Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Lessons from the original Rogue

1,223 views
Skip to first unread message

pende...@gmail.com

unread,
Jan 21, 2009, 2:13:43 PM1/21/09
to
After playing other roguelikes (okay, just Angband and Nethack) and
being drawn inevitably back to what I see as the superior experience
of the original Rogue, I felt compelled to jot down my thoughts about
what makes that original game so great, and what makes the other two
fall short.

The difficulty:
I'm on a Mac, so the only version I have been able to find that runs
natively is a version of Rogue that has wands of confusion, scrolls of
protect armor, and no dark rooms. I'm not sure how people commonly
refer to this version, but it is not quite the same as the "smiley
face" Epyx DOS version. It's somewhat easier. Having played probably
thousands of games, I've actually won twice -- something I was never
able to do in decades of DOS Rogue back on the Pentium 286 in the
basement.

Still, only two wins. I like that. Each is really meaningful when it
happens. The insane difficulty also means that as a player, you have
to adjust your expectations. You can't view winning as the sole
yardstick by which to judge a character's worth, since it is so rare,
even, I imagine, for highly skilled players. The score -- which in
Rogue is the amount of gold you accumulate -- becomes meaningful. How
many other games can truly claim that? There is no way to accumulate
arbitrarily high scores, either; the only way to find more gold is to
dive deeper, and the difficulty increases exponentially as you do.

The urgency:
No character, no matter how powerful, can ever feel comfortable. The
food system is a meaningful constraint that forces you to delve ever
deeper and precludes any attempt to "stat grind," and no character can
face down every monster it encounters. Even the few characters that
become powerful enough to defeat a Dragon (the most fearsome of
dungeon-dwellers) in hand-to-hand combat will balk at the prospect of
two dragons, or a dragon and a Jabberwock in quick succession. And at
those levels, there will frequently be another dragon or Jabberwock
around the next corner, waiting for a weakened player to stumble into
it.

Every character will get into situations that require a wild card.
Sometimes that means high-tailing it for the staircase. Sometimes it
means dropping a scroll of Scare Monster, firing a wand of slow
monster, or quaffing a potion of extra healing. Sometimes it means
fleeing through a loop in the level long enough for hitpoint
regeneration to tip the balance.

But none of these methods is completely reliable. Fleeing for the
stairs or around a loop entails the possibility of running into
another monster while the first is at your back, leaving you in a
worse situation than before. Slow monster is only a trump card if you
have room to kite backwards as you attack on alternating turns, and
even then is useless if the player is confused by a Medusa's gaze.
Even Scare Monster can be overcome by the Dragon, who can breathe fire
onto or across the protected square -- or, if deployed in an open
space, by the invisible Phantom which can lurk unseen until the player
steps off of the scroll.

Most of these wild cards also require the use of a non-renewable
resource -- an item or a charge on a wand. There being no shops in the
game and no vertical progression except ever downwards, wild card
items are found only serendipitously, and will run out if used more
frequently than they are found.

And since overwhelming encounters become more frequent as one delves
deeper, a huge part of the game's strategy entails the careful
management of this precious resource of escape items. No other video
game that I have ever played has an experience as intense as choosing
whether to risk another attack with a weakened character or parachute
to safety with an item that, when used, is forever gone.

In any case, the incredible difficulty means that every character will
be tested to its limit and usually beyond. Really good and
sufficiently patient players can usually coast through a game of
Angband or Nethack; as long as they check all the boxes and proceed
carefully, they can be relatively confident that every obstacle of the
depths can be overcome. Not so in Rogue; since no one can reliably
win, there is no point past which additional skill becomes
superfluous.

Part of the reason Rogue can remain fun with such a punishing
difficulty curve is its shortness; no game will last more than an hour
or so at the most. Dying is a temporary setback. In Nethack, or
especially Angband, a single death can mean days or even weeks of lost
effort, so developers have compromised by making it all but completely
avoidable for relatively skilled players.

The simplicity:
Aside from movement commands, the entire command set of the game is
short, fairly intuitive, and requires only a single keystroke. There
are no macros (though the built-in fight and repeat commands could
perhaps be seen as hard-wired macros) and there are no commands that
require any modifier key besides Shift.

There is also no need to consult a spoiler file. With one exception
(the correct use of the Scare Monster scroll), the capabilities of
every item and monster in the game quickly become apparent with some
experimentation. Nethack is notorious for requiring consultation with
spoiler files to play well (who can remember all the alchemy
combinations, or the probabilities of various events when you kick a
sink?), but Angband isn't that much better; certain items are found
only at certain depths, and some (like stat gains and resistances) are
all but required to proceed past other depths. Proceeding in ignorance
of these various thresholds means the nearly-certain death of a weeks-
old character. Life is just not long enough to figure all of these
things out by experimentation. While Rogue does have a few quirks
(throwing potions at monsters will sometimes inflict a status effect
on the monster, and throwing a spent wand at a monster will sometimes
wring one last use out of it), these are not essential.

There are also no resistances or for that matter elemental damage.
Angband and Nethack at a certain point devolve into a period of
checking the boxes, where each of a number of resistances must be
attained before a character can adventure further. Angband is
notorious for this, including even "high-level resists" like nether,
nexus, sound, and something called "shards." Both games all but
require a period of stat gain, and Nethack even includes an explicit
collection game that is required to complete the main quest. (Seven
candles?) In Rogue, you need three things for a winning game: an
enchanted two-handed sword, good enchanted armor, and a big pile of
various escape items.


Rogue shortcomings:
This is not to say that the game is perfect. There are a few changes
that would probably improve the game -- in fact, I'm using this list
as a recipe for a roguelike that I'm coding.

Changes that would offer a more complex game without making the game
harder to understand:
-The levels are boring. A 3x3 grid, each of which may or may not be a
room, connected by tunnels, was exciting back in the early UNIX days
but is boring now. We can do better without making the game harder to
understand. In fact, we could make it easier to understand, since an
essential skill in Rogue is understanding the general workings of the
level creation algorithm so you can figure out where secret doors
probably are at a glance. Better level design would eliminate the
necessity of this not-very-fun skill.
-Monsters could be more diverse. While twelve of the twenty-six
monsters have interesting features, only one can attack from a
distance, none can potentially become an ally, none can summon another
monster, and none has any effect on or interaction with any other
monster except to get in its way. More can be done here without making
the game counterintuitive.

Changes that would reduce the capriciousness of the game without
reducing the difficulty:
-Some items should appear in every game. In particular, if you never
find a two-handed sword, your character is doomed no matter your
skill. In my game, no weapon will be generated for a second time until
every other weapon has been generated once. Maybe I'll think of a
better way to do this, but if a single item is essential to progress
beyond a certain point, every character should be able to find it
sometime before that point.
-Party rooms are problematic. As implemented, any room can be randomly
chosen as a party room, which is filled to the brim with monsters and
items. This means that you can descend a flight of stairs into the
middle of a party room, or even start the game in the middle of a
party room. Both are obviously unfair and should be eliminated. More
controversially, perhaps party rooms should not be so easy to stumble
across without warning. I'll need to give this one some more thought.

Changes that would give the game more atmosphere:
-Dungeon generation in Rogue does change slightly as one goes deeper,
but the only change is to make rooms more likely to be mazes -- which
aren't even particularly enjoyable to navigate. Adding different
terrain types, doing interesting things with lighting, and tweaking
the dungeon generation algorithm in subtle ways as one goes deeper
would add a lot to the atmosphere of the dungeon.


Anyway, those are my thoughts. Rogue is a seriously overlooked game
considering that the entire genre is named after it, and I think there
is a lot of potential for roguelikes that are closer to Rogue in terms
of simplicity than to Angband or Nethack.

msa...@gmail.com

unread,
Jan 21, 2009, 2:37:09 PM1/21/09
to
You should spend some time with Crawl if you haven't already. In my
opinion, it captures the feel of rogue a little bit better than the
hacks and bands do, while modernizing it significantly. It does
require spoilers, and I think your points about simplicity are well
made. But it is an example of a roguelike that drives you deeper with
food demands, does not demand stair-scumming (or any scumming for that
matter), and doesn't have you returning to a town over and over.

Its still gameable, and certain points in the game do violate a lot of
the rogue flavor (random shops, no food demands after the hive,
stashes, etc). But it might give you some inspiration too.

Stefan O'Rear

unread,
Jan 21, 2009, 2:49:01 PM1/21/09
to
On 2009-01-21, msa...@gmail.com <msa...@gmail.com> wrote:
> You should spend some time with Crawl if you haven't already. In my
> opinion, it captures the feel of rogue a little bit better than the
> hacks and bands do, while modernizing it significantly. It does
> require spoilers, and I think your points about simplicity are well
> made. But it is an example of a roguelike that drives you deeper with
> food demands, does not demand stair-scumming (or any scumming for that
> matter), and doesn't have you returning to a town over and over.

Winnable without spoilers is an explicit design goal of Crawl - would it
be possible to narrow down the "requirement" to the point where it is
bug-reportable?

msa...@gmail.com

unread,
Jan 21, 2009, 3:00:50 PM1/21/09
to
On Jan 21, 2:49 pm, Stefan O'Rear <stefa...@cox.net> wrote:
>
> Winnable without spoilers is an explicit design goal of Crawl - would it
> be possible to narrow down the "requirement" to the point where it is
> bug-reportable?

Fair enough. I rescind the term 'requires'. But I do claim that its
pretty tricky to play without them. There are a number of areas where
I think spoilers are almost required (that spell that turns corpses
into potions? I needed a spoiler to figure that out). But I would
accept that its winnable without them.

pende...@gmail.com

unread,
Jan 21, 2009, 3:33:41 PM1/21/09
to
On Jan 21, 2:37 pm, msal...@gmail.com wrote:
> You should spend some time with Crawl if you haven't already.

I'd like to. As far as I can tell, no version exists for Mac OS X, and
I don't care to install Linux or Windows or for that matter build a
machine-dependent front-end for it when I'm not even sure what success
would look like.

Is there such a version out there?

For what it's worth, since I know Crawl makes use of at least classes,
I wish I had added a section about classes and stats. Rogue has no
classes and the only stats -- HP and STR -- are the same every time
you start a new game. It's a wise move, I think; classes mean you have
several games bound together that all need to be tested and balanced
separately even though they generally share the same monster and item
database, and rolling for stats just encourages players to scum for
good stats, which is highly un-fun and strictly inferior to starting
out with the same stats every time.

I think a better method to classes is to have a variety of items, only
a subset of which will be found in any given game, and let the classes
evolve from there. If a character finds tons of scrolls of enchant
armor and weapon, their playing style evolves to be like a fighter. If
a character finds lots of wands and scrolls but not as many weapons
and armor, they will play more like a mage. And if, as I am
contemplating in my game, a character finds items to boost his
persuasiveness and lots of gold and food, that character will play
more like a "diplomat" or bard, charming and bribing monsters and
having them do most of the heavy fighting for him. The part of this
design I don't feel completely comfortable with yet is how to make
sure that every character does not do all of these things more or less
equally. If a single powerful item opens one of these paths, finding
three of the items makes the character three times as powerful --
which is too capricious. And if each class requires the concerted
effort of a number of specific items, then you'd end up with some
characters who never draw a complete hand. I don't want to secretly
choose the player's class and generate only those items intentionally,
since that's cheating and will be transparent to the player. I am
planning to try to solve this dilemma with items that must be
"committed" to one purpose or another -- perhaps a generic "enchant
item" scroll which can be used to strengthen an armor or weapon as
usual, but can alternatively be used to recharge a wand or boost the
potency of a ring.

David Ploog

unread,
Jan 21, 2009, 3:40:32 PM1/21/09
to
On Wed, 21 Jan 2009, Stefan O'Rear wrote:
> On 2009-01-21, msa...@gmail.com <msa...@gmail.com> wrote:

>> You should spend some time with Crawl if you haven't already.

>> It does require spoilers


>
> Winnable without spoilers is an explicit design goal of Crawl - would it
> be possible to narrow down the "requirement" to the point where it is
> bug-reportable?

Yes. Actually, the game _is_ winnable without spoilers (lorimer of
Sporkhack did it, for example).

Pointing out spoiler information is always welcome. There is a line
between under- and overinformation. We're reluctant to hand out lots of
numbers but we are quick to give qualitative information.

An example: in trunk, monster descriptions show certain intrinsics
(resistances, susceptibilities, speed) but not health or attack damage.

David

msa...@gmail.com

unread,
Jan 21, 2009, 3:43:22 PM1/21/09
to
telnet crawl.akrasiac.org

From a terminal. You may have to fiddle with your terminal settings by
control-clicky on the terminal and selection 'window settings' to get
the color and commands working :(

Chephren

unread,
Jan 21, 2009, 4:12:20 PM1/21/09
to
On Jan 21, 7:13 pm, penderpr...@gmail.com wrote:
> After playing other roguelikes (okay, just Angband and Nethack) and
> being drawn inevitably back to what I see as the superior experience
> of the original Rogue, I felt compelled to jot down my thoughts about
> what makes that original game so great, and what makes the other two
> fall short.
>

Having a played a fair bit of rogue, I agree with many of your
observations. I've mainly played a port of the original unix version,
which I got off the roguelike restoration project. Word of warning, I
actually recommend the 5.2 release, its pretty similiar to the 3.6 one
as far as I recall - the main difference is that the spacebar isn't
used for both clearing messages and skipping turns (thats a lot of
pain). 5.4 actually changes a bit of stuff, like having separate
identity scrolls for armour, weapons, scrolls etc.

The main advantage I feel that rogue has over its modern descendants
is actually balance. Its the only roguelike I know that not only
never lets up, but actually gets progressively more challenging as it
goes on. I stopped playing crawl a while back because my dwarven
warrior (to be fair one of the easiest combos) reached a point where
nothing was really threatening him so long as he was careful, and all
I had to look forward to was an extended tedious middlegame. I
couldn't be bothered to continue playing that character, but couldn't
just give up (I want to ascend!); so i'm still waiting to reload that
save. It was disappointing because I'm a terrible roguelike player,
and a complete crawl noob. I've had the same experience with nethack
as well. After overcoming the interesting early game, there just
seems to be a lot of mindless bumping of monsters, and I get bored.

The other nice thing about rogue, as you rightly pointed out is the
simplicity. Not only does it make it accessible, and easily
comprehensible, but it makes all the the elements distinctive. Wands,
potions, and scrolls don't have any overlapping effects, there is
actually a meaningful delineation between the categories.

As far as I can tell the early unix version is pretty similar to the
epyx version, except some of the names are different, and the unix one
is all in black and white. You mentioned the epyx version doesn't
have the dark rooms as you descend, which sounds like it would make
the game much easier. Its sort of cool the way you have increasingly
more dark rooms the further you descend, as though you really are
delving into the dark heart of the world. At the same time it makes
ranged weapons useless, because you never spot a monster until you
bump into him.

I've had ideas for a purified rogue, that maintains the balance and
simplicity of the original. Changes I would like to see are:
Some discernible difference in damage between shooting an arrow and
throwing one (and a convenient way to switch between bow and melee
weapon)
A lighting system that maintains the angst of descending into
darkness, without wrecking ranged combat, and the possibility of any
tactical maneuvering.
Maybe a movement system that takes speed into account.

Oh yeah, I can't believe I never knew you had to drop the scroll of
scare monster, and now i've read it twice in one day! :-)

Stefan O'Rear

unread,
Jan 21, 2009, 4:31:42 PM1/21/09
to
On 2009-01-21, pende...@gmail.com <pende...@gmail.com> wrote:
> On Jan 21, 2:37 pm, msal...@gmail.com wrote:
>> You should spend some time with Crawl if you haven't already.
>
> I'd like to. As far as I can tell, no version exists for Mac OS X, and
> I don't care to install Linux or Windows or for that matter build a
> machine-dependent front-end for it when I'm not even sure what success
> would look like.
>
> Is there such a version out there?

You're presumably looking at the official Dungeon Crawl, which was
abandoned by the developers around 2003. All modern development,
including a fully functional OS X port, occurs on the Stone Soup
branch.

Radomir Dopieralski

unread,
Jan 21, 2009, 4:47:32 PM1/21/09
to
At Wed, 21 Jan 2009 11:13:43 -0800 (PST), pende...@gmail.com wrote:

> Aside from movement commands, the entire command set of the game is
> short, fairly intuitive, and requires only a single keystroke. There
> are no macros (though the built-in fight and repeat commands could
> perhaps be seen as hard-wired macros) and there are no commands that
> require any modifier key besides Shift.

I don't agree here. There is a number of special cases and non-intuitve
item uses that let you survive a little easier: mostly by adding new
escape and recovery tacticts, but also by improving your stats faster.
You still don't need a spoiler file, as these can be learned by just
normally playing (sooner or later you will discover them randomly, as
they don't require doing anything particularly silly).


> -Monsters could be more diverse. While twelve of the twenty-six
> monsters have interesting features, only one can attack from a
> distance,

I definitely think one is enough :)

> none can potentially become an ally,

There is a way to make monsters fight each other.

> none can summon another
> monster,

Are you a masohist? Besides, that would allow grinding.

> and none has any effect on or interaction with any other
> monster except to get in its way.

Each monster that has a status-affecting attack interacts with others.
In particular, my approach to Ice Monster, Floating Eye or Venus Flytrap
is completely different depending on whether there are monsters around
or not (although I admit this changed when they added hypothermia).

> More can be done here without making
> the game counterintuitive.

I think it's difficult enough already, thank you.

> -Party rooms are problematic. As implemented, any room can be randomly
> chosen as a party room, which is filled to the brim with monsters and
> items. This means that you can descend a flight of stairs into the
> middle of a party room, or even start the game in the middle of a
> party room. Both are obviously unfair and should be eliminated. More
> controversially, perhaps party rooms should not be so easy to stumble
> across without warning. I'll need to give this one some more thought.

Well, if you die just at the beginnig, then it's a small lose -- you just
start over. On the other hand, if you somehow manage to escape such
a room, you get a huge boost of adrenaline and ego :)

> Anyway, those are my thoughts. Rogue is a seriously overlooked game
> considering that the entire genre is named after it, and I think there
> is a lot of potential for roguelikes that are closer to Rogue in terms
> of simplicity than to Angband or Nethack.

I completely agree here, it's a marvelous game and I think there should
be more roguelikes like this. Actually, that's probably why I like the
console roguelike-likes so much -- they are definitely based on Rogue more
than on any other game.

I have three versions of rogue installed (3.6, 5.4 and DOS) and play them
regularly. It's really great fun.

--
Radomir Dopieralski, http://sheep.art.pl

pende...@gmail.com

unread,
Jan 21, 2009, 5:53:36 PM1/21/09
to
On Jan 21, 4:47 pm, Radomir Dopieralski <n...@sheep.art.pl> wrote:

> I don't agree here. There is a number of special cases and non-intuitve
> item uses that let you survive a little easier: mostly by adding new
> escape and recovery tacticts, but also by improving your stats faster.

Not sure what you're talking about -- do you mean quaffing healing
potions when you're at max health to boost your HP by 1-2 points? If
so, it's a pretty marginal difference... but fair enough. The scare
monster scrolls and the healing potions are both counterintuitive in
that sense.

> > -Monsters could be more diverse. While twelve of the twenty-six
> > monsters have interesting features, only one can attack from a
> > distance,
>
> I definitely think one is enough :)

If "attack from a distance" has to mean FIRE BREATH OF DOOM then yes,
one is enough -- but an enemy that shoots arrows at you and tries to
keep a minimum distance would be a very interesting addition to the
game.

> There is a way to make monsters fight each other.

This is news to me. What's the secret? There are no rings of conflict
or anything like that in my version.

> > none can summon another monster,
>
> Are you a masohist? Besides, that would allow grinding.

I'm planning to prevent summoned monsters from giving experience or
dropping items.

> > and none has any effect on or interaction with any other
> > monster except to get in its way.
>
> Each monster that has a status-affecting attack interacts with others.
> In particular, my approach to Ice Monster, Floating Eye or Venus Flytrap
> is completely different depending on whether there are monsters around
> or not (although I admit this changed when they added hypothermia).

That's not the monsters interacting with each other; that's the
monsters mutually interacting with you. What if orcs and centaurs
instinctively hated one another and would fight each other? What if
Flytraps would trap and attack any creature that came close? What if
an Orcish Tamer would boost the morale and damage of nearby animals?
More potential for strategy.

> > More can be done here without making
> > the game counterintuitive.
>
> I think it's difficult enough already, thank you.

I didn't say more difficult. If you changed Centaurs into ranged arrow-
shooting attackers, that would increase the strategic complexity of
the game without necessarily making it more or less difficult; its
effect on the difficulty would depend entirely on what stats you gave
the new Centaur.

> > -Party rooms are problematic. As implemented, any room can be randomly
> > chosen as a party room, which is filled to the brim with monsters and
> > items. This means that you can descend a flight of stairs into the
> > middle of a party room, or even start the game in the middle of a
> > party room. Both are obviously unfair and should be eliminated. More
> > controversially, perhaps party rooms should not be so easy to stumble
> > across without warning. I'll need to give this one some more thought.
>
> Well, if you die just at the beginnig, then it's a small lose -- you just
> start over. On the other hand, if you somehow manage to escape such
> a room, you get a huge boost of adrenaline and ego :)

Yes, but I think no game should be obviously lost before the first
move. It's also just unfair to have a great character descend to level
26 into a room full of Dragons and Jabberwocks. Unless he has a scroll
of teleport (and having one is not strictly a matter of preparedness,
since an awful lot of luck figures into the equation), he's instantly
dead with no questions asked. No other item in the entire game will
save him, and the only blunder he made was walking down a flight of
stairs, which is mandatory.

There's no way to make sure that EVERY character death is "fair," per
se, since luck is a big part of the game, but we should definitely
eliminate specific unfair deaths when it is so easy to do so.

pende...@gmail.com

unread,
Jan 21, 2009, 5:55:04 PM1/21/09
to
On Jan 21, 4:31 pm, Stefan O'Rear <stefa...@cox.net> wrote:

Oh, wonderful! Yes, I see that you're right. Thanks for the tip; I'll
check it out tonight.

Gelatinous Mutant Coconut

unread,
Jan 21, 2009, 6:12:59 PM1/21/09
to
> > none can summon another
> > monster,
>
> Are you a masohist? Besides, that would allow grinding.

No arguments on the masochism front, but summoned monsters don't need
to be implemented in a way that allows grinding. For example, a demon
summoning cultist monster capable of summoning one, and only one,
demon. The risks and rewards are precisely constrained: if the cultist
succeeds in summoning the demon, the player has a much more difficult
fight, but may potentially get XP and loot from both the cultist and
the demon, while if the player is quick and kills the cultist before
the demon is summoned, they get an easier fight, but XP and loot from
only the cultist. You end up getting an interesting tactical decision
without giving the player the opportunity to endlessly grind.

Another idea would be to go with a monster that actually summoned
other monsters, instead of creating monsters and saying that they have
been summoned with flavor text. How about a monster that, when it
first encounters the player, teleports the strongest monster still
alive on a dungeon level nearer to the start (Sidestepping the
question of whether those are 'higher' or 'lower' levels) to its side?
There isn't any opportunity to grind, because the player isn't
introduced to any additional monsters that they would not have been
able to encounter by going back towards the beginning of the dungeon.
Players can then either try to kill that monster in a new environment
now that they are presumably a bit stronger, or decide that the
encounter is still out of their league and head back to an earlier
part of the dungeon until they feel that they are strong enough. If
they do decide to go back, they might be able to access new items or
dungeon regions that were previously guarded by the presence of the
summoned monster, giving them the opportunity to power up in
preparation for when they do want to take the monster on. (Or, they
could try to sneak past the summoned monster and its summoner,
possibly to have it pop up again later in the game the next time they
encounter a monster summoner.)

Radomir Dopieralski

unread,
Jan 21, 2009, 6:46:39 PM1/21/09
to
At Wed, 21 Jan 2009 14:53:36 -0800 (PST), pende...@gmail.com wrote:

> On Jan 21, 4:47 pm, Radomir Dopieralski <n...@sheep.art.pl> wrote:
>
>> I don't agree here. There is a number of special cases and non-intuitve
>> item uses that let you survive a little easier: mostly by adding new
>> escape and recovery tacticts, but also by improving your stats faster.
>
> Not sure what you're talking about -- do you mean quaffing healing
> potions when you're at max health to boost your HP by 1-2 points? If
> so, it's a pretty marginal difference... but fair enough. The scare
> monster scrolls and the healing potions are both counterintuitive in
> that sense.

<rot13>
Fnzr tbrf sbe gur cbgvba bs ertnva fgeratgu. Gurer ner nyfb fbzr rssrpgf
bs gur cbgvba bs urnyvat gung ner abg vzzrqvngryl boivbhf, urnyvat rssrpg
bs gur cbgvba bs cbvfba (fvp!) naq zber :)
</rot13>

>> > -Monsters could be more diverse. While twelve of the twenty-six
>> > monsters have interesting features, only one can attack from a
>> > distance,
>>
>> I definitely think one is enough :)
>
> If "attack from a distance" has to mean FIRE BREATH OF DOOM then yes,
> one is enough -- but an enemy that shoots arrows at you and tries to
> keep a minimum distance would be a very interesting addition to the
> game.
>
>> There is a way to make monsters fight each other.
>
> This is news to me. What's the secret? There are no rings of conflict
> or anything like that in my version.

Try throwing a potion of confusion at a group of monsters in a corridor.

>> > none can summon another monster,
>>
>> Are you a masohist? Besides, that would allow grinding.
>
> I'm planning to prevent summoned monsters from giving experience or
> dropping items.

By the way, I forgot to mention one more cool idea from Izuna in the other
thread: a relatively weak monster (frog) that drops an egg when killed.
This egg won't hatch until attacked, but the tadpole that is created is
at least 3 levels stronger than any frog. You can kill it at that level
after some preparations, but if you make it hatch by mistake while
fighting with other monsters then you are practically dead.

>> > and none has any effect on or interaction with any other
>> > monster except to get in its way.
>>
>> Each monster that has a status-affecting attack interacts with others.
>> In particular, my approach to Ice Monster, Floating Eye or Venus Flytrap
>> is completely different depending on whether there are monsters around
>> or not (although I admit this changed when they added hypothermia).
>
> That's not the monsters interacting with each other; that's the
> monsters mutually interacting with you. What if orcs and centaurs
> instinctively hated one another and would fight each other? What if
> Flytraps would trap and attack any creature that came close? What if
> an Orcish Tamer would boost the morale and damage of nearby animals?
> More potential for strategy.

Somehow I don't like those ideas, but I can't really put a finger to it.
It's similar with the more shooting monsters above. Maybe it's because
then it makes the game too dependant on the exact positions of monsters
and the combinations of monsters generated, which you simply can't
control -- it seems too random.

b0rsuk

unread,
Jan 22, 2009, 4:15:29 AM1/22/09
to
On 21 Sty, 21:33, penderpr...@gmail.com wrote:

> For what it's worth, since I know Crawl makes use of at least classes,
> I wish I had added a section about classes and stats. Rogue has no
> classes and the only stats -- HP and STR -- are the same every time
> you start a new game. It's a wise move, I think; classes mean you have
> several games bound together that all need to be tested and balanced
> separately

(...)


> I think a better method to classes is to have a variety of items, only

(...)

Crawl has no real classes. Class merely defines your starting items,
initial stats and - in some cases - initial god. The god part is not
such a big deal - Ecumenical Temple can be found quite early and it
contains almost all altars so you can start worshipping any god as
soon as you find the Temple, sometimes earlier. Spellbooks are
trickier - some are rather uncommon (Geomancy, and the one with
transmigrations).
I'm more and more convinced that naming starting kits 'class' was a
mistake and and has many negative effects on players. I keep running
into people like you who, upon hearing the word 'class' jump into
conclusions about Crawl, the way it is played, the way it should be
played. Crawl has no classes except Lugonu and Beogh.
Several alternatives have been proposed, like 'proffesion' or
'background'.

By the way, I'm going to mention that Crawl on several occasions
compromises its design goals. Crawl isn't meant to support infinite
play and grinding, yet Pandemonium and the Abyss exist. One of recent
proposals on crawl-ref discuss mailing list was to disable 'Detect
Items' spell in the Abyss - this can be done, but searching for the
rune will become much more boring and grindy, especially with Abyss
staying the way it is. Even if Abyss is finally changed (getting
harder as the time passes) searching for the rune is going to involve
repeated attempts, 'diving in'.

Crawl is supposed to be anti-spoiler, but one of fairly recent
features - vaults designed by hand which occasionally appear in
dungeons - highly reward spoiled players. Players who remember vault
layout know what monsters to expect, where to find loot and secret
passages. There's not enough randomisation in custom vaults. One of
examples I can think of is the square box made of secret doors. An
inexperienced player may continue to open the doors and search for
more, while another one will know the box contains stone giants,
shadow dragons, titans etc. This box can appear as early as D:13 (in
one of my games) or in Vaults branch (which is not that bad).
Experienced player who's not well prepared will run away at the first
sight of the box, and leave the level as fast as possible.

b0rsuk

Stefan O'Rear

unread,
Jan 22, 2009, 5:33:09 AM1/22/09
to
On 2009-01-22, b0rsuk <jaz...@gmail.com> wrote:
> Crawl is supposed to be anti-spoiler, but one of fairly recent
> features - vaults designed by hand which occasionally appear in
> dungeons - highly reward spoiled players. Players who remember vault
> layout know what monsters to expect, where to find loot and secret
> passages. There's not enough randomisation in custom vaults. One of
> examples I can think of is the square box made of secret doors. An
> inexperienced player may continue to open the doors and search for
> more, while another one will know the box contains stone giants,
> shadow dragons, titans etc. This box can appear as early as D:13 (in
> one of my games) or in Vaults branch (which is not that bad).
> Experienced player who's not well prepared will run away at the first
> sight of the box, and leave the level as fast as possible.

Recent? That vault is at least 10 years old. I have a copy of it in
maps.cc from a DC 3.0.2 tarball, the file being timestamped 1998-12-23.

b0rsuk

unread,
Jan 22, 2009, 6:16:14 AM1/22/09
to
On 22 Sty, 11:33, Stefan O'Rear <stefa...@cox.net> wrote:
> Recent?  That vault is at least 10 years old.  I have a copy of it in
> maps.cc from a DC 3.0.2 tarball, the file being timestamped 1998-12-23.

Ok, i didn't know about that particular vault, but my point still
stands. Each new custom vault increases Crawl's spoiler potential.
Vaults are not random enough and an experienced player knows what to
expect. The gap between smart/spoiled and smart/unspoiled widens. Some
powergamers could even refer to Crawl's source code upon encountering
a new vault.
That said, Crawl is still much better in these areas than other
roguelikes I played. I just don't want it to degenerate.

Back to Rogue. I haven't played it, but I like what others said about
it in this thread. In particular, it confirms my gut feeling that a
game (not just roguelike) doesn't need to have any explicit rewards
for the player (like experience levels, or gods forbid badges and
medals). A game needs to be fun and provide interesting challenges.
Ideally progressively harder challenges. Explicit rewards like medals
and goals are used by modern mainstream game developers to hide the
fact that only one in many games is actually very fun to play to the
point people will keep coming back to it. Achievements are used to
glue people to a game.
There's almost no character progression in DooM (except for weapons,
but ammo runs out too and most powerups are temporary). Yet DooM
remains very fun game to play. Contrary to what some people say, DooM
II (one I'm more familiar with) provides very interesting tactical
challenges and situations. Monsters of various types near and far,
combinations of monsters, high variety in terrain. And each level is
winnable on Ultra Violence difficulty level even if you start with NO
weapons (for example with a cheat code). We used to run contests "who
can beat this level first with no previously collected items". Pretty
soon in any level you can find a weapon, and you keep finding more.

b0rsuk

Darren Grey

unread,
Jan 22, 2009, 6:26:26 AM1/22/09
to
On Jan 22, 11:16 am, b0rsuk <jaze...@gmail.com> wrote:

> Explicit rewards like medals
> and goals are used by modern mainstream game developers to hide the
> fact that only one in many games is actually very fun to play to the
> point people will keep coming back to it. Achievements are used to
> glue people to a game.

And yet roguelike players go out of their way to invent all sorts of
new challenges or milestones for each other (your Doom II example fits
perfectly). I think Achievements can be done well if they encourage
the player to try out different playstyles, or increasingly harder
challenges for when they've mastered different levels of play. A
problem many games have is staying difficult after you know the game
inside-out. Encouraging players to aim for higher or different
Achievements is a way of reinventing the game for them and thus
extending how much fun can be had. It's also great as a community-
driven thing.

Of course many commercial games use Achievements for ridiculous
things, like character creation or completing a bog-standard quest.
Or even worse simply kill x of y or do z 100 times. These sort of
simple, monotonous achievements should be avoided. One good way to
use them actually could be for negative things - scumming and so on
would get negative Achievements. Kill too many of one monster, stay
in one area too long, etc - all marks of shame on players that rely
excessively on one thing. But then perhaps that would put players
off...

--
Darren Grey

pau...@mbnet.fi

unread,
Jan 22, 2009, 6:33:18 AM1/22/09
to
On 22 tammi, 13:26, Darren Grey <darrenjohng...@gmail.com> wrote:
> Kill too many of one monster, stay
> in one area too long, etc - all marks of shame on players that rely
> excessively on one thing.  But then perhaps that would put players
> off...

It's best not tell the playes how they should play the game.

b0rsuk

unread,
Jan 22, 2009, 8:18:23 AM1/22/09
to
On 22 Sty, 12:26, Darren Grey <darrenjohng...@gmail.com> wrote:
> On Jan 22, 11:16 am, b0rsuk <jaze...@gmail.com> wrote:

> And yet roguelike players go out of their way to invent all sorts of
> new challenges or milestones for each other (your Doom II example fits
> perfectly).

These are all fully optional and coming from players' initiative. Such
challenges are invented because usually they are fun. If they're not,
people ignore them. If a game has achievement mechanic, it has very
different psychological effect. Player will have that nagging feeling
"I haven't completed the game fully, there's still stuff I'm expected
to do."
I value player initiated challenges much more than built-in
achievement system. "Let's take two engies and build a level3
sentrygun in enemy spawn on Well map!" and so on. This is stuff it's
fun to tell people about. Achievements are too often used as an excuse
to push players towards grind, and the distinction is not clear.


>I think Achievements can be done well if they encourage
> the player to try out different playstyles, or increasingly harder
> challenges for when they've mastered different levels of play.

The same can be done with a text file, a list of suggestions. I don't
mind as long as the game doesn't keep track of what you've
accomplished and what not.

 A
> problem many games have is staying difficult after you know the game
> inside-out.

I don't call that a 'problem'. For me it's a feature I'm looking for
in games. I expect all games to be challenging except those with a lot
of story - then it's optional as long as the story is good.

 Encouraging players to aim for higher or different
> Achievements is a way of reinventing the game for them and thus
> extending how much fun can be had.  It's also great as a community-
> driven thing.

You can do the same by posting a bunch of amazing screenshots or a
wicked youtube video to document your exploits. As a general rule, I
watch interesting gameplay videos as long as they have something new
to show. I wouldn't watch a vid showing a guy killing 10 people with
knife in a match. I would deffinitely NOT watch 10 such vids. Player
initiated contests often end once they're beaten.

b0rsuk

Gerry Quinn

unread,
Jan 22, 2009, 10:24:39 AM1/22/09
to
In article <f4991a0c-3ad0-4caf-8f59-
6d6648...@v18g2000pro.googlegroups.com>, jaz...@gmail.com says...

> On 22 Sty, 11:33, Stefan O'Rear <stefa...@cox.net> wrote:
> > Recent?  That vault is at least 10 years old.  I have a copy of it in
> > maps.cc from a DC 3.0.2 tarball, the file being timestamped 1998-12-23.
>
> Ok, i didn't know about that particular vault, but my point still
> stands. Each new custom vault increases Crawl's spoiler potential.
> Vaults are not random enough and an experienced player knows what to
> expect. The gap between smart/spoiled and smart/unspoiled widens. Some
> powergamers could even refer to Crawl's source code upon encountering
> a new vault.
> That said, Crawl is still much better in these areas than other
> roguelikes I played. I just don't want it to degenerate.

Another thing to consider is that all roguelikes have a large element of
'learning what to expect', and Rogue is no exception. You know what
levels rattlesnakes or aquators are likely to appear, you know that food
will be an issue.

Yes, custom maps are a bit more blatant in this regard, I agree.

> Back to Rogue. I haven't played it, but I like what others said about
> it in this thread. In particular, it confirms my gut feeling that a
> game (not just roguelike) doesn't need to have any explicit rewards
> for the player (like experience levels, or gods forbid badges and
> medals).

Rogue gives levels, and a nice tombstone.

> A game needs to be fun and provide interesting challenges.
> Ideally progressively harder challenges. Explicit rewards like medals
> and goals are used by modern mainstream game developers to hide the
> fact that only one in many games is actually very fun to play to the
> point people will keep coming back to it. Achievements are used to
> glue people to a game.

Are we back on the MMORPG thread? The new achievement system in WoW has
driven players bonkers, for a fact :-)

- Gerry Quinn
--
Lair of the Demon Ape (a coffee-break roguelike)
<http://indigo.ie/~gerryq/lair/lair.htm>

Gerry Quinn

unread,
Jan 22, 2009, 10:39:20 AM1/22/09
to
In article <1993fdd8-3962-457b-baef-39c71080e154
@s9g2000prg.googlegroups.com>, jaz...@gmail.com says...

> On 22 Sty, 12:26, Darren Grey <darrenjohng...@gmail.com> wrote:
> > On Jan 22, 11:16 am, b0rsuk <jaze...@gmail.com> wrote:
>
> > And yet roguelike players go out of their way to invent all sorts of
> > new challenges or milestones for each other (your Doom II example fits
> > perfectly).
>
> These are all fully optional and coming from players' initiative. Such
> challenges are invented because usually they are fun. If they're not,
> people ignore them. If a game has achievement mechanic, it has very
> different psychological effect. Player will have that nagging feeling
> "I haven't completed the game fully, there's still stuff I'm expected
> to do."

I think it's half and half. WoW's system shows the good an bad points
of such a system, as some of the achievements are easy to do and rather
pointless. Whereas some of them provide a focus fot players or guilds
to set themselves a challenge.

Sartharion is the easiest raid boss in the game; killing the 25-man
version with 20 players or fewer is something that experienced raiders
will do as a matter of course, if they happen to be lacking in players
one evening. But I would guess that the achievement is a nice spur to
less confident players. There are also a good number of achievements
related to how you kill certain bosses, some of them whimsical. I think
they add something in the shape of a target that a group can try for.

Another thing related to achievements but not exactly the same is also
present in WoW and might be adapted to roguelikes. The aforementioned
Sartharion is only easy when you kill his three lieutanents before you
kill him. You can leave one, two, or all three alive before you engage
him, making for a more difficult fight; if all three are left, he
becomes the hardest boss in current content. If you don't kill one or
more drakes, you not only get an achievement, but he drops better loot.
A more useful option for a single-player game such as a roguelike, I
think, and worth thinking about for roguelikes.

Gerry Quinn

unread,
Jan 22, 2009, 10:41:10 AM1/22/09
to
In article <a37396b9-97a9-4479-a817-
38e1f1...@p2g2000prf.googlegroups.com>,
kabusho_mus...@hotmail.com says...

> Oh yeah, I can't believe I never knew you had to drop the scroll of
> scare monster, and now i've read it twice in one day! :-)

I never knew it when I played Rogue (MSDOS version). Still beat it
twice, but it would have been useful to know...

Gerry Quinn

unread,
Jan 22, 2009, 10:44:11 AM1/22/09
to
In article <Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de>,
pl...@mi.fu-berlin.de says...

I think handing out all the numbers is a reasonable option too,
especially if the game involves difficult tactical combat. One big
bonus of this is that there is nothing the player has to remember.

Message has been deleted

Gerry Quinn

unread,
Jan 22, 2009, 10:53:51 AM1/22/09
to
In article <f407a70e-e2d1-4913-bb62-88fcad715384
@e1g2000pra.googlegroups.com>, GelatinousM...@gmail.com says...

> No arguments on the masochism front, but summoned monsters don't need
> to be implemented in a way that allows grinding. For example, a demon
> summoning cultist monster capable of summoning one, and only one,
> demon. The risks and rewards are precisely constrained: if the cultist
> succeeds in summoning the demon, the player has a much more difficult
> fight, but may potentially get XP and loot from both the cultist and
> the demon, while if the player is quick and kills the cultist before
> the demon is summoned, they get an easier fight, but XP and loot from
> only the cultist. You end up getting an interesting tactical decision
> without giving the player the opportunity to endlessly grind.

A good way to implement the 'extra loot for more difficulty' mechanic,
if the cultist has little loot but the demon tends to gave something
good.

> Another idea would be to go with a monster that actually summoned
> other monsters, instead of creating monsters and saying that they have
> been summoned with flavor text. How about a monster that, when it
> first encounters the player, teleports the strongest monster still
> alive on a dungeon level nearer to the start (Sidestepping the
> question of whether those are 'higher' or 'lower' levels) to its side?
> There isn't any opportunity to grind, because the player isn't
> introduced to any additional monsters that they would not have been
> able to encounter by going back towards the beginning of the dungeon.
> Players can then either try to kill that monster in a new environment
> now that they are presumably a bit stronger, or decide that the
> encounter is still out of their league and head back to an earlier
> part of the dungeon until they feel that they are strong enough. If
> they do decide to go back, they might be able to access new items or
> dungeon regions that were previously guarded by the presence of the
> summoned monster, giving them the opportunity to power up in
> preparation for when they do want to take the monster on. (Or, they
> could try to sneak past the summoned monster and its summoner,
> possibly to have it pop up again later in the game the next time they
> encounter a monster summoner.)

You could have a named nasty 'monster captain' at the start whom a low
level player can only flee from (make him move slowly, or be slow to
react to seeing the player. But until he's killed, certain monsters
will periodically shout out "He's here, boss!" and summon him...

David Ploog

unread,
Jan 22, 2009, 10:54:22 AM1/22/09
to
On Thu, 22 Jan 2009, Gerry Quinn wrote:
> In article <Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de>,
> pl...@mi.fu-berlin.de says...

>> Actually, the game _is_ winnable without spoilers (lorimer of


>> Sporkhack did it, for example).
>>
>> Pointing out spoiler information is always welcome. There is a line
>> between under- and overinformation. We're reluctant to hand out lots of
>> numbers but we are quick to give qualitative information.
>>
>> An example: in trunk, monster descriptions show certain intrinsics
>> (resistances, susceptibilities, speed) but not health or attack damage.
>
> I think handing out all the numbers is a reasonable option too,
> especially if the game involves difficult tactical combat. One big
> bonus of this is that there is nothing the player has to remember.

It is an option, but the numbers are many and their meaning not obvious.
In other words, we have to weigh between giving useful and full
information for well informed players and lots of confusing garbage for
the rest (not to mention intimidating information for newbies).
Showing lots of numbers also has the drawback of tempting players into
reading all the garbage, slowing everything down.

My own stance is towards the new players. Nothing will beat full
spoilerage in the end, and the developers should not try to explain all
the formulae within the game. To me, it is good enough if either the naive
approach generally works as expected (then little or no additional text is
required); in other cases (e.g. the resistances of monsters) we hand them
out. For judgement of monsters as threats, I prefer colours or one-line
comments to number sets.

As an example: it seems custom to not tell the player exactly how damaged
a monster is (neither Nethack nor Crawl do this). In Crawl, you get a
qualitative hint ("near death" etc.) I am sure that the game would lose
appeal by showing current monster HP. But if we are opaque with this
number (arguably the most important one), why should we print the others?

David

Perdurabo

unread,
Jan 22, 2009, 11:07:28 AM1/22/09
to
On Jan 22, 3:54 pm, David Ploog <pl...@mi.fu-berlin.de> wrote:

<snip>

>
> As an example: it seems custom to not tell the player exactly how damaged
> a monster is (neither Nethack nor Crawl do this). In Crawl, you get a
> qualitative hint ("near death" etc.) I am sure that the game would lose
> appeal by showing current monster HP. But if we are opaque with this
> number (arguably the most important one), why should we print the others?
>
> David

I happen to think Crawl is perhaps a smidgeon *too* vague with
information like this, but however, this could merely demonstrate that
I'm not paranoid enough for Crawl's style of gameplay (my lack of
success in the game would also probably back this up).

A happy, but consistent, medium is probably called for. Defining what
that medium is, on the other hand....

Best,
P.

Gerry Quinn

unread,
Jan 22, 2009, 11:20:32 AM1/22/09
to
In article <428ebf89-8b3c-4309-b4fb-
c8e6ab...@g3g2000pre.googlegroups.com>, pende...@gmail.com says...

> On Jan 21, 4:47 pm, Radomir Dopieralski <n...@sheep.art.pl> wrote:
>
> > I don't agree here. There is a number of special cases and non-intuitve
> > item uses that let you survive a little easier: mostly by adding new
> > escape and recovery tacticts, but also by improving your stats faster.
>
> Not sure what you're talking about -- do you mean quaffing healing
> potions when you're at max health to boost your HP by 1-2 points? If
> so, it's a pretty marginal difference... but fair enough. The scare
> monster scrolls and the healing potions are both counterintuitive in
> that sense.

The scare monster scroll is counterintuitive, but you will probably
discover the secondary effect of healing and strength recovery potions
by randomly quaffing potions for identification when at full health.

- Gerry Quinn

David Damerell

unread,
Jan 22, 2009, 12:26:57 PM1/22/09
to
Quoting <pende...@gmail.com>:
[Rogue versus Angband, NetHack]
>Still, only two wins. I like that. Each is really meaningful when it
>happens.

Try Crawl.

>No character, no matter how powerful, can ever feel comfortable. The
>food system is a meaningful constraint that forces you to delve ever
>deeper and precludes any attempt to "stat grind,"

Try Crawl.

>or so at the most. Dying is a temporary setback. In Nethack, or
>especially Angband, a single death can mean days or even weeks of lost
>effort,

This is simply not true of NetHack. You can win the game in an
all-nighter, and only extinctionists play for weeks.

>experimentation. Nethack is notorious for requiring consultation with
>spoiler files to play well (who can remember all the alchemy
>combinations, or the probabilities of various events when you kick a
>sink?)

... and who needs to? Not that NetHack isn't aided by reference to
spoilers, but a) bad examples and b) you can certainly win without
having to look at spoilers during the game.

>Angband and Nethack at a certain point devolve into a period of
>checking the boxes, where each of a number of resistances must be
>attained before a character can adventure further.

This is technically true of NetHack but a) the number is 1 (magic
resistance) and b) if you don't have it there isn't much to do _but_
adventure further and hope.

>nexus, sound, and something called "shards." Both games all but
>require a period of stat gain

Total nonsense, of NetHack. I've ascended 22 times and never once stopped
to grind my stats up to maximum. They get there, or close enough, just in
the course of normal play.

>and Nethack even includes an explicit collection game that is required
>to complete the main quest. (Seven candles?)

I'm starting to think you haven't actually _played_ NetHack through. Yes,
seven candles. A lot of the time, these can be obtained by the fiendishly
difficult operation of going to the guaranteed lighting shop and saying
"Please sell me these candles, Izchak". If Izchak doesn't have seven -
unlikely but not impossible - the odds against the others not turning up
are very high. But, if they don't, you are guaranteed at least four wishes
(Castle wand is 0:1, no scrolls of charging) each of which can be used to
wish for a full seven candles.

Hardly an intense grind-fest!

NetHack obviously has downsides, but to conflate it with Angband's "recall
to $depth, kill 800 monsters, analyse loot, sell loot, rinse and repeat"
game is just silly.
--
David Damerell <dame...@chiark.greenend.org.uk> flcl?
Today is Saturday, January - a weekend.

pende...@gmail.com

unread,
Jan 22, 2009, 1:00:02 PM1/22/09
to
On Jan 22, 12:26 pm, David Damerell <damer...@chiark.greenend.org.uk>
wrote:

> >Dying is a temporary setback. In Nethack, or
> >especially Angband, a single death can mean days or even weeks of lost
> >effort,
>
> This is simply not true of NetHack. You can win the game in an
> all-nighter, and only extinctionists play for weeks.

"Weeks" was directed at Angband -- and I don't find it particularly
persuasive in the context of this argument that rather than play for
days, one could conceivably compress all of those gameplay hours into
a single 24-hour period by enduring sufficient sleep deprivation.

> >experimentation. Nethack is notorious for requiring consultation with
> >spoiler files to play well (who can remember all the alchemy
> >combinations, or the probabilities of various events when you kick a
> >sink?)
>
> ... and who needs to? Not that NetHack isn't aided by reference to
> spoilers, but a) bad examples and b) you can certainly win without
> having to look at spoilers during the game.

You CAN, it's just substantially harder. And this is about a game
where an "unspoiled ascension" is a notable rarity; it's nearly
impossible to play the game well without first digesting a few FAQs.
In fact, it seems to me like that's almost the point of the game: once
you have amassed an encyclopedic knowledge of how all the pieces work,
it's pretty simple to manipulate them in a way that leads to victory,
so it's predominantly the monastic willingness and capability to
memorize all of this information that separates the great players from
the mediocre players, not skill or intuition.

> >Angband and Nethack at a certain point devolve into a period of
> >checking the boxes, where each of a number of resistances must be
> >attained before a character can adventure further.
>
> This is technically true of NetHack but a) the number is 1 (magic
> resistance) and b) if you don't have it there isn't much to do _but_
> adventure further and hope.

Or use one of your wishes. Or, I don't know, somehow summon a gray
dragon, kill it, wear the scales, and read a scroll of enchant armor.
And what about spell reflection, and elemental resists / cancellation?
Oh, and I think resist poison is essential too if you don't want insta-
death from an invisible trap. Not to mention all the non-essential
boxes that perhaps you don't have to check individually but still
present mindless little rituals that you'd be a fool to forego. Like
eating a floating eye corpse. My point is, the game is full of chores,
even if not quite as many as Angband; Rogue has none.

> >and Nethack even includes an explicit collection game that is required
> >to complete the main quest. (Seven candles?)
>
> I'm starting to think you haven't actually _played_ NetHack through. Yes,
> seven candles. A lot of the time, these can be obtained by the fiendishly
> difficult operation of going to the guaranteed lighting shop and saying
> "Please sell me these candles, Izchak". If Izchak doesn't have seven -
> unlikely but not impossible - the odds against the others not turning up
> are very high. But, if they don't, you are guaranteed at least four wishes
> (Castle wand is 0:1, no scrolls of charging) each of which can be used to
> wish for a full seven candles.

You're proving my point -- this candles thing is a silly collection
chore that involves no risk or interesting gameplay. First, you have
to read a FAQ to find out that you need seven candles in the first
place and then to find out where the shop is, then you trek over there
and buy the candles. And why on earth, from a game designer
perspective, shouldn't at least seven candles be guaranteed to be in
the shop? It probably require one line of code to make sure Izchak
always has at least seven candles; how is the playing experience
enhanced by basically levying a random tax of one wish on one out of
twenty players?

> NetHack obviously has downsides, but to conflate it with Angband's "recall
> to $depth, kill 800 monsters, analyse loot, sell loot, rinse and repeat"
> game is just silly.

Perhaps you're right that I went too far, but the point remains that
Rogue is at one end of the continuum and both NetHack and Angband are
way down the continuum in the other direction -- it just happens that
Angband is even further than NetHack.

David Damerell

unread,
Jan 22, 2009, 1:49:09 PM1/22/09
to
Quoting <pende...@gmail.com>:
>David Damerell <damer...@chiark.greenend.org.uk>

>>>Dying is a temporary setback. In Nethack, or
>>>especially Angband, a single death can mean days or even weeks of lost
>>>effort,
>>This is simply not true of NetHack. You can win the game in an
>>all-nighter, and only extinctionists play for weeks.
>"Weeks" was directed at Angband -- and I don't find it particularly
>persuasive in the context of this argument that rather than play for
>days, one could conceivably compress all of those gameplay hours into
>a single 24-hour period by enduring sufficient sleep deprivation.

I mean, eight hours. I picked an all-nighter because that's how I did it.

>>... and who needs to? Not that NetHack isn't aided by reference to
>>spoilers, but a) bad examples and b) you can certainly win without
>>having to look at spoilers during the game.
>You CAN, it's just substantially harder.

No, it's not. It's slightly harder. I've done it both ways and I actually
know what I'm talking about.

>In fact, it seems to me like that's almost the point of the game: once
>you have amassed an encyclopedic knowledge of how all the pieces work,

You are exaggerating grossly here. I wrote the NetHack Object ID spoiler,
which is about twelve hundred lines of minituae. How much of that do I
actually use?

Bless scrolls of ID. #name object types. Bless/curse ID from altars,
pets. Known objects in Sokoban, Vlad's, etc. ID scrolls are cheapest.
Price-ID spellbooks (100 x level). Unicorn horn dipping in potions.
Engrave-ID wands. Sink-ID rings. Uncursed amulets are harmless. Blessed
spellbooks are harmless. Types of armour, bag, whistle, lamp, etc. Hard
gems are valuable.

Believe me, I can hold all that in my head. It's not exactly a ninja feat
of memorisation.

>>>Angband and Nethack at a certain point devolve into a period of
>>>checking the boxes, where each of a number of resistances must be
>>>attained before a character can adventure further.
>>This is technically true of NetHack but a) the number is 1 (magic
>>resistance) and b) if you don't have it there isn't much to do _but_
>>adventure further and hope.
>Or use one of your wishes. Or, I don't know, somehow summon a gray
>dragon, kill it, wear the scales, and read a scroll of enchant armor.

Thank you, Captain Obvious. Maybe I regard having no MR item and a way of
getting one as equivalent to having MR, inasmuch as the next thing you do
in that situation is to get one?

>And what about spell reflection, and elemental resists / cancellation?

Reflection is not vital, and I don't mean in the _technical_ sense that
MR-less is possible because it has been done at least once by very clever
people, I mean in the sense that reflectionless is a bit of a nuisance.
You're certainly not going to find yourself unable to adventure further
because you lack reflection (albeit that you might pick the direction you
adventure further based on that lack).

Elemental resistances can all be dispensed with save fire resistance in
Gehennom... but to get to Gehennom without eating something appropriate
takes some work. You're not, ordinarily, going to find yourself unable to
adventure further.

Any vaguely sensible cloak choice has MC2 or MC3, you can easily ascend
with MC2; you aren't going to find yourself unable to "adventure further"
for the lack of magic cancellation.

>Oh, and I think resist poison is essential too if you don't want insta-
>death from an invisible trap.

But you are not going to find yourself unable to "adventure further".
Those traps can turn up in so much of the dungeon that there's no sense in
hiding in the rump looking for poison resistance.

>present mindless little rituals that you'd be a fool to forego. Like
>eating a floating eye corpse.

Again, you're not going to stop advancing because you haven't done it.

The idea portrayed - that NetHack has a bunch of resistances like Angband's
"ooh, can't go below 1000' without Free Action" situations - is simply not
true.

[I gather a lot of the hardcore Angbanders regard those limits as pretty
flexible, too, but I'll let them speak to that.]

>>>and Nethack even includes an explicit collection game that is required
>>>to complete the main quest. (Seven candles?)
>>I'm starting to think you haven't actually _played_ NetHack through. Yes,
>>seven candles. A lot of the time, these can be obtained by the fiendishly
>>difficult operation of going to the guaranteed lighting shop and saying
>>"Please sell me these candles, Izchak".

>You're proving my point -- this candles thing is a silly collection
>chore that involves no risk or interesting gameplay.

Except that your point was originally that it's an "explicit collection
game", and that's nonsense. It may involve no risk or interesting
gameplay, but it also involves no effort. You might as well say that Rogue
requiring a two-handed sword to win is an "explicit collection game" - in
both cases, it's "find desired item, pick up desired item". Hardly
onerous; the difference is in NetHack it is absolutely guaranteed that you
will get the desired item, and without spending any time grinding at all.

>place and then to find out where the shop is, then you trek over there
>and buy the candles.

"Find the shop" and "trek over there" being part of "exploring the
dungeon". A terrible thing to have to do in a dungeon exploration game!

>the shop? It probably require one line of code to make sure Izchak
>always has at least seven candles; how is the playing experience
>enhanced by basically levying a random tax of one wish on one out of
>twenty players?

If you'd read the post you were replying to you'd know that's nothing
_like_ one in twenty.

>Perhaps you're right that I went too far, but the point remains that
>Rogue is at one end of the continuum and both NetHack and Angband are
>way down the continuum in the other direction -- it just happens that
>Angband is even further than NetHack.

Except that if you're actually looking at repeated grinding, Angband is
all the way over with World of Warcraft, Rogue has none, and NetHack lets
you do it if you want but doesn't actually require any at all.

rdc

unread,
Jan 22, 2009, 2:43:56 PM1/22/09
to
"David Ploog" wrote :

> As an example: it seems custom to not tell the player exactly how damaged
> a monster is (neither Nethack nor Crawl do this). In Crawl, you get a
> qualitative hint ("near death" etc.) I am sure that the game would lose
> appeal by showing current monster HP. But if we are opaque with this
> number (arguably the most important one), why should we print the others?

Since I didn't know any better, with the monsters in my RL, Deep Deadly
Dungeons, I displayed a health bar above the monsters. I never had any
negative comments on it, and it actually seemed to be a feature that people
liked. It also added quite a bit of tension to battles. The player character
had a health bar as well, and in the heat of the battle, as both bars were
shrinking, it brought an extra dimension to battles that I don't feel in
many RL battles.

Rick Clark

David Ploog

unread,
Jan 22, 2009, 3:05:30 PM1/22/09
to
On Thu, 22 Jan 2009, rdc wrote:

> "David Ploog" wrote :
>> As an example: it seems custom to not tell the player exactly how damaged a
>> monster is (neither Nethack nor Crawl do this). In Crawl, you get a
>> qualitative hint ("near death" etc.) I am sure that the game would lose
>> appeal by showing current monster HP. But if we are opaque with this number
>> (arguably the most important one), why should we print the others?
>
> Since I didn't know any better, with the monsters in my RL, Deep Deadly
> Dungeons, I displayed a health bar above the monsters. I never had any
> negative comments on it, and it actually seemed to be a feature that people
> liked.

Oh, health bars are cool: they convey the information in a manner (I like)
much better than numbers. Incidentally, we added health bars for the
player Health not so long ago. The reaction was generally positive. (We
did not remove the numerical player Health which would probably have
caused riots :)

> It also added quite a bit of tension to battles. The player character
> had a health bar as well, and in the heat of the battle, as both bars
> were shrinking, it brought an extra dimension to battles that I don't
> feel in many RL battles.

Well, in Crawl the monster has a number of health states (lightly
damaged, moderately damaged,...almost dead) which provides some tension,
too.

David

Gerry Quinn

unread,
Jan 22, 2009, 7:26:16 PM1/22/09
to
In article <Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de>,
pl...@mi.fu-berlin.de says...
> On Thu, 22 Jan 2009, Gerry Quinn wrote:
> > In article <Pine.LNX.4.64.09...@lobster.mi.fu-berlin.de>,
> > pl...@mi.fu-berlin.de says...

> >> example: in trunk, monster descriptions show certain intrinsics


> >> (resistances, susceptibilities, speed) but not health or attack damage.
> >
> > I think handing out all the numbers is a reasonable option too,
> > especially if the game involves difficult tactical combat. One big
> > bonus of this is that there is nothing the player has to remember.
>
> It is an option, but the numbers are many and their meaning not obvious.

Depends on the game.

> In other words, we have to weigh between giving useful and full
> information for well informed players and lots of confusing garbage for
> the rest (not to mention intimidating information for newbies).
> Showing lots of numbers also has the drawback of tempting players into
> reading all the garbage, slowing everything down.

Perhaps the solution is to eliminate stats that are irrelevant.

> My own stance is towards the new players. Nothing will beat full
> spoilerage in the end, and the developers should not try to explain all
> the formulae within the game. To me, it is good enough if either the naive
> approach generally works as expected (then little or no additional text is
> required); in other cases (e.g. the resistances of monsters) we hand them
> out. For judgement of monsters as threats, I prefer colours or one-line
> comments to number sets.

Handing out resistances is rather unusual, actually - monster health is
much more commonly handed out.

> As an example: it seems custom to not tell the player exactly how damaged
> a monster is (neither Nethack nor Crawl do this). In Crawl, you get a
> qualitative hint ("near death" etc.) I am sure that the game would lose
> appeal by showing current monster HP. But if we are opaque with this
> number (arguably the most important one), why should we print the others?

If you are opaque with this number, perhaps you shouldn't. But there's
nothing wrong with games that show monster health exactly, either in the
form of a health bar or a number.

r.sh...@gmail.com

unread,
Jan 25, 2009, 5:11:05 PM1/25/09
to
Rogue was my first roguelike, and I still play it from time to time
using the Hexatron java implementation. To me, its one significant
flaw is that games frequently end simply due to lack of good
equipment. It is stingy with items to the point that every use of an
item is significant (this is a good thing) but also stingy to the
point that you just don't have the gear to continue. If you can't
handle trolls by troll depth, you're pretty much screwed, and the only
way I know how to mitigate this is to scum level 12 to whatever extent
food permits.

It's more nerve wracking than its followers (with the possible
exception of raiding Hell in Crawl, my favorite part of that game),
but at the price of a lot of death by RNG.

Radomir Dopieralski

unread,
Jan 25, 2009, 7:52:45 PM1/25/09
to
At Sun, 25 Jan 2009 14:11:05 -0800 (PST), r.sh...@gmail.com wrote:

> It's more nerve wracking than its followers (with the possible
> exception of raiding Hell in Crawl, my favorite part of that game),
> but at the price of a lot of death by RNG.

Personally I think it's a fair price for a roguelike. Ok, it's not
nice being punished for something that's not your fault. But who says
it's a punishment?

I remember from pen & paper RPG sessions that a death of your character
might be much more rewarding and fun than survival, as long as it's in
accordance with the plot and your character's... well, character. Although
"computer RPGs" are not really RPGs, I think it's fair to say that as long
as your death follows your expectations, it's justified. It could be even
jarring to *not* die when you are supposed to!

0 new messages