Meta discussion of optimal and sub-optimal

[dowan]

We talk so much here about optimal play vs sub-optimal play, but where do we draw the lines?

For example, it's pretty well accepted that some races are easier than others. But we generally don't say "Playing as anything except (insert easiest race here) is sub-optimal.", instead we tend to look at race as a difficulty setting, in a sense.
Backgrounds are the same thing, generally.
However, for some reason this changes when we talk about gods. People will call chei or qazlal sub-optimal choices, rather than looking at them as yet another difficulty modifier. Is this because they're chosen in the course of a game, rather than before starting it?

And then everything after that is subject to optimal/sub-optimal labels. Skilling, killdudes methodology, branch order, spell choices, weapon choices, etc.

Yet, if a player is trying to win the game, clearly there are optimal race and class choices. If a player is trying to win the game as a specific race, there are optimal classes. If a player wants to win with a weird specific build, there are still optimal ways to do so.

I guess my question is, why do we draw the line where we do, and does it make sense?

[Sar]

I guess it might make sense to look at it from score/streak playing perspective. Competitive players like that will often take weaker races and backgrounds to either contest the highscore for those or to add an unique combo to a streak. In that sense, their choice isn't optimal or suboptimal. However, god choice happens in the game and doesn't influence score directly.

[dowan]

I see your point. So it's really about the score db. Winning with a character who had firestorm castable in lair doesn't get you any extra points or a unique entry in your score file, while winning a MuSk does. A CK who abandons Xom on turn 1 is counted just the same as a CK who stuck with Xom all game.

Well, that's a very good explanation for why we do it the way we do. But, does that really make sense, when you consider most people coming here for advice probably aren't even playing online, and if they are, they're probably not trying to set high scores?

[Sar]

That's why most people are advised against playing weak races, combos and gods.

[dowan]

True, but if someone says "I am trying to win as a spellcasting high elf who doesn't use filthy weapons" they're usually told they're playing wrong. People who say "I am trying to win a MuMo because I want to be a greatermummy usually aren't told such things.

At least, that's how it seems to me, although I don't have any examples to back it up. Basically it seems that all tavern advice is given from the implied position that the player cares about their online score page, rather than playing in some particular specific way.

Obviously that's not the same as someone saying "I want to win as easy as possible so I'm only playing DEWz and only training spellcasting because that's the easiest way to win, but I keep losing?!? What do?" Obviously in that case you'd tell them that what they think is the easiest way to win actually is not.

[Sar]

I think most of the time when people do things like completely avoiding melee on casters it's because they have those preconceptions about how magic characters should play. These characters don't play like that in Crawl. I don't think I've ever seen somebody starting a thread in DCA with "hey I really want an advice on this char except don't advise me X and Y because I don't want to do those under any circumstances".

[dowan]

Well in DCA you'll often see threads where someone is repeatedly advised to train a weapon skill, and they'll keep saying they don't want to use weapons, they would have played a fighter if they wanted to use weapons. But nobody ever says "Well, if you insist on not using weapons, which makes the game harder than it would otherwise be, here's the best way to go about it". Instead they'll usually say "Well then enjoy losing because that's not how crawl is played".

But people don't say that to the guy trying to win the MuMo, they try to find the best way to win in the constraints of the terrible character that is a MuMo. Is that because a tourney might give extra points for winning a nem's choice MuMo, but there will never be extra points for not training any weapon skills as a DEWz?

[Berder]

Maybe it's because they want to actually be a mage instead of always a fighter and don't want to get pushed into the fighter role!

When you start a new RPG you have never played before and are presented with character selection, if you choose a mage presumably it's because you want to roleplay as an actual mage. If you wanted to hit things with a stick you would have chosen fighter. Crawl is failing these new players by not letting them roleplay the way they'd like.

[dynast]

I choose to believe there is no optimal gameplay, only choice and consequence. I dont like the idea of optimal e sub-optimal, im fine with "hey, dont read/quaff id consumables you dont have stacks of", im not fine with "hey, dont wake up this unique because he is kinda of a threat, now go and explore the rest of the floor with this map limitation and probly meet this then awoken unique 5 turns later when he has the upper hand".

[dowan]

Maybe it's because they want to actually be a mage instead of always a fighter and don't want to get pushed into the fighter role!

When you start a new RPG you have never played before and are presented with character selection, if you choose a mage presumably it's because you want to roleplay as an actual mage. If you wanted to hit things with a stick you would have chosen fighter. Crawl is failing these new players by not letting them roleplay the way they'd like.

Well maybe it's the tavern failing. Crawl doesn't literally force you to use a weapon, even if using a weapon makes the game easier, just like it doesn't force you to select a strong race, even if using one makes the game easier. The tavern forces you to use a weapon if you want advice, but it doesn't force you to choose an easy race.

[dynast]

When you start a new RPG you have never played before and are presented with character selection, if you choose a mage presumably it's because you want to roleplay as an actual mage. If you wanted to hit things with a stick you would have chosen fighter. Crawl is failing these new players by not letting them roleplay the way they'd like.

Then you play pillars of eternity, create a wizard and the game tells your main attribute is strength.

[byrel]

I think this is really a good point. Optimal is only meaningful with a set of assumptions. For instance, in the caustic shrike thread, the op was playing an OpWz 'suboptimally' (ie by not training melee, and by roleplaying a brave character.) It IS meaningful to ask for advice on optimal play given those assumptions (well, you'd have to define 'brave' play better than he did. But still...) It's completely valid to say 'I want to kill all the popcorn with spells. How do I do that optimally?' By the same token, I think a lot of people who ask questions like 'I'm playing a blasty mage DECj of Veh, but I keep dying! Help!' are really trying to figure out the optimal way to win a DECj. Which probably means advice about easier gods, training defenses and melee, and adapting to weapon drops is welcome.

It's hard to know what assumption someone's coming in with, but we should probably do a better job of catering to people who have 'I want to play suboptimally in this one way, but optimally otherwise' questions.

[dowan]

Maybe a separate forum for advice like that would help, like "Advanced exercises in sub-optimal silliness". Because Aesos is a cool acronym anyway...

[tedric]

congratulations, this thread has arrived at the inevitable conclusion that the meaning of words is context-sensitive: http://www.iep.utm.edu/wittgens/#H5

[Wahaha]

By default optimal means best way to win. If there is context then optimal is related to that context. It's really not hard to understand this.

[all before]

People give themselves arbitrary challenges or restrictions all the time to make the game more difficult or to give it variety. Speedrunning is a version of this, as is greaterplayer since it makes people play weaker backgrounds and classes. I don't see people here trying to dissuade players from doing that or not giving them advice for how to win under these conditions. What I do see is new players who don't understand that backgrounds are designed as starting kits or that melee/spells are not designed as binary options. Those players really do need the advice that crawl is a game about doing whatever you can to survive and not about imposing false notions of a "role" on a game that has no storytelling element.

[Sar]

Then you play pillars of eternity, create a wizard and the game tells your main attribute is strength.

It's Might, and it's one of main attributes, but you really don't want to dump Int either.

[byrel]

Yeah, It's kind of silly all the people who bitch about 'strength powering wizards' when it's not even called strength. It's an interesting design choice to be sure, but it's not as silly as they're trying to make it out to be.

[byrel]

congratulations, this thread has arrived at the inevitable conclusion that the meaning of words is context-sensitive: http://www.iep.utm.edu/wittgens/#H5

Like... that's the first time I've seen anyone on the crawl forums link to a source so arcane as to be harder to understand than onget. While speaking perfect English. Who is this Witty character anyhow?

[dynast]

Yeah, It's kind of silly all the people who bitch about 'strength powering wizards' when it's not even called strength. It's an interesting design choice to be sure, but it's not as silly as they're trying to make it out to be.

Its not silly if you dont have a berserker that can cast stronger spells than your wizard using scrolls.

[byrel]

Yeah, It's kind of silly all the people who bitch about 'strength powering wizards' when it's not even called strength. It's an interesting design choice to be sure, but it's not as silly as they're trying to make it out to be.

Its not silly if you dont have a berserker that can cast stronger spells than your wizard using scrolls.

The silly part is using the word 'strength' which the games doesn't use, in order to make the game sound more ridiculous. Disagreeing with their stat system is fine; using the wrong word to describe it is duplicitous and silly.

[dowan]

congratulations, this thread has arrived at the inevitable conclusion that the meaning of words is context-sensitive: http://www.iep.utm.edu/wittgens/#H5

This thread has more to do with the contexts in which the crawl community is willing to discuss the concepts represented by those words than the meaning of the words themselves. Nobody talks about the optimal way to play a weaponless spellcaster, despite the fact that many players want to play that way. Hell, it's rare you see people talk about how best to play a transmuter of chei, and at least a few people seem to enjoy that playstyle.

[dowan]

People give themselves arbitrary challenges or restrictions all the time to make the game more difficult or to give it variety. Speedrunning is a version of this, as is greaterplayer since it makes people play weaker backgrounds and classes. I don't see people here trying to dissuade players from doing that or not giving them advice for how to win under these conditions. What I do see is new players who don't understand that backgrounds are designed as starting kits or that melee/spells are not designed as binary options. Those players really do need the advice that crawl is a game about doing whatever you can to survive and not about imposing false notions of a "role" on a game that has no storytelling element.

Yeah, there's an important distinction there. I'm not saying you just should go along with people's mistaken ideas when giving them advice.

[ydeve]

Just wanted to point out that sub-optimal playstyles aren't necessarily hard. For example, it's actually quite easy to play a weaponless DECj. New players are quite capable of winning suboptimal characters if you give them the chance . Tell them it's suboptimal, but if they want to do it, they can try <this, etc>.

[dynast]

The silly part is using the word 'strength' which the games doesn't use, in order to make the game sound more ridiculous. Disagreeing with their stat system is fine; using the wrong word to describe it is duplicitous and silly.

I forgot it was Might ok? I stopped playing that game a long time ago. Also, nice job putting words inside my mouth, i had to go back to see where i called it silly to find out i didnt.

[Wahaha]

Nobody talks about the optimal way to play a weaponless spellcaster, despite the fact that many players want to play that way. Hell, it's rare you see people talk about how best to play a transmuter of chei, and at least a few people seem to enjoy that playstyle.

Because no one asks "hello, how can I win a weaponless spellcaster?". If someone started a thread with this question there would be plenty of real answers and a few posts strongly bashing the idea. The latter are inevitable and somewhat necessary. Either the OP learns that wizards should be melee characters or says "I don't care" and the thread continues.

[byrel]

Also, nice job putting words inside my mouth, i had to go back to see where i called it silly to find out i didnt.

Wait whut? I never said you called it silly. I did say you 'were making it out to be silly', which is my interpretation of

Then you play pillars of eternity, create a wizard and the game tells your main attribute is strength.

If I misread you, I'm sorry. That was obvious to me, but intent can be a bit hard to read over the internet.

In any event, I think we've derailed this thread enough...

[dynast]

My point was that Might in that game breaks some of its roleplaying factor.

[Sar]

You can treat it like a setting quirk.

[dowan]

You've all ruined my thread forever and I'm taking my ball and going home. Damn kids and their pillars of various intangible concepts...

[onget]

That the optimal play is present, It is bad.

It is optimal to always focus on the early game, There are no exceptions.
Besides there are absolutely optimal choice.

It is always a force the same choise, it is boring.

[Sar]

In pretty much any game if there are as much as two options, one is gonna be optimal.

[dynast]

In pretty much any game if there are as much as two options, one is gonna be optimal.

But how do you know which one is the optimal one? Its like exploring on civilization... *proceeds to derail the thread again*

[Arrhythmia]

congratulations, this thread has arrived at the inevitable conclusion that the meaning of words is context-sensitive: http://www.iep.utm.edu/wittgens/#H5

how does wittgenstein keep coming up in this forum. wtf

[Igxfl]

There is a particular playstyle, or narrow set of playstyles, which is considered optimal.

An optimal choice is any choice that moves the character towards the optimal playstyle.

Note that using the optimal playstyle is not the same as maximizing winrate; there exist choices outside of the optimal playstyle that increase one's chances of winning.

These choices are called "optimal play" and disdained.

[bel]

Well in DCA you'll often see threads where someone is repeatedly advised to train a weapon skill, and they'll keep saying they don't want to use weapons, they would have played a fighter if they wanted to use weapons. But nobody ever says "Well, if you insist on not using weapons, which makes the game harder than it would otherwise be, here's the best way to go about it". Instead they'll usually say "Well then enjoy losing because that's not how crawl is played".

But people don't say that to the guy trying to win the MuMo, they try to find the best way to win in the constraints of the terrible character that is a MuMo. Is that because a tourney might give extra points for winning a nem's choice MuMo, but there will never be extra points for not training any weapon skills as a DEWz?

You need not play optimally to win. You can win the game without hitting things with a stick at all. One of my first wins was a DEFE of Vehumet, who basically went the whole game without hitting things with weapons, or using them very rarely. It's perfectly possible, though a bit painful.

Presumably if one comes to DCA, it is to get some ideas on how other people play the game. Most learning is by imitation. Also, if you totally cut off one option because of roleplay or whatever, it is plausible to assume that you will make things harder on yourself.

(random story: I once played a game where I met a guy who always played atheist characters because he had some weird ideas about conflicting with real-life religion. I did ask him whether his religion approved of him murdering everything in sight, but didn't get a response. But he managed to play atheist characters fine, that's the point.)

[Sandman25]

Nobody talks about the optimal way to play a weaponless spellcaster, despite the fact that many players want to play that way. Hell, it's rare you see people talk about how best to play a transmuter of chei, and at least a few people seem to enjoy that playstyle.

Because no one asks "hello, how can I win a weaponless spellcaster?". If someone started a thread with this question there would be plenty of real answers and a few posts strongly bashing the idea. The latter are inevitable and somewhat necessary. Either the OP learns that wizards should be melee characters or says "I don't care" and the thread continues.

I don't see "plenty of real answers" in viewtopic.php?f=12&t=12409&p=177421
I didn't ask any advices though, it is expected that player does not need basic advices when trying to do something as hard/silly as this.

[tedric]

congratulations, this thread has arrived at the inevitable conclusion that the meaning of words is context-sensitive: http://www.iep.utm.edu/wittgens/#H5

This thread has more to do with the contexts in which the crawl community is willing to discuss the concepts represented by those words than the meaning of the words themselves. Nobody talks about the optimal way to play a weaponless spellcaster, despite the fact that many players want to play that way. Hell, it's rare you see people talk about how best to play a transmuter of chei, and at least a few people seem to enjoy that playstyle.

Wittgenstein's insight is that words do not represent concepts; instead, the meaning of a word lies in its use, which consists of a pattern of expectations learned by experience and shared among a community. When Alice uses the word 'optimal' in a way that clashes with Bob's expectations about how that word is generally used, they are bound to misunderstand each other. Similarly, if it is your observation that "nobody talks about the optimal way to play a weaponless spellcaster", the reasonable conclusion is that within this community weaponless spellcasters are not considered optimal. In fact, I would conclude that for the Tavern community, the phrase 'optimal weaponless spellcaster' is a meaningless, nonsensical oxymoron.

Asking, as you did in your OP, "why do we draw the line where we do?" is a fool's errand. You are bound to get as many different answers as Tavern posters, each of whom will have invented her own (essentially superstitious) explanation based on her own unique life history and set of experiences. But that's OK, because it does not matter whether we agree on why the line exists -- only that we all agree (more or less) on where it is. And you can pretty clearly tell where the line is by observing our collective behavior.

[dowan]

Well that's silly. If I decide I'm going to play a weaponless spellcasting character, within the confines of that I can play a variety of different ways. I could choose to train nothing but spellcasting throughout the whole game. Or I could more evenly train defenses and various magic skills. One of those is clearly more optimal than the other. Whether the concept of the original character is optimal doesn't matter, just as it doesn't matter that playing dungeon crawl is a suboptimal use of your time vs whatever else you could be doing when we talk about how best to play it. There's still room to discuss the best way to do this thing, even if this thing isn't the best way to win at crawl, get a job, or make millions of dollars.

Collecting the various different answers is my goal. Just because they aren't all the same answer doesn't mean it's pointless to ask.

[Lasty]

I haven't really seen the described scenario of "person asks for advice, making it clear that they have chosen to forsake weapons even though they know it's bad for their chances of winning, and then only gets advice to pick up a weapon."

Very rarely I've seen "person asks for advice, making it clear that they have chosen to forsake weapons even though they know it's bad for their chances of winning, and then gets advice about how to succeed despite forsaking weapons".

Very, very often I see "person asks for advice, is told to pick up a weapon, says that they're a caster so they don't want to, gets told it's better for chances for winning, and then conversation devolves into bickering about whether or not weapons increase one's odds of winning".

[dowan]

Eh, maybe I'm coloring things through my own recollection, but either way, we both seem to agree it's very rare you'll see advice on how best to succeed when forsaking weapons, or doing other weird stuff that's clearly sub-optimal for winning. I guess you could call it voluntary challenges. On the other hand, on the wiki it's nothing but strategies for winning with really weird sub-optimal strategies. I'd just like to see the really skilled players ideas on how best to play within certain constraints. I know there are people here that enjoy playing in clearly sub-optimal ways, like statue nagas of chei and so forth, but you don't see much in the way of discussions on how to do it.

On the other hand, if you look at a Tome4 or nethack forum, there's all sorts of discussions of weird and wild strategies. So I supposed I was wondering why we don't really have much of that here.

Again, I'm not saying people shouldn't tell the new guy that 'mages' are allowed to use weapons and train defensive skills like everyone else, or correct them if they claim some wildly sub-optimal strategy is in fact optimal for winning.

[tedric]

We talk so much here about optimal play vs sub-optimal play, but where do we draw the lines?

...

why do we draw the line where we do, and does it make sense?

if someone says "I am trying to win as a spellcasting high elf who doesn't use filthy weapons" they're usually told they're playing wrong. People who say "I am trying to win a MuMo because I want to be a greatermummy usually aren't told such things.

At least, that's how it seems to me, although I don't have any examples to back it up. Basically it seems that all tavern advice is given from the implied position that the player cares about their online score page, rather than playing in some particular specific way.

If I decide I'm going to play a weaponless spellcasting character, within the confines of that I can play a variety of different ways. I could choose to train nothing but spellcasting throughout the whole game. Or I could more evenly train defenses and various magic skills. One of those is clearly more optimal than the other. Whether the concept of the original character is optimal doesn't matter, just as it doesn't matter that playing dungeon crawl is a suboptimal use of your time vs whatever else you could be doing when we talk about how best to play it. There's still room to discuss the best way to do this thing, even if this thing isn't the best way to win at crawl, get a job, or make millions of dollars.

Collecting the various different answers is my goal. Just because they aren't all the same answer doesn't mean it's pointless to ask.

I'd just like to see the really skilled players ideas on how best to play within certain constraints. I know there are people here that enjoy playing in clearly sub-optimal ways, like statue nagas of chei and so forth, but you don't see much in the way of discussions on how to do it.

I think you have moved your own goalposts.

Your original question was about how and why Tavern defines what's "optimal," and I've just pointed out that asking us to talk about it is going to be less productive than reflecting on the observations that led you to ask the question in the first place.

Somewhere along the way, you raised the perceived issue of people wanting helpful advice on playing the game with self-imposed conditions but not receiving it -- and on that, I side with the several posters above who disagree with your perception that that's a common issue worthy of our collective self-reflection.

Now you seem to be asking "Given X self-imposed condition, what is the best way to play the game, for all possible values of X?" which is a literally infinite inquiry. Which is not to say it isn't interesting! Nothing wrong with exploring these on a case-by-case basis. But if that's your goal, you're no longer having a "meta-discussion of optimal and sub-optimal" and instead having a series of discussions on the practicalities of getting by with various specific handicaps. And you should probably just, you know, start one or two of those threads and see how it goes.

[dowan]

Yeah, well, as I'm sure you've noticed by now I'm not always that great at getting my point across, or even knowing what my point is from time to time. So it shouldn't surprise you that the discussion has moved a bit from the OP. It'd be rather silly of me to stick with my original assertion if it was correctly shown to be flawed. It would also be silly to keep making new threads every time the conversation changes direction. I do realize we can't discuss all the crazy ways one could play the game in a single thread. And as you say, we may have reached the end of the useful life of this thread, and it may be time to just start a new thread about such an example.

I like to fully explore ideas, even when they lead far from their starting point. I like to try to gauge the community's collective viewpoint. I like to see when two people who thought they agreed on something actually have very different ideas when you dig into the details. I like to make assertions based on my current understanding, and have Duvessa tell me when I'm completely off my rocker.

Aside from that, a very important discussion about Columns of A Pretty Long Time occurred, so even if the rest of the thread was worthless, we got that. Plus I apparently inspired Berder to make a GDD thread about magic melee spells, so that's something.

[Aule]

You're just not asking the right question.

The meta discussion should be about the meta game.

You see, there is the Crawl game, which is a turn-based RPG with a really nifty tiles version that some people like to play, and there is the ##crawl game, which is the meta game that is all about the win rates, statistics, greatplayer, greaterplayer, and like-minded camaraderie related to the Crawl game as reported via !lg. The two games (call them rpg and meta) are distinct and separate in reality.

The problem (if it is one) is that most of the development of crawl is taking place in that meta game atmosphere, which as a rule usually won't acknowledge the rpg aspect of the game itself, and on those rare occasions where it is mentioned, it is done so dismissively in favor of the meta game. This is unfortunate, in my opinion, because it is insular and creates a self-sustaining feedback loop, which all programmers, whether they be clever or not, should know is detrimental in the end.

And then for what has all your hard work been?