary appeal factors of any roguelike is that it is _not_ a
> twitch game; it is a turn-based game. You will find little support for
> enforced real-time time limits here, I suspect.
I was actually thinking of a turn limit, in this context. Real-time
limits in turn-based games don't make much sense to me either.
Derek Ray
Turn limits are certainly possible, but require some explanation as to
why they are not necessarily completely arbitrary. And, there is one
final problem: Guide Dang It syndrome (tvtropes), which Nethack suffers
from to an unbelievable degree.
In short, if your new players can learn both that there a) is a
time limit, and b) what they can do to minimize its effects such that c)
they complete the game reasonably on time, then it is very likely that
your experienced players will shrug off the time limit as meaningless.
Paul Donnelly
> On 2009-08-16, Paul Donnelly <paul-d...@sbcglobal.net> wrote:
>> Derek Ray <de...@moot.its.only.a.spamtrap.org> writes:
>>> On 2009-08-15, Paul Donnelly <paul-d...@sbcglobal.net> wrote:
>>>> We talk about "food clocks" and other ways to force progress so much,
>>>> but we never mention plain old clocks. Is that just because they're
>>>> obvious? Personally, I think racing to beat a time limit is a lot of
>>>> fun, and even derive some satisfaction about taking a playthrough to the
>>>> end even though I know I don't have time to win. Plus time limits are
>>>> easy to make up a justification for.
>>> One of the primary appeal factors of any roguelike is that it is _not_ a
>>> twitch game; it is a turn-based game. You will find little support for
>>> enforced real-time time limits here, I suspect.
>> I was actually thinking of a turn limit, in this context. Real-time
>> limits in turn-based games don't make much sense to me either.
>
> Turn limits are certainly possible, but require some explanation as to
> why they are not necessarily completely arbitrary. And, there is one
> final problem: Guide Dang It syndrome (tvtropes), which Nethack suffers
> from to an unbelievable degree.
An explanation is easy enough. Instead of finding the MacGuffin you have
to defuse the ticking MacGuffin. Or you have to get out of the bomb's
blast radius before it blows, or save a hostage in time. Whatever fits.
> In short, if your new players can learn both that there a) is a
> time limit, and b) what they can do to minimize its effects such that c)
> they complete the game reasonably on time, then it is very likely that
> your experienced players will shrug off the time limit as meaningless.
If the limit's purpose is to compel a minimum rate of forward progress,
I'd say that it's already done its job at that point. I think one
advantage of a time limit (and this includes turn limits in games where
turns = time) is that unlike a food clock, if you're playing at the rate
it's intended to suggest, there's no extra work required on your part to
comply. If we're only out to prevent dilly-dallying, it doesn't even
need to be an especially difficult limit to meet. As long is it prevents
whatever kind of behavior you're interested in preventing.
Billy Bissette
news:87vdkot...@plap.localdomain:
>
> We talk about "food clocks" and other ways to force progress so much,
> but we never mention plain old clocks. Is that just because they're
> obvious? Personally, I think racing to beat a time limit is a lot of
> fun, and even derive some satisfaction about taking a playthrough to
> the end even though I know I don't have time to win. Plus time limits
> are easy to make up a justification for.
Hard time limits have an annoyance how they tend to be a flat
win/lose situation. Regardless of your condition, and often of the
enemy's condition, you just simply lose. Something like a food clock
isn't so binary, as games tend to send you through a starving period
where various deliterious effects are applied (HP loss, random
unconsciousness, or the like.)
If the game is long enough, game over by time limit can be
extremely frustrating.
The player may also be annoyed if they feel that they have no
control over the time limit. With a food clock, you can keep
extending the time limit as long as you have and can keep obtaining
food. With the idea I floated of a pursuing end boss, you might
can see the end boss arrive, and you can still try to escape to the
next level. (Even a bit of randomness to the limit can make a
player feel better, as long as they expect the game to be balanced
towards beating even a tight clock. In that case, they may view
anything extra to the limit as "bonus" time instead of any
reduction of the limit as a penalty.)
The meer presence of a hard time limit can annoy a player, even
if it is so lenient as to not be a factor. Thing is, you don't know
*how* long the game will take you on your first play, or possibly
even after several plays. You cannot judge how important a resource
time will be. Food clocks have an advantage here in that they
divide time management into smaller chunks. You don't think "I have
20,000 turns left, so I can afford to take this detour," you think
"I have five pieces of bread, which will last long enough until I
can find some more food."
None of which of course presents such time limits. And games do
get made with them.
David Ploog
> We talk about "food clocks" and other ways to force progress so much,
> but we never mention plain old clocks. Is that just because they're
> obvious?
Yes. At least Crawl features such clocks. They are only used locally but
let me explain them anyway: There are two dungeon features (bazaars and
labyrinths) which will be announced when you enter the level for the first
time. (One is about coins being counted, the other one about a snort.) The
duration is random and there are hints whether you are close or far away.
(Crawl levels are of size 80x70, so rather big.)
The reasons for making these features times: Labyrinths don't feature any
food and we don't want player to safely stock up on food before entering
the labyrinth. Bazaars feature a number of shops and we want buying what's
there to be an ad-hoc decision.
> Personally, I think racing to beat a time limit is a lot of fun.
I agree :)
David
Ray
> I was actually thinking of a turn limit, in this context. Real-time
> limits in turn-based games don't make much sense to me either.
Right. If the adventure begins with "The Rukhs are massing arms
and levying troops to the East" there's an implicit time limit (as
turn limit); if you take more than $TOO_LONG the invaders will
reach your village and do all the nasty things conquering armies
do.
Bear
Paul Donnelly
> Paul Donnelly <paul-d...@sbcglobal.net> wrote in
> news:87vdkot...@plap.localdomain:
>>
>> We talk about "food clocks" and other ways to force progress so much,
>> but we never mention plain old clocks. Is that just because they're
>> obvious? Personally, I think racing to beat a time limit is a lot of
>> fun, and even derive some satisfaction about taking a playthrough to
>> the end even though I know I don't have time to win. Plus time limits
>> are easy to make up a justification for.
>
> Hard time limits have an annoyance how they tend to be a flat
> win/lose situation. Regardless of your condition, and often of the
> enemy's condition, you just simply lose. Something like a food clock
> isn't so binary, as games tend to send you through a starving period
> where various deliterious effects are applied (HP loss, random
> unconsciousness, or the like.)
That's generally true, but it can be ameliorated. It depends on what
happens when time runs out. Maybe if you're tough and well-equipped it's
still possible to win.
> If the game is long enough, game over by time limit can be
> extremely frustrating.
Yes. Ideally it's at least possible to predict well in advance whether
you're going to make it in time. Partial clocks are another option: a
time limit for each level, or a limit that gets extended upon a goal
completion. An example of the former would be levels that must be
completed before security lockdown completes (it could be made softer if
it were still possible to fight through the door), and the latter might
be progressive destruction of the doomsday device's components, each one
earning you another thousand turns, or whatever.
> The player may also be annoyed if they feel that they have no
> control over the time limit. With a food clock, you can keep
> extending the time limit as long as you have and can keep obtaining
> food.
That's one thing that can't be designed away. Personally, I find food
clocks annoying, so I think there's room for both kinds of clocks.
> With the idea I floated of a pursuing end boss, you might can see the
> end boss arrive, and you can still try to escape to the next level.
> (Even a bit of randomness to the limit can make a player feel better,
> as long as they expect the game to be balanced towards beating even a
> tight clock. In that case, they may view anything extra to the limit
> as "bonus" time instead of any reduction of the limit as a penalty.)
That's a good idea too.
> The meer presence of a hard time limit can annoy a player, even
> if it is so lenient as to not be a factor. Thing is, you don't know
> *how* long the game will take you on your first play, or possibly
> even after several plays. You cannot judge how important a resource
> time will be. Food clocks have an advantage here in that they
> divide time management into smaller chunks. You don't think "I have
> 20,000 turns left, so I can afford to take this detour," you think
> "I have five pieces of bread, which will last long enough until I
> can find some more food."
IMO, that's an advantage of a time limit. You know *exactly* how long
the game will take. You don't know if you'll win.... You can estimate
the number of turns a detour will take and compare it to your limit as
well as you can guess when you'll next find food.
And a P.S. on time-limit representation: even if a game proceeds in
turns, it's probably better to map that to absolute game time. Otherwise
fast characters (including slow characters who have quaffed speed
potions) won't have the expected time advantage over slow ones.
Krice
> Personally, I find food clocks annoying, so I think there's
> room for both kinds of clocks.
Food "clock" is transparent, you don't feel like you are
fighting agaist a time limit.
If your game has a simple clock (time limit) that kills the
player let me know. I wont download that game.
Cuboidz
> Food "clock" is transparent, you don't feel like you are
> fighting agaist a time limit.
> If your game has a simple clock (time limit) that kills the
> player let me know. I wont download that game.
A time limit can be invoked gradually. For example:
Every dungeon level is populated by a limited amount of monsters
appropriate for its depth, but once a player enters the level, the
game would slowly but steadily generate ever more difficult out of
depth monsters - like Davig Ploog suggested.
This has the advantage of not killing the player right away, yet
making a full exploration of the level non-trivial.
You'll know that if you spend too much time pillar dancing a dangerous
foe, that you nevertheless insist on killing, you might miss a few
items lying on the floor, due to an emergency retreat to the next
dungeon level.
Kenneth 'Bessarion' Boyd
> I kind of wonder if this debate is worthwhile, but whatever:
It's useful for terminating mental blocks in other readers.
> On Aug 15, 1:26 am, "Kenneth 'Bessarion' Boyd" <zaim...@zaimoni.com>
> wrote:
>
> > -n items from a *finite* supply: you're out n items, period. This is
> > what practically happens in NetHack, Crawl, and POWDER. (Not true for
> > Crawl if Pan is survivable, and only somewhat false when the Abyss is
> > survivable [Abyss doesn't generate food, Pan does])
>
> If you don't find N items that were generated, you're out N items
> relative to the world in which you DID find the N items that were
> generated. But you've lost nothing relative to the world in which the
> N items were never generated in the first place (and in which,
> consequently, the finite supply was N items smaller).
Which amounts to choosing which finite supply of items. This choice
is nonexistent in *bands -- which is why both Andrew and Derek are
correct, for their respective specializations.
> But if you accept the stated assumptions, then I'm not sure what
> you're arguing. The whole set of items FOUND is not affected at all by
> extra unfound items. To the extent you're claiming that the size of
> the finite item pool stays the same, then yes, failing to find more of
> those items means fewer total items are found.
Which is, assuming that more map is easily explorable, (un?)
intentionally choosing a harder game that is technically suboptimal in
NetHack/Crawl/POWDER. (Of course, technically suboptimal may still be
more fun for some definitions of entertainment.)
> But again, that violates the independent generation assumption ....
No point arguing with a real-time falsifiable, analytically false,
clause like that.
Especially when it's completely irrelevant to the material issue
(actually being able to choose how many items are findable whole-
game). The corresponding choice in *bands is which probability
distribution to use for the items.
J�rgen Lerch
On Mon, 10 Aug 2009 14:49:21 -0500, Paul Donnelly wrote:
> And I wonder how the existence of an automap affects this. Without a
> map, would people still feel the urge to fill it in?
I guess I would, though automapping certainly makes it a lot
easier. I still have those old paper maps from Bard's Tale, or
those from IF games. I want to explore.
Ad Astra!
JuL
--
jyn...@gmx.de / Reality is a crutch for those who can't
J�rgen ,,JuL'' Lerch / cope with fantasy
Gerry Quinn
donn...@sbcglobal.net says...
> We talk about "food clocks" and other ways to force progress so much,
> but we never mention plain old clocks. Is that just because they're
> obvious? Personally, I think racing to beat a time limit is a lot of
> fun, and even derive some satisfaction about taking a playthrough to the
> end even though I know I don't have time to win. Plus time limits are
> easy to make up a justification for.
Larn has such a a clock.
- Gerry Quinn
Gerry Quinn
de...@moot.its.only.a.spamtrap.org says...
> On 2009-08-11, David Ploog <pl...@mi.fu-berlin.de> wrote:
> > On Tue, 11 Aug 2009, Derek Ray wrote:
> > (By the way, it would be trivial to make stair dipping the bad choice:
> > simply assume every other level had a timed, unannounced portal vault with
> > great loot.)
>
> Sure, but then you swing the predictability factor the other way; if
> portal vaults are _that_ common, I avoid touching stairs except in dire
> emergency, and I do what I can to prepare for the rush-to-loot. :D (I
> know a happy medium isn't easy to obtain, but it's still vaguely
> amusing to watch the pendulum swing.)
There are lots of possibilities. For example, some could be linked
with a particular staircase. Or the timer on some could run only when
you are on the level (so you can stair-dip freely but must explore fast
to find the vault). I'm sure there are others, and vaults could be a
mix of different options.
- Gerry Quinn
Gerry Quinn
@i18g2000pro.googlegroups.com>, andre...@gmail.com says...
> You should always design your game so the amount of time you spend
> deeper in the dungeon always generates higher rewards than earlier
> in the dungeon. So the incentive is to dive, leaving levels
> incomplete,
> as much as possible.
>
> Level completionism is grinding by another name...
Maybe playing - or learning - any game is grinding by another name?
What I mean is, let's not be too purist about these issues. All games
involve some repetition. The alternative to repetition is insanity.
To learn a game you have to be in similar situations a lot and find out
what sort of actions give what sort of outcomes. Is this grinding? In
an abstract way it could be described as such.
Playing satisfaction comes in large part from learning how to execute
the game's tasks well... if some of the tasks are (say) mapping a
level, or ably pillar-dancing, they will be enjoble, so long as they
are not repeated to death.
Maybe level completing should be made harder, but a bonus given for it
sometimes (someone has asked you to make a map for a reward, say, or
find six lost items)? Heh, nice quest idea anyway.
- Gerry Quinn
Gerry Quinn
de...@moot.its.only.a.spamtrap.org says...
> On 2009-08-13, Andrew Doull <andre...@gmail.com> wrote:
> >
> > Level completionism should be strongly discouraged by the game design.
>
> That's easy to say, much less easy to do.
Not that hard if you make it a focus of your game design... you could
have infinitely levels with dangerous monster density increasing
linearly with time, for example. It just depends on how much you want
to focus on it, at the cost of other design considerations.
- Gerry Quinn
Gerry Quinn
@z28g2000vbl.googlegroups.com>, pende...@gmail.com says...
> On Aug 13, 1:10 pm, David Ploog <pl...@mi.fu-berlin.de> wrote:
>
> > As I said, your conclusion is spot on, to me. To me, grindless also means
> > these:
> > * get rid of innate healing -- it leads to grinding on the miniscule level
> > (optimal play: rest after every fight)
> > * no ascension kits -- make sure that winning characters must look
> > different (this achieves that players cannot work down a laundry list of
> > things to get/do)
> > * all resources must be finite -- in particular, persistent levels, no
> > unlimited death drops
> > * have at least one clock that drives the game (this need not be food:
> > I've seen action game [1] where the level would slowly flood with water,
> > keeping you running up and up)
>
> All of these can be accomplished with a real food clock. None of the
> rest are necessary.
>
> 1) Innate healing takes time, which consumes food. If there's only so
> much food, it's no longer optimal to rest after every fight. A real
> food clock means HP is only a semi-renewable resource.
Or if you want resting to be quick, you could make health regeneration
consume a lot of food per unit of health recovered.
> Definitely agree with David Damerell that eliminating grind falls to
> the game designer, not the player. It's always baffled me that Angband
> and Nethack both shamelessly broke the food clock when they split off
> from Rogue -- and to the extent that the food clock is the only thing
> preventing grind, getting rid of it had predictable effects.
Rogue's food clock was overtuned, maybe that was part of it. (At least
MSDOS Rogue, maybe others are different.) Ring of Slow Digestion was
the nearest thing to a Rogue ascension kit.
> If the gameplay remains interesting, why is vertical exploration
> intrinsically preferable to horizontal exploration? And if the levels
> are perfectly safe for the player, then vertical exploration is just
> as mindless and grindy as horizontal exploration. Level completionism
> is only a symptom of the problem in Angband, not the problem itself.
It comes from another desihgn consideration, the need to balance the
number of items found (or experience gained)Qagainst the difficulty of
winning.
- Gerry Quinn
Gerry Quinn
bai...@coastalnet.com says...
> Pender <pende...@gmail.com> wrote in
> news:97da10d3-0d1f-44b7...@z28g2000vbl.googlegroups.com:
> > On Aug 13, 1:10 pm, David Ploog <pl...@mi.fu-berlin.de> wrote:
> >
> >> As I said, your conclusion is spot on, to me. To me, grindless also
> >> means these:
> >> * get rid of innate healing -- it leads to grinding on the miniscule
> >> level
> >> (optimal play: rest after every fight)
>
> healing through effectively self-replinishing resources. Anything
> which allows "optimal" play where you spend your game time recovering
> rather than advancing the game.
>
> If you remove innate healing but retain innate magic point recovery
> and have healing spells, then you've only really changed the specific
> thing that you sit around recovering. Same for if you give the player
> the ability to pray every thousand turns to heal.
>
> Even things like vampiric weapons/attacks could theoretically lead
> to degenerate gameplay where you regularly spend time topping off
> your character on weaklings rather than facing serious threats. Or
> an extremely limited amount of healing items could cause players to
> try to farm weak creatures that can produce them.
Fixable by making ALL healing, magical or otherwise, consume a certain
amount of food per health point recovered.
- Gerry Quinn