Display monster AC as range

[jwoodward48ss]

a. Disputable.
b. There are too many options. And options that change the game difficulty are IIRC frowned upon.

[HardboiledGargoyle]

they need no further context, they're self-explained. That's very good feedback.

it is qualitative feedback that needs further context, and a strike against pips and such, whereas numbers like 1d20 are self-explained

Note how simply printing a bunch of numbers is the simplest (i.e. laziest) solution from both a coding and a design point of view. We have higher standards.

...simply printing a bunch of exclamation marks?

[VeryAngryFelid]

a. Disputable.
b. There are too many options. And options that change the game difficulty are IIRC frowned upon.

a. Not really. Check crawl philosophy section.
b. autofight_warning, force_more_message etc. Though I would be happy to get this reply from devs, because then it will be easier to convince them to make the numbers visible by default, if the numbers are that important for decision-making.

[CanOfWorms]

If nothing else, lets not pretend that more information leads to worse choices.

i'm pretty sure no one actually claimed this

There's a very good reason you do research before you make an important decision.

people may do research before making an important decision, but no one actually makes a important decision with perfect information. eventually the benefit/information ratio becomes too marginal

[dowan]

If nothing else, lets not pretend that more information leads to worse choices.

i'm pretty sure no one actually claimed this

"I think this" (not knowing precise numbers) "might, if anything, lead me to play overly cautiously, rather than leading to extra deaths."

people may do research before making an important decision, but no one actually makes a important decision with perfect information. eventually the benefit/information ratio becomes too marginal

In this case, it's less information with more cognitive load (!!! count, conversion rate, range) vs more information with less cognative load (the number itself you were trying to figure out in the previous steps).

[tabstorm]

What are the current obstacles to integrating Sequell data into gameplay as a submenu of x-v? I assume this is considered bad because of cognitive load, but in reality people will just go to the wiki or beem and check what features enemies have. There are also a few instances in the game where enemies can do a lot more damage than you would expect, the typical offenders are af:cold users. It's also useful to know when enemies can double-turn you or under what conditions you can outspeed them, as in, with swiftness, haste, swifthaste, up to energy randomization, and so on.

Also, I don't think AC has much of an effect on decision-making. What are you going to do, change weapons/spells based on enemy AC? On the other hand, enemy evasion is important to know, enemy speed is definitely important to know, and enemy damage is extremely important to know. I think it's most important for player ghosts, for instance if someone manages to leave an early AC+++++++ gargoyle ghost or EV+++++ spriggan ghost, you can know it's not worth the effort to fight it.

[VeryAngryFelid]

Also, I don't think AC has much of an effect on decision-making. What are you going to do, change weapons/spells based on enemy AC?

Yes

[CanOfWorms]

"I think this" (not knowing precise numbers) "might, if anything, lead me to play overly cautiously, rather than leading to extra deaths."

damn it siegurt

In this case, it's less information with more cognitive load (!!! count, conversion rate, range) vs more information with less cognative load (the number itself you were trying to figure out in the previous steps).

the cognitive load only exists for people who think that the precision leads to noticeable benefits

[Psieye]

So to sum up this debate, people fail to acknowledge that there's a spectrum of players. In other words, "my solution will satisfy everyone" if not "surely there must be a silver bullet solution (that we haven't found yet) which will satisfy everyone".

In broad strokes, there are two extremes of the spectrum. For simplicity, I will call it the 'organic' vs 'inorganic' scale. The inorganic side actively derives enjoyment out of number crunching. Parsing lots of arcane data correctly where others would stumble to get favourable outcomes feels good for them. This side aspires to ascend to 'the hypothetically optimal player who never miscalculates'.

The organic side wants to hone instincts and intuition. Through shared anecdotes and direct experience (which may or may not involve actual character death), they fine-tune their danger sense. Guessing very accurately and much faster than others is how this side derives enjoyment. I won't say they aspire to be playtime-based speedrunners, but they will gravitate in that direction.


These descriptions may be off (the natural counter = "do people really enjoy number crunching? They enjoy winning and it happens to require number crunching") but it should get people thinking.

[tabstorm]

Also, I don't think AC has much of an effect on decision-making. What are you going to do, change weapons/spells based on enemy AC?

Yes

I can't think of many situations where you would actually do this. I can only think of a few contrived situations like having both a highly enchanted great mace and demon whip and switching to the whip against low-AC enemies,or switching to the mace against very-high AC enemies. Most enemies have very low AC, a few have high AC. The precise value of AC isn't so important, just that some enemies have very high AC and you might not want to use LRD or Sandblast on them. On the other hand you already asked for trouble by playing EE...

[archaeo]

So to sum up this debate, people fail to acknowledge that there's a spectrum of players. In other words, "my solution will satisfy everyone" if not "surely there must be a silver bullet solution (that we haven't found yet) which will satisfy everyone".

The problem is that I've seen devs flatly reject some kind of "more numbers" option that could be toggled on and off, and I think it's fair to say that most of the disagreement here is about what's better, not what we could do to satisfy the biggest part of that spectrum of players.

Part of me wishes the game would turn these numbers on in wizmode (which it partially does), which would at least offer some form of numbercrawl in the wild. I don't think this would satisfying everyone, though.

[dpeg]

Note how simply printing a bunch of numbers is the simplest (i.e. laziest) solution from both a coding and a design point of view. We have higher standards.

...simply printing a bunch of exclamation marks?

Yes. You make it sound like !!! is all of the feedback Crawl gives to players, so I'll not go along that route. But the exclamation marks themselves are a nifty tool: they are not intrusive, they indicate meaningful information, and they are flavourful. It suffices if the players parses them as "Whoa, lotsa damage!"

Did you miss the part where he specifically said he didn't look up those numbers (fwiw i don't either, i have never looked up a monster's stats while playing, have never used beem, and only occasionally used knowledge bots while responding to some forum post about something.)

No, I didn't miss it. As far as I understand dpeg is not trying to play optimally and does not care about deaths much (no offense, dpeg, I play this way sometimes too).

You indeed miss my point. Yes, I play non-optimal. I do this on purpose, because that's my idea of fun. I think you could interpret this as: "I want to win the game just with the information it gives." Under this premise, I try to avoid mistakes. (Don't dare to call my playing style "non-spoiled optimal", because too often I just want bloodshed on the screen.)

What are the current obstacles to integrating Sequell data into gameplay as a submenu of x-v? I assume this is considered bad because of cognitive load, but in reality people will just go to the wiki or beem and check what features enemies have.

There is a problem before cognitive load: by doing so, we tell players, "this is relevant information, use it!". If you have ever seen a new player paralysed because they don't know which species to choose, or which spell to learn, or which consumable to use, then this is the ultimate paralyser. They get access to a bucketful of numbers, almost of all of which make no sense to them, but which are official because they can be seen in-game. In my opinion, this would be make Crawl much, much worse as a game, with just a single small change. If another game did this, I'd mention it as the pinnacle of horrible design decisions anywhere.

There are also a few instances in the game where enemies can do a lot more damage than you would expect, the typical offenders are af:cold users. It's also useful to know when enemies can double-turn you or under what conditions you can outspeed them, as in, with swiftness, haste, swifthaste, up to energy randomization, and so on.

Yes, and it is a good idea to think about how to convey these situations. This can be done! If there are ideas, please say so. Sequell data inside DCSS is definitely not the solution, however.

If nothing else, lets not pretend that more information leads to worse choices. There's a very good reason you do research before you make an important decision.

I don't think anyone is saying that. If you know what you're doing, more information will lead to better play and more wins.
I claim that more information is not more fun for everyone. Granted, more (even full) information is fun for some. For myself, it'd be less fun. Hence, someone has to decide where to draw the line. I hear you say, "then just print the numbers and you foolhardy ignoramuses can pretend they're not there". However, this is not true because I'll always assume that whatever feedback the game gives to me, it is relevant. I will try to parse and use it, thereby decreasing my fun.
In other words, I think it is much better if there is only a very controlled set of information the game gives out, and everyone who wants more has to look it up. Again, changes can happen, did happen, and will happen. But I want better reasons, for every single piece of information that we provide. And after that, a discussion about how to present that information.

The numbers inform the decision. If the decision is meaningless with numbers, it's meaningless now, the meaninglessness is just hidden. If the decision is important with numbers, it's important without them, it's just that you don't have the proper tools to make the decision.

It doesn't work like this. In real life, you make informed decisions all the time without knowing exact numbers (think about health, purchases, relationships etc.)

About !!! which came up several times: I hope you guys realise that the exclamation marks were never intended to be disassembled by players. They are a very natural and convenient way to display overkill, and the additional messages do the same. They allow you to get a feeling about your current damage output. Feel free to count them, but rest assured that you're reading more into this than ever was intended.

I'm terrified that dpeg is saying he'd rather have 20% of characters killed off unavoidably, but that's a whole other subject.

It's right in the philosophy section.
Setup A: 100% winrate, actually achievable (this is almost the case with current Crawl).
Setup B: even optimal play only gets you 80% winrate; there are that many unavoidable deaths.
Which setup allows for greater game depth, i.e. meaningful distinctions in playing skill?

[Psieye]

The problem is that I've seen devs flatly reject some kind of "more numbers" option that could be toggled on and off

A small problem from my perspective. The devs don't want that option in official crawl. There's nothing stopping people from putting that in an 'underground branch' that always pulls from trunk. Heck, with enough willpower, they could then build that branch and run a 'not-official-crawl' server somewhere so players aren't blocked by a technical know-how wall.

[Shard1697]

Now we tell players to get a local copy of crawl, fire up wizard mode, create a copy of their character and equipment, and run fsim to get that information to decide which weapon is better.

Imagine if players could just see how much damage they were doing instead of having to learn all of wizard mode's awkward commands to replicate a character+enemy and then fsim it

About !!! which came up several times: I hope you guys realise that the exclamation marks were never intended to be disassembled by players. They are a very natural and convenient way to display overkill, and the additional messages do the same. They allow you to get a feeling about your current damage output. Feel free to count them, but rest assured that you're reading more into this than ever was intended.

What else are we supposed to do when we don't get more specific info about how much damage we're doing?

I don't think anyone is saying that. If you know what you're doing, more information will lead to better play and more wins.
I claim that more information is not more fun for everyone. Granted, more (even full) information is fun for some. For myself, it'd be less fun. Hence, someone has to decide where to draw the line. I hear you say, "then just print the numbers and you foolhardy ignoramuses can pretend they're not there". However, this is not true because I'll always assume that whatever feedback the game gives to me, it is relevant. I will try to parse and use it, thereby decreasing my fun.
In other words, I think it is much better if there is only a very controlled set of information the game gives out, and everyone who wants more has to look it up. Again, changes can happen, did happen, and will happen. But I want better reasons, for every single piece of information that we provide. And after that, a discussion about how to present that information.

So why can't it be a rcfile option so people who don't like it can keep it turned off?

[dpeg]

What I am claiming is not just theory, I've seen it in practise.

When a new player asks for help, one of the first things to be recommended usually is "turn off auto skilling, use manual skilling". Now you can say many bad things about automatic skilling (and they might be true), but manual skilling is hard to get right for a new player: they have to learn the commands, the monsters, and some basic tactics. All of this works well with auto skilling. Switching to manual skilling that early makes their life harder, not easier.

If we'd provide in-game access to Sequell data, say, in whatsoever form, then I predict recommendations such as "Turn on option XY and check the numbers."

It's an art to trim down the information needed into manageable chunks, and simply heaping it upon the player is the sloppiest and worst way to do it. This will not happen, no matter how much the enjoyment or winrate of some advanced players may profit from it.

Feedback is fine, but it needs much more thought than this. If we'd brainstorm along that route instead of having this very same discussion over and over again, progress would be faster.

[Sprucery]

If I have two (or more) weapons and want to know which one is the best, what am I supposed to do?

[VeryAngryFelid]

I can't think of many situations where you would actually do this. I can only think of a few contrived situations like having both a highly enchanted great mace and demon whip and switching to the whip against low-AC enemies,or switching to the mace against very-high AC enemies. Most enemies have very low AC, a few have high AC. The precise value of AC isn't so important, just that some enemies have very high AC and you might not want to use LRD or Sandblast on them. On the other hand you already asked for trouble by playing EE...

Sandblast on early levels.

[VeryAngryFelid]

If we'd provide in-game access to Sequell data, say, in whatsoever form, then I predict recommendations such as "Turn on option XY and check the numbers."

IMHO it's like saying "Our restaurant does not give knife and fork to clients because I've seen one guy kill another with a dagger. Use tea spoons to eat your fish and pasta".

Nothing currently stops the same players from saying "use knowledge bots to check monsters and fsim to check weaposn". And the most funny thing is that some players are grateful for such advices (I remember one such thread recently).

[dpeg]

I think this discussion has run its course. You guys won't convince me, I am pretty sure I don't reach you, so that's the end of story.

I want to leave this thread by saying that (tavern) discussions about single gameplay-information pieces are useful and welcome.

[Siegurt]

I can't think of many situations where you would actually do this. I can only think of a few contrived situations like having both a highly enchanted great mace and demon whip and switching to the whip against low-AC enemies,or switching to the mace against very-high AC enemies. Most enemies have very low AC, a few have high AC. The precise value of AC isn't so important, just that some enemies have very high AC and you might not want to use LRD or Sandblast on them. On the other hand you already asked for trouble by playing EE...

Sandblast on early levels.

So you have mentioned early sandblast a couple times now, but what is the actual choice you are making here, "sandblast vs..... an untrained unenchanted short sword"? "Kill it with sand blast or run away"?

What is the actual other option that you are puporting would be informed by having acutal ac numbers vs the current situation, because i don't feel like you have made that clear..

[VeryAngryFelid]

So you have mentioned early sandblast a couple times now, but what is the actual choice you are making here, "sandblast vs..... an untrained unenchanted short sword"? "Kill it with sand blast or run away"?

What is the actual other option that you are puporting would be informed by having acutal ac numbers vs the current situation, because i don't feel like you have made that clear..

"Sandblast with last few stones and you have 2 adjacent monsters, both are shown to have the same number of AC pips but one of them has AC 0 and another has AC 4 (12! vs Sandblast)."
So you use sandblast with stones vs one monster and sandblast without stones vs another monster. Now depending on which monster is attacked with stones (AC 0 or 12) you will be either dead or alive.

[Siegurt]

So you have mentioned early sandblast a couple times now, but what is the actual choice you are making here, "sandblast vs..... an untrained unenchanted short sword"? "Kill it with sand blast or run away"?

What is the actual other option that you are puporting would be informed by having acutal ac numbers vs the current situation, because i don't feel like you have made that clear..

"Sandblast with last few stones and you have 2 adjacent monsters, both are shown to have the same number of AC pips but one of them has AC 0 and another has AC 4 (12! vs Sandblast)."
So you use sandblast with stones vs one monster and sandblast without stones vs another monster. Now depending on which monster is attacked with stones (AC 0 or 12) you will be either dead or alive.

But ac is only a small part of "if i can kill these two critters before they kill me" you might whiff your small number of attacks with stones and die, you might hit for max damage and one shot the thing with 4 ac, you might kill the 0ac first with stones because it has a whole lot more hps and does crap tons more damage (lets say the 0 ac critter is an ogre with a gsc, and the 4 ac critter is a goblin with ring mail and no weapon)

In the vast majority of cases what the best action to take there has nothing to do with AC, even if you are using sandblast as your primary killdudes and have a very limited supply of stones, and that is the absolute *most* impactful monster AC is, in literally every other situation AC is less meaningful.

And honestly if i only have like 3-5 stones left, and i see two critters who have any meaningful chances of killing me, the correct thing to do is *not fight them* you should retreat, go around, ditch them on some stairs, or if you are really desparate with no other options, kill the more dangerous one with stones, and kite the other one without, and "more ac" might or might not have anything to do with "more dangerous"

[VeryAngryFelid]

Siegurt
I am not sure what's your point. Do you mean decisions which give 1% chance to survive and 99% chance to survive are identical because both can result in death? Crawl is about making decisions, no decision guarantees survival so it is a game about maximizing survival probabilities.

You can assume that you are in corridor and you need to kill one of those monsters to be able to retreat because 5 orc priests are coming. Don't ask me how it happened, teleportation traps are still in the game.

[Shard1697]

just because something makes a small difference does not mean it makes no difference.

also, none of this changes that knowing damage numbers is important.

[Shard1697]

*Zardoz voice* THE PIP IS GOOD, THE NUMBER IS EVIL

[Siegurt]

Siegurt
I am not sure what's your point. Do you mean decisions which give 1% chance to survive and 99% chance to survive are identical because both can result in death? Crawl is about making decisions, no decision guarantees survival so it is a game about maximizing survival probabilities.

You can assume that you are in corridor and you need to kill one of those monsters to be able to retreat because 5 orc priests are coming. Don't ask me how it happened, teleportation traps are still in the game.

So your prerequisite situation for "monster AC is a meaningful number to display" is: 1. You are using sand blast to kill monsters 2. You are early enough in the game that you haven't found enough stones to cast sandblast with stones enough times to kill two creatures 3. You must kill both of those two creatures, in order to not die, because circumstances dictate that you are in a situation that will rapidly deteriorate if you try to escape qithout killing them (or it is impossible). 4. One of the two creaures is not already inherently more dangerous than the other. 5. They are at the same level of AC pip, but there is enough difference between the two that rocky sandblast and not non rocky sandblast will kill the one with more ac.

Note that in the actual described situation, if you only need to kill one critter to escape, and can kill that one with a rocky sandblast, and it doesn't matter which one has the higher AC, you need the requirement that you must kill both in order for AC to be relevant. (Or i suppose you could modify the scenario to say that rocky sandblast could kill the one with less ac, and you could not kill the one with more ac at all)

[duvessa]

Are we debating showing AC vs. not showing AC, or showing AC as a number vs. showing AC as +++?

[Siegurt]

Are we debating showing AC vs. not showing AC, or showing AC as a number vs. showing AC as +++?

I assumed it was as a number vs as +++

[VeryAngryFelid]

I am arguing that for low AC you need to display AC as number because of sandblast and that for high AC you need to display AC as numeric range because it is very inconvenient and weird to count pips in +++++++++++.

[CanOfWorms]

I am arguing that for low AC you need to display AC as number because of sandblast and that for high AC you need to display AC as numeric range because it is very inconvenient and weird to count pips in +++++++++++.

it's a good thing AC display doesn't actually go up to that amount of pips (it stops at 8)

[VeryAngryFelid]

it's a good thing AC display doesn't actually go up to that amount of pips (it stops at 8)

So a player ghost with AC 35 has the same number of pips as player ghost with AC 80, right?

[goodcoolguy]

It's right in the philosophy section.
Setup A: 100% winrate, actually achievable (this is almost the case with current Crawl).
Setup B: even optimal play only gets you 80% winrate; there are that many unavoidable deaths.
Which setup allows for greater game depth, i.e. meaningful distinctions in playing skill?

Reducing the winrate alone isn't what you want, of course. You want a deeper skill curve where the difference between winning 50% of the time and 100% of the time is a lot. Right now, even new players can achieve very high winrates just by listening to people give them advice in chat for the first few games then playing a very cautiously. There was a player recently who showed up apparently never having played before, but with an unusual willingness to take things slow and follow advice. He immediately achieved a winrate around 80%. The fact that the essentials of the game can be conveyed so completely through a few pages of text in a chat box should be worrying. I mean, yes, this guy was getting pro-tips, but come on.

What you want is a lot of avoidable deaths that are actually plausible given the resources available to the player. That's the problem with crawl today, lack of plausible deaths and it mostly comes from an excess of resources.

As for the OP and the AC discussion, the game could always just give you a damn histogram of damage rolls plus the probability of hitting at all. No need to get out your adding machine and statistics textbook to figure out what AC does.

[CanOfWorms]

it's a good thing AC display doesn't actually go up to that amount of pips (it stops at 8)

So a player ghost with AC 35 has the same number of pips as player ghost with AC 80, right?

yes

[nago]

From my point of view, putting the +++ ac value in xv as currently is was an useless and bad addition because that screen has already enough clutter on it.

I mean, what kind of clear or just useful information provides?

Does + stand for 1? For 1\8 of max value? (how many pipes are there?) Does full pipe mean max value?

And even if I knew that or there was the exact value, what the fuck I'd give about?
The only case I can think it does matter to know how much ac a monster have is early on as EE so you may choose if use sandblast, rock sandblast or try that 30% stone arrow and retreat if it fizzles two times in a row.
Oh and when you see that +++++++ ac "learn DU or leave me alone" Gr ghost. But you already know that players ghost are a nuisance.

All 99.9% of times you just hit things with your stronger base attack anyway, no matter their ac because in dcss there isn't nothing developed to differentiate things around that.

The display of ev is as much awful as ac but at least it may be not totally useless because players often have some things to do to play better against high ev monster so knowing beforehand that could lead to meaningful decisions, like using a lower damage higher accuracy weapon early on, or an undodgeable attack later on.

That said, if ac must be displayed I think numeric values would be better. At least that would provide a clear, not ambiguous, hard to read or wrong kind of data.

[Hurkyl]

AC also matters to characters prone to firing around magic darts and hitting things with poorly enchanted short blades.

[VeryAngryFelid]

All 99.9% of times you just hit things with your stronger base attack anyway, no matter their ac because in dcss there isn't nothing developed to differentiate things around that.

Some characters have different tools for different AC: Lightning Bolt (half AC) ,Tornado (less than half AC), OoD (just very high damage).

Spoiler: show

Recent example https://crawl.project357.org/morgue/San ... 035455.txt

[dpeg]

It's right in the philosophy section.
Setup A: 100% winrate, actually achievable (this is almost the case with current Crawl).
Setup B: even optimal play only gets you 80% winrate; there are that many unavoidable deaths.
Which setup allows for greater game depth, i.e. meaningful distinctions in playing skill?

Reducing the winrate alone isn't what you want, of course. You want a deeper skill curve where the difference between winning 50% of the time and 100% of the time is a lot. Right now, even new players can achieve very high winrates just by listening to people give them advice in chat for the first few games then playing a very cautiously. There was a player recently who showed up apparently never having played before, but with an unusual willingness to take things slow and follow advice. He immediately achieved a winrate around 80%. The fact that the essentials of the game can be conveyed so completely through a few pages of text in a chat box should be worrying. I mean, yes, this guy was getting pro-tips, but come on.

What you want is a lot of avoidable deaths that are actually plausible given the resources available to the player. That's the problem with crawl today, lack of plausible deaths and it mostly comes from an excess of resources.

Exactly! When I say "lower winrate", that's very crude. I really mean everything you say. Underlying this is my assumption that more depth is easier to achieve if the "perfect winrate" (assuming very best play) is significantly less than 100%, rather than almost 100% (as it is now).

[tabstorm]

It's right in the philosophy section.
Setup A: 100% winrate, actually achievable (this is almost the case with current Crawl).
Setup B: even optimal play only gets you 80% winrate; there are that many unavoidable deaths.
Which setup allows for greater game depth, i.e. meaningful distinctions in playing skill?

Reducing the winrate alone isn't what you want, of course. You want a deeper skill curve where the difference between winning 50% of the time and 100% of the time is a lot. Right now, even new players can achieve very high winrates just by listening to people give them advice in chat for the first few games then playing a very cautiously. There was a player recently who showed up apparently never having played before, but with an unusual willingness to take things slow and follow advice. He immediately achieved a winrate around 80%. The fact that the essentials of the game can be conveyed so completely through a few pages of text in a chat box should be worrying. I mean, yes, this guy was getting pro-tips, but come on.

What you want is a lot of avoidable deaths that are actually plausible given the resources available to the player. That's the problem with crawl today, lack of plausible deaths and it mostly comes from an excess of resources.

Exactly! When I say "lower winrate", that's very crude. I really mean everything you say. Underlying this is my assumption that more depth is easier to achieve if the "perfect winrate" (assuming very best play) is significantly less than 100%, rather than almost 100% (as it is now).

How would more unavoidable deaths, and I mean actually unavoidable, not something like dying on U:3 as the result of an error made on D:10, make it easier to create a deeper game? This is not a sarcastic question, I'm just curious what ideas you have in mind here.

[dpeg]

tabstorm: currently, there's a ceiling which is not hard to reach. I should say that "unavoidable death" is imprecise -- I don't think of Xom deaths for hapless characters. However, if the game becomes considerably harder, then there will be many situations that *feel* like unavoidable deaths. However, there will be players who can prevent some of these deaths. That's a skill level differentiation that's currently visible (not needed because 100% winrate can be achieved).

So I admit that "unavoidable death" is imprecise. I really mean "harder game", assuming that winrate does sink.

edit: corrected "precise" to "imprecise", a bad typo

[Lasty]

Our brain cannot easily differentiate ******** from **********

why are we forcing players to count the !!!!!

As dpeg said, I don't think anyone expected players to count !s and then use that as a crude numeric comparison. The increasing numbers of !s is, as I understand it, simply an unobtrusive way to say, "hey, you hit that monster really hard just now!" As for number of pips in AC/EV displays, humans are bad at counting a long sequence, but are fine at counting short sequences. *** *** ** is really easy to count. We should just add some breaks to the display to make them easier to count. Again, the point isn't perfect accuracy, but instead to let players go "huh, that's a lot of pips. That one must have lots of AC."

How much science is behind this? While I recognize the noble intent, my impression of the actual execution is that obfuscation is done for the sake of having obfuscation rather than being tuned for usefulness -- and the quality of the result judged by people who don't really care about having the information presented to them.

please remove knowledge bots and close source on monster data. otherwise how do we know that if many good players check ac, ev and damage out of game?

This is root cause of the problem IMHO. All devs have easy access to monster data, they know all those monster AC/EV/HP/damage/spell damage etc. That's why I suggested to remove that access, add some new monsters and see how you will like fighting it knowing nothing about its damage and not that much about AC/EV (in stable it would be even worse because average HP is not shown). You should try playing like you are unspoiled player, stop using beem/knowledge bots etc.

Nothing currently stops the same players from saying "use knowledge bots to check monsters and fsim to check weaposn". And the most funny thing is that some players are grateful for such advices (I remember one such thread recently).

There are some pretty inaccurate (and, to me, strange) assumptions here: That devs have more information about monster stats because they are devs than other players, and that we design the game around the assumption that others should have the same level of knowledge. This is just not true. First, any player has the option of getting perfect info about monster stats from out-of-game sources, not just devs. Second, at least two devs in this thread (dpeg and I) really don't use out-of-game monster stats and haven't memorized many if any monster stats.

I have used the knowledge bots a few times in game over the many years I've been playing, but doing so taught me that it almost never gives me info I needed that I couldn't have figured out myself, except during the periods where certain monster characteristics weren't well-documented in-game. I also don't have many monsters' stats memorized. I'm pretty sure the only monster stats I know precisely offhand are 1) hydras have 0 AC; 2) death yaks hit for 30 and yaks for 20; 3) stone giants can deal 45 in melee and 65 with a thrown rock; 4) juggernauts attack for 90 and 30 (because I worked on them recently). I don't think I know with any degree of certainty any other monster damages, AC, EV, HP, etc. And it turns out that that is just fine. I have a ballpark idea of how dangerous and tough the involved monsters are based on having had encounters with them before, and that is more than enough to make good decisions. I play in sudden-death challenges like CDSC and dieselrobin, and when I do I'm no more likely to seek out monster data than normal.

Really, playing well has almost nothing to do with knowing any exact monster stats. It has a lot more to do with mastering basic tactics, being able to predict how a situation will develop, having an escape plan, and leveraging all your resources creatively as needed.

One could raise the argument that this is well and good for very experienced players and it screws over new players, and there's some truth to that. But that's more or less how crawl is intended to work. And, more importantly, that's about equally true whether you have the numbers or not. Looking at xv and seeing "this monster has a fair amount of AC" gives a new player just as much to work with as checking a knowledge bot and seeing "this monster has 15 AC". Both have very little context until you have the experience of fighting a monster with that much AC in a variety of conditions.

Imagine if there were some resource in-game that could tell you either a monster's accurate stat or a wildly wrong one: "this monster has 0 or 10 AC". I would have no difficulty telling which one was accurate using only in-game feedback. Wouldn't you?

As for the science behind this, this discussion isn't about science. It's about user experience, and how to present information to a player in the ideal way to allow them to best interact with the game. AFAIK, no one is obscuring the numbers just for the sake of obscuring them. The question is "what's the best thing to display for the quality of the game experience?" I'm sure there's room to improve, but I'm also sure that it doesn't mean displaying every number a player might think to ask about.

Experienced players may well suggest that fsim and knowledge bots are the correct way to learn about the game, and new players may well be grateful for the advice, but that's not evidence for anything except that some experienced players feel that knowing certain numbers helps them, and some new players are grateful to hear an experienced voice tell them that once they know the numbers their problems will go away. The knowledge bots are a useful tool, but IMO not because they're giving specific monster numbers for use in live play. fsim is a great testing tool for devs making potentially balance-affecting changes, but my experience is that it isn't at all necessary to play at a top level.

For a long time I thought that giant frogs are tougher than spiny frogs!

One of my recent deaths was a death to Snorg, I didn't expect to lose 72 HP in a single attack.

There are a lot of specific examples in this thread, and in general my response is going to be that it's easy to cherry-pick a specific moment and say that yes, at that moment if you knew that a monster could do exactly fatal damage to you, you might blink instead of walking away. However, it's usually correct to blink away in those circumstances anyway; randomized energy, a wandering monster, or any number of other things could finish you off if you let a monster hit you down to extremely low HP. Good play involves trying to avoid going below a significant fraction of your max hp.

For example, a player can easily see that Snorg hits 3x/turn, that normal trolls can do considerable damage, and that berserk monsters do extra damage and can act twice per player action. You look up all the numbers of monsters, and somehow they made you feel it was safe to stand next to berserk Snorg, even though it's not. Giving you more numbers wouldn't have helped, but reminding you that berserk shit is scary probably would have.

As for the giant frog/spiny frog thing, I'm not sure what to make of that.

Uncertainty changes your experience of the game

We have uncertainty from random rolls - is it insufficient? What, we want uncertainty from ignorance too? If so, shouldn't we deprive the player of the feedback/information rhat we so generously bequeath? Hide resists, hide pips, let them figure it out.

Uncertainty from ignorance isn't a bad mechanic in some games -- survival horror games tend to trade heavily on it. But no, crawl isn't trying to use uncertainty-from-ignorance, except the standard ignorance-of-what-rolls-will-happen and -what-the-future-holds. That said, knowing the specific values for a lot of things changes your attitudes towards them, even if you can't meaningfully leverage those values. It gives a feeling of control even in the absence of control. I think that the absence of that feeling of control is a happy side-effect of displaying the most useful amount of information in crawl.

[VeryAngryFelid]

As for number of pips in AC/EV displays, humans are bad at counting a long sequence, but are fine at counting short sequences. *** *** ** is really easy to count. We should just add some breaks to the display to make them easier to count. Again, the point isn't perfect accuracy, but instead to let players go "huh, that's a lot of pips. That one must have lots of AC."

I find it weird that you do not want to provide a number if player is expected to count those ***. It's like writing XL 5 5 3 when player has XL 13.
If player is not expected to count them, then just replace them with a description AC none/very low/low/average/high/very high

[VeryAngryFelid]

One could raise the argument that this is well and good for very experienced players and it screws over new players, and there's some truth to that. But that's more or less how crawl is intended to work.

Exactly, you have just confirmed my point that crawl devs know monster stats or just have a "feel" of them due to playing a lot so they don't want to add what they don't need.

For example, a player can easily see that Snorg hits 3x/turn, that normal trolls can do considerable damage, and that berserk monsters do extra damage and can act twice per player action. You look up all the numbers of monsters, and somehow they made you feel it was safe to stand next to berserk Snorg, even though it's not. Giving you more numbers wouldn't have helped, but reminding you that berserk shit is scary probably would have.

The problem is that I am often lazy and in this particular case I didn't check Snorg's damage, that's my complaint, it should be listed in the game. I don't think I should use a scroll of blinking from every monster when I am at about 80% HP assuming I have no idea how much damage it deals. Do you remember a thread where I was complaining that it is very weird that players die to Ogre at 30 HP, then again at 35 HP, then again at 40 HP. Where should they stop? Should they blink at 70 or 100 HP also?

I am arguing as a new player who is unspoiled and wants to make good (not even optimal, just good) decisions using in-game info only. currently it is almost impossible because of hidden damage numbers, unclear AC/EV is just a minor thing. If I use consumables to escape when there is no need, I run out of consumables and eventually die. If I don't escape, I die to a monster when it gets high damage roll.

[goodcoolguy]

tabstorm: currently, there's a ceiling which is not hard to reach. I should say that "unavoidable death" is imprecise -- I don't think of Xom deaths for hapless characters. However, if the game becomes considerably harder, then there will be many situations that *feel* like unavoidable deaths. However, there will be players who can prevent some of these deaths. That's a skill level differentiation that's currently visible (not needed because 100% winrate can be achieved).

So I admit that "unavoidable death" is precise. I really mean "harder game", assuming that winrate does sink.

Should really give hellcrawl a look. Part of it is strengthened spawns, but a huge piece of what you're looking for is a matter of cutting experience and items.

[dpeg]

I find it curious how you get long, in-depth replies, and then pick some detail to bicker about. I realise it's not our job to convince anybody, but this form of communication seems pretty pointless to me altogether.

As for AC: I think it shouldn't have been displayed in the first place, and certainly not as a number. Words might be a good fit ("no armour", "lightly armoured", "heavily armoured"), but, as always, that's actually *more* work for us.

Sandman: Can you please stop this nonsense about "devs don't give numbers because they know and players don't deserve them". It's wrong, preposterous, and other adjective that momentarily escape me.

[Hurkyl]

As for the science behind this, this discussion isn't about science. It's about user experience, and how to present information to a player in the ideal way to allow them to best interact with the game. AFAIK, no one is obscuring the numbers just for the sake of obscuring them. The question is "what's the best thing to display for the quality of the game experience?" I'm sure there's room to improve, but I'm also sure that it doesn't mean displaying every number a player might think to ask about.

There's a lot of science behind user experience (such as, for example, actually gathering data about user experiences). But let me rephrase my point. For example, in regards to displaying monster AC:

By what mechanism were candidates for how to display monster AC generated?

By what mechanism was it decided that one pip per 5 AC was the most promising candidate to implement? The only candidate to test?

By what mechanism was it judged whether one pip per 5 AC really is a good enough way to display the information? That it is the best way?

I would be honestly surprised if the whole process was much more than "We like to obfuscate things. 5 is a nice round number. We'll show one pip per 5 AC and call it a day."

[VeryAngryFelid]

I find it curious how you get long, in-depth replies, and then pick some detail to bicker about. I realise it's not our job to convince anybody, but this form of communication seems pretty pointless to me altogether.

If I don't reply to something, it means I have nothing to say or I agree.

Sandman: Can you please stop this nonsense about "devs don't give numbers because they know and players don't deserve them". It's wrong, preposterous, and other adjective that momentarily escape me.

Do you have any other reason why you don't want to make everyone happy? Please don't say again that experienced players will give a "bad" advice to turn on the new option, it is not an argument (you ignored my reply about it too by the way). I don't think worse about devs than they seem to think about players, we are all people and can be wrong anytime.

[Lasty]

You're wrong there. We A/B tested this across a broad cross-section of demographics in a series of a 300-person trials! After three years and hundreds of thousands of dollars, we discovered that this was the best possible scenario.

C'mon man, this is a labor-of-love open-source project. Of course we're not running lengthy studies. But listen to what I'm saying: no one is starting from "let's hide all the numbars!" People are trying to determine as best they can which numbers to show directly and which to show as ballparks and which to not show at all.

[sanka]

The thing about hidden numbers is that I have played crawl since version 4.5 and I still do not have any idea without fsim which weapon is better. When I do try fsim the difference sometimes very big (read: one weapon does about 80% more damage). It is pretty annoying.

[Hurkyl]

But listen to what I'm saying: no one is starting from "let's hide all the numbars!" People are trying to determine as best they can which numbers to show directly and which to show as ballparks and which to not show at all.

It's not evident to me.

But, admittedly, after being told so many times "It's a matter of principle. I like it this way. Your user experience is invalid." it's hard to see the things as being any other way than that.

[sanka]

I also do not understand any argument against showing monster damage numbers. Yeah, after playing a lot I have a sense about monster damage but part of the good strategy is to know which monster you let walk next to you. Thi mostly depends on the damage of the monster, so it's also pretty annoying to need to check it on IRC, even if I very rarely do it. Monster description contains a lot of rarly matters things like resistances etc., but not the most important one: can this monster kill you?