Learning from death


Although the central place for design discussion is ##crawl-dev on freenode, some may find it helpful to discuss requests and suggestions here first.

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Thursday, 3rd September 2015, 20:45

Learning from death

I believe there is some misunderstanding about learning from death and about roguelikes in general (maybe it is my misunderstanding but whatever).
Any good roguelike is reliably winnable by any smart person in one attempt who has never played roguelikes before.
Let me describe some examples.

1) You see an unknown monster called Ogre with Giant Spiked Club and you are a melee character with 25 HP.

1a) Good roguelike will display that the Ogre can deal up to 39 damage in a single attack. If player tries to attack the monster and dies in a single attack, the player learns "I was stupid, I should escape (use scroll of blinking etc.) when there is a monster who can one-shot me. And of course I should not melee it if I have any choice". You learn it once and it works with all monsters (Ettins, Hydra simulacrum, Azure Jelly, Jugernaut etc.)

1b) Bad (you can treat it as "not so good" if you like) roguelike will not display max damage. If player tries to attack the monster, gets 27 damage and dies, the player learns "I should not melee Ogre when I have less than 28 HP. I still can try to melee it if I can kill it with 2 attacks, I have more than 27 HP and the ogre misses me during first attack. I should escape (use scroll of blinking etc.) if I am adjacent to Ogre and I have less than 28 HP".
You learn 27 damage for Ogre, then you learn 39 damage for Ogre, then you learn 70 damage for Ettin, then you learn 116 damage for Etting, then you learn 130 damage for Azure Jelly, then you learn 272 damage for Azure Jelly, then you learn 105 damage for Jugernaur, then you learn 120 damage for Jugernaut etc. That's lots of unnecessary deaths to learn something that can be see in crawlwiki or #crawl.
Should the player think "I was stupid, I am guilty in attacking extremely dangerous monsters"? If your answer is yes, continue reading very carefully.

2) You see an unknown monster which is extremely dangerous.

2a) Good roguelike either has very few extremely dangerous monsters so you can avoid them almost always or it has not so few extremely dangerous monsters and then it provides a detailed description of them because you are expected to kill at least some of them, you cannot avoid them for too long. Good roguelike will list that extremely dangerous Boggart has 13-28 HP so you can try to kill it with a wand or other relatively weak attack. Good roguelike will list that extremely dangerous Caustik Shrike has 95-141 HP so using a wand or another relatively weak attack is not a very good idea. If player dies to Caustik Shrike, he learns "I was stupid, I should not use a wand which deals 30 max damage on a monster with potentially 141 HP".

2b) Bad roguelike has not so few extremely dangerous monsters and does not give detailed description of them. So you cannot avoid them for too long but you cannot choose a good target for killing either. If you try to attack an extremely dangerous Caustik Shrike with a wand, you will get 114 damage in a single turn and will die. Because you had no idea that Caustik Shrike can have 141 HP, deal 57 damage per attack and attack twice per turn. After death player learns "Strange, Caustik shrike wasn't damaged that much, probably it has high HP, or maybe I was just unlucky and it is not realy much tougher than Boggart"

3) You see an unknown monster which is suspectible to Control Undead, Mass Confusion, Mephitic Cloud, Dazzling Spray, fear scroll, blinking scroll during constriction etc.

3a) Good roguelike will provide a way to see how likely you are to affect the monster. It will show something like "Skeletal Warrior has 15% chance to be affected, Centaur Zombie has 56% chance to be affected. Continue casting Control Undead? y/N"
Good roguelike tries to be nice to all players so if some players don't want to see the chances, there will be a way to hide it (even as default probably). After death player learns "I was stupid, casting the spell with 15% chance was a serious mistake" or "I was stupid, fear had only 45% chance to save me while blinking had 75% chance, why did I save blinking?"

3b) Bad roguelike will hide this chance, really bad ("really not so good", remember?) roguelike will have some hidden parameter like HD which makes it impossible to affect some monsters (Frederic and Mephitic Cloud, for example). After death player learns "Probably I was unlucky or Frederick is special-cased to be resistant to Mephitic Cloud, let's write it down for future games"
Did I say hidden parameteres? Let's talk about it some more.

4) You see a monster, it shoots at you with arrows.

4a) Good roguelike will treat all monsters with arrows in the same way. If you got adjacent to Centaur and it switched to melee immediately, then you safely assume that when you get adjacent to Deep Elf Master Archer, it will switch to melee too exactly like Centaur did. If Deep Elf Master Archer is special, you will clealry see it in its description, something like "This monster continues shooting arrows at you even when adjacent". Without any dark and fog which can be problematic to non-native English speakers. After death player learns "I was stupid, description clearly states that it does not make much sense to get adjacent to Master Archer, why didn't I read description carefully, I should have escaped instead".

4b) Bad roguelike will have several types of monsters, one archer will switch to melee (no "Archer" flag) and another will not, one monster will have increased accuracy ("Fighter" flag) and another will not, one monster will be able to open doors ("Open doors" flag) and another will not and so on. You will get nice descriptions like "The blade it wields is ethereal, passing easily through shields and armour, and rendering those it strikes more susceptible to magic." instead of clearly saying something like "their attacks completely ignore shields and are only affected by half of your AC, also your MR has 50% chance to be halved after successful attack for 5-22 turns".
Edit. To make it clear having all these flags is fine, I love it. I don't love it that they are not visible in monster description because they can be quite important.

5) You enter a new branch

5a) Good roguelike will describe every branch, providing a list of typical monsters. So when you are going to enter Elven Halls from Orcish Mines at XL 10, you will see something like "The branch can contain Deep Elf Sorcerers, Deep Elf Annihilators, Deep Elf Blade Masters etc.". and you will be able to check every monster and think "Uh, I have 45% chance to be banished by the sorcerer. U, the annihilator can deal 102 damage with LCS. I should probably avoid this branch for a while". Ok, with Elven Halls it is kind of expected that it is populated with Elves so you can probably use "?/melf" and be fine, but you would get the same description for Depths, Vaults, Crypt etc.

5b) Bad roguelike will describe branches like this: "Possibly constructed by some powerful madman, the Vaults are said to be home to great treasures. The blood spattered liberally around the entrance suggests that there is a certain amount of violence, too.". Nice, I know that I should be prepared for 60% Mark, 100% door seal and V5 ambush, don't I???

6) You enter a new floor

6a) I will start with bad roguelike first. Bad roguelike will kill you for entering a new floor with wrong equipment. After the death you will learn "I see, so Erolcha can generate on D9, I will need to wear ring of MR instead of robustness/shaolin when entering this floor". Or "I see, so Ijyb is special-cased to have a wand of fire on D2", I should wear ring of rF+ instead of ring of MR/+5 AC when entering the floor. Or "Wow, Ijyb can have wand of draining too. Not sure what I should wear now".

6b) Good roguelike will not effectively kill you from the edge of LoS. It will give you a chance to do something vs banishment, it will let you know that Ijyb has a wand of cold in inventory and the wand can do 1-37 damage and that you can get it to 50% by rC+, to 33% by rC++ and to 20% by rC+++". You will learn something from the death, because it was your mistake, you knew what you were against and you didn't do everything you could (I am ignoring unavoidable deaths here)

7) Monsters with unknown properties

7a) Good roguelike will warn you that you can be paralyzed by melee attack of a panlord which ignores everything. If the same Panlord can cast LCS or Fire Storm, you will see it in its description and will act accordingly.

7b) Bad roguelike will describe the monster as "A formidable demon lord inhabiting the endless halls of Pandemonium. Each one is different, with its own set of strengths and weaknesses." and then will write "You die..." after unavoidable death of a character you spent literally hours on.
If you put stasis one, you can get Singularity or Fire Storm and just die because you are unable to blink. If you don't put stasis on, you can get summons with paralyzing eyes or chaos melee attack so there is no safe choice here but at least you knew what you were against.

8) Damage

8a) Good roguelike will have a clear feedback for player to let him know how he's doing. That includes 66% chance for Throw Flame vs a goblin as XL 1 DEFE with 1d10 roll. Or if player wants, he can get just damage numbers "Flame hits goblin (3 damage)". Or if player wants, no damage numbers at all (as default). Those numbers make it much easier to choose what spell to use, what weapon to use, what launcher to use, what wand to use etc. Good roguelikes show something like "Throw Icicle hits Ice Dragon but it partially resists (damage 15 instead of 43). This makes it obvious for players how training magic skills increases damage from spells, how training Evocations increases damage from wands/rods, how training Invocations increases damage from abilities etc.

8b) Bad roguelike will hide everything, you will get only two types of message "You hit foo" and "You miss foo". No info about current monster HP also. Too unplayable? Ok, let's add "You hit foo but do no damage". What, still unplayable? Ok, let's add "You hit foo!" , "You hit foo!!" and "You hit foo!!!". No, you don't need to know what is the difference between "!" and "!!!", the latter is better and that's all you need to know. If you want to compare weapons, count the "!" for yourself. No, "!!" is not the same as "!" + "!" and the damage is randomized and non-linear too. Don't be too demanding, it is a free game. Why do you need to know how damaged the monster is? You know that you can spend literally dozens of turns without dealing any damage, don't you? Just out of curiosity? Ok, we can split monster HP in 4-5 categories and we will call them "lightly wounded", "heavily wounded", "almost dead" etc. What, that's too easy for players? Ok, let's randomize monster HP, 9-21 HP for a monster 7-17 smiting attack should be good enough. We will add different hidden stepdowns (for Dex, for spellpower, for piety) while ignoring OP items and spells which are used by 70+% winning characters.

TL;DR. Learning from death means player should learn what he did wrong, not that some monster can deal 27 damage.

This is my final post. Thank you everyone, I enjoyed my years on the forum. Happy crawling!
Last edited by Sandman25 on Thursday, 3rd September 2015, 21:55, edited 2 times in total.

For this message the author Sandman25 has received thanks: 21
Arrhythmia, byrel, duvessa, le_nerd, mopl, Rast, Sar, scorpionwarrior, Shard1697, WalkerBoh and 11 more users

Sar

User avatar

Ziggurat Zagger

Posts: 6418

Joined: Friday, 6th July 2012, 12:48

Post Thursday, 3rd September 2015, 21:18

Re: Learning from death

Sandman25 wrote:Bad roguelike will have several types of monsters, one archer will switch to melee (no "Archer" flag) and another will not, one monster will have increased accuracy ("Fighter" flag) and another will not, one monster will be able to open doors ("Open doors" flag) and another will not and so on.

I was kind of on your wave until this part.
Sandman25 wrote:This is my final post.

That's pretty sad to hear. Good luck.

For this message the author Sar has received thanks: 3
MainiacJoe, Rast, Sandman25
User avatar

Barkeep

Posts: 4435

Joined: Tuesday, 11th January 2011, 12:28

Post Thursday, 3rd September 2015, 21:20

Re: Learning from death

Posts: 10211

Come on, you need another 90 posts to get back to palindrophy.
I am not a very good player. My mouth is a foul pit of LIES. KNOW THIS.
User avatar

Abyss Ambulator

Posts: 1194

Joined: Friday, 18th April 2014, 01:41

Post Thursday, 3rd September 2015, 21:22

Re: Learning from death

DIsplaying something like the equivalent of Sequell data will absolutely never be done because of "paralysis by analysis", as far as I know. To be fair, there is a lot of memory demand on the player with all the statuses, resists, weird behaviors and interactions that are different for players and monsters (like potion effects), 500+ (?) different enemies with different speeds varying dam based on wielded gear among other things, randomized hp, combat data that work differently for monsters and players (and almost no one knows exactly how these work on either side anyway because the formulae are very complicated), etc.

You could argue that if you gave the player all this data it would be sort of empty of value (since it's not comparable to player data of the same type), and all you can hope to do is get a qualitative understanding of the impact of e.g. monster AC values. This is sort of what we have now though, but without the number listed. Because of all this ambiguity I imagine the goal is to give only a qualitative understanding to the player, to avoid paralysis by analysis. Then again, a lot of players query Sequell and aren't paralyzed by analysis, so whatever. Anyway, such a thing would be a lot of effort to code and keep up to date which is another reason it won't ever be done.
remove food

Vestibule Violator

Posts: 1601

Joined: Sunday, 14th July 2013, 16:36

Post Thursday, 3rd September 2015, 21:27

Re: Learning from death

tabstorm wrote:and keep up to date which is another reason it won't ever be done.

Actually, for the most part, keeping this up-to-date would be far easier than any other means of presenting this information, since it's taken directly from the game. You even get the added advantage of getting the information for the precise version you're playing at the time.

For this message the author Hurkyl has received thanks: 3
duvessa, Sandman25, Sar
User avatar

Abyss Ambulator

Posts: 1194

Joined: Friday, 18th April 2014, 01:41

Post Thursday, 3rd September 2015, 21:30

Re: Learning from death

I stand corrected then...
remove food

Tartarus Sorceror

Posts: 1694

Joined: Tuesday, 31st March 2015, 20:34

Post Thursday, 3rd September 2015, 22:41

Re: Learning from death

Well I hope you decide not to leave. You have always had a unique viewpoint, and you usually have good justifications, even if I disagree with you a lot of the time. Just take a break for a while.

For this message the author dowan has received thanks: 9
all before, Brannock, byrel, MainiacJoe, Rast, rchandra, rockygargoyle, Sandman25, TeshiAlair
User avatar

Ziggurat Zagger

Posts: 4478

Joined: Wednesday, 23rd October 2013, 07:56

Post Thursday, 3rd September 2015, 22:55

Re: Learning from death

I disagree with many of the things in the OP (I also agree with some!), but I thanked because it was a good post. You made it quite clear what you think makes a game good or bad.

I sincerely hope you will come back, I for one have enjoyed reading your posts (and I dare to say I have learned a lot about Crawl from them).
DCSS: 97:...MfCj}SpNeBaEEGrFE{HaAKTrCK}DsFESpHu{FoArNaBe}
FeEE{HOIEMiAE}GrGlHuWrGnWrNaAKBaFi{MiDeMfDe}{DrAKTrAMGhEnGnWz}
{PaBeDjFi}OgAKPaCAGnCjOgCKMfAEAtCKSpCjDEEE{HOSu
Bloat: 17: RaRoPrPh{GuStGnCa}{ArEtZoNb}KiPaAnDrBXDBQOApDaMeAGBiOCNKAsFnFlUs{RoBoNeWi

For this message the author Sprucery has received thanks: 3
byrel, Sandman25, TeshiAlair
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Thursday, 3rd September 2015, 22:57

Re: Learning from death

The thing I really don't get about this argument is that it seems predicated on the idea that "good game design" involves giving the player every bit of pertinent information about the game when, in fact, virtually every beloved game I can think of (including several roguelikes) hides significant amounts of information from the player. Obviously, part of designing a game is deciding what information should be hidden and what needs to be provided, but Crawl is hardly an outlier in the things it chooses to keep hidden.

Spider Stomper

Posts: 234

Joined: Thursday, 30th April 2015, 08:29

Post Thursday, 3rd September 2015, 23:21

Re: Learning from death

I learned a rant from death.

For this message the author onget has received thanks:
Arrhythmia

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Friday, 4th September 2015, 00:48

Re: Learning from death

I read somewhere that tl;dr means "thinking lazily; dumb reply." Whatever it is, it's unquestionably childish and trollish when it forms the entire contents of a post.

@Sandman25: sorry you're going through this man. You've contributed a lot, fought the good fight. Props, here.

For this message the author Aule has received thanks: 3
byrel, rockygargoyle, Sandman25

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 4th September 2015, 00:57

Re: Learning from death

Aule wrote:I read somewhere that tl;dr means "thinking lazily; dumb reply." Whatever it is, it's unquestionably childish and trollish when it forms the entire contents of a post.

@Sandman25: sorry you're going through this man. You've contributed a lot, fought the good fight. Props, here.

It actually means "too long; didn't read", sometimes people post a synopsis of their own long post after "tl;dr" for the benefit of those who aren't willing to read an entire long post.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks:
Arrhythmia

Dungeon Master

Posts: 634

Joined: Sunday, 22nd September 2013, 14:46

Post Friday, 4th September 2015, 01:59

Re: Learning from death

Siegurt wrote:
Aule wrote:I read somewhere that tl;dr means "thinking lazily; dumb reply." Whatever it is, it's unquestionably childish and trollish when it forms the entire contents of a post.

@Sandman25: sorry you're going through this man. You've contributed a lot, fought the good fight. Props, here.

It actually means "too long; didn't read", sometimes people post a synopsis of their own long post after "tl;dr" for the benefit of those who aren't willing to read an entire long post.

Aule is making the point that a post consisting in its entirety of 'tl;dr' isn't constructive. While I agree, there's also a reason there are report buttons. If a post isn't constructive, this is GDD, and there are rules against that.

Zot Zealot

Posts: 1031

Joined: Friday, 26th April 2013, 19:52

Location: AZ, USA

Post Friday, 4th September 2015, 06:29

Re: Learning from death

Seems to me that the last 4 posts were all contentless and could stand to be deleted.

I also thanked Sandman's post for being detailed and well-written. This is also like the 10th time I've seen him say it's his last post, so I wouldn't read much into that if I were you guys.

tabstorm wrote:DIsplaying something like the equivalent of Sequell data will absolutely never be done because of "paralysis by analysis", as far as I know. To be fair, there is a lot of memory demand on the player with all the statuses, resists, weird behaviors and interactions that are different for players and monsters... etc.

In a lot of ways, the current situation regarding "paralysis by analysis" doesn't seem any worse than it would be with Sequell data available to the player in-game. This is especially true for new players, who have a ton of information to learn and absorb as it is. I think Sandman correctly points out that "memory demand" (good phrase) is made worse by the fact that players are not encouraged to efficiently learn from mistakes, and must learn and re-learn information based essentially on trial and error. This wouldn't be so bad except that, as Sandman points out, at some point in the game you're learning by trial and error after investing multiple hours into a character.

I wouldn't want the equivalent of Sequell data popping up on the bottom of the screen during play. But having more data available on the xv menu regarding hidden monster mechanics, or explanations for how non-intuitive mechanics work (or hell, even as a sub-menu on the xv menu called "Detailed Information", for example) seems non-controversial to me.
Last edited by WalkerBoh on Friday, 4th September 2015, 23:17, edited 1 time in total.

For this message the author WalkerBoh has received thanks: 4
all before, dowan, Kismet, Sar

Dungeon Dilettante

Posts: 2

Joined: Thursday, 3rd September 2015, 10:41

Post Friday, 4th September 2015, 11:40

Re: Learning from death

The only roguelike I know that matches all points described by OP as "good roguelike" is ToME, is there some other game that fits those descriptions? So we can see if it really makes the game better.
Edit: except no, even ToME doesn't do point 5 - describing the typical monsters in the branch.

Dungeon Master

Posts: 634

Joined: Sunday, 22nd September 2013, 14:46

Post Friday, 4th September 2015, 13:12

Re: Learning from death

I think that Brogue is generally held up as a standard for the kind of game that is very open with information about internals.
User avatar

Abyss Ambulator

Posts: 1111

Joined: Monday, 18th March 2013, 23:23

Post Friday, 4th September 2015, 13:20

Re: Learning from death

I read the description for the subway and it totally did not prepare me for the percentage chance I would have of seeing a homeless guy that smelled like pee masturbating.

Real life needs better warning signs.

For this message the author prozacelf has received thanks: 2
Arrhythmia, WalkerBoh

Vestibule Violator

Posts: 1508

Joined: Monday, 21st November 2011, 07:40

Post Friday, 4th September 2015, 13:26

Re: Learning from death

Sandman, I'll really miss you; I've loved your input, and your invariably positive attitude to those trying to learn the game and get better at it. You're an inspiration in many ways. Don't take moderation too seriously; just back out of certain conversations if it's not actually going to accomplish anything. I get that retributive moderation is adversarial, but it's a really common paradigm anyhow. Don't let it get you down. ;)
Usual account: pblur on kelbi

For this message the author byrel has received thanks: 3
dowan, rockygargoyle, Sandman25

bel

Cocytus Succeeder

Posts: 2184

Joined: Tuesday, 3rd February 2015, 22:05

Post Friday, 4th September 2015, 13:41

Re: Learning from death

About transparency, Sil is pretty transparent (or marvinpa-sil actually, which has some improvements in interface and transparency department, like monster recall). Almost everything is defined in the manual (except smithing) and the game, and combat rolls are displayed in its own window.

I have only recently started Brogue, so I don't know too much about it.

About OP, it is a mish-mash of various things, some ok, some not. It is pretty clear that Sandman wants a more chess-like game, with more numbers, and less RNG luck (no matter how improbable it is: I remember some thread about being shafted multiple times in quick succession) Or take the current concern with max damage and HP, even though it is very rare that monsters will hit exactly for max damage, which is just below your HP). Actually one should compare twice the max damage with HP in many cases, because of energy randomization. Nobody makes calculations like this as matter of routine. It would be just silly, not to mention impractical.

Some other points about more transparent mechanics are good, especially HD-based mechanics.

Sandman has "last post"-ed before, so I don't take it too seriously. Though it would be sad if it happens for real this time.
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Friday, 4th September 2015, 15:38

Re: Learning from death

WalkerBoh wrote:In a lot of ways, the current situation regarding "paralysis by analysis" doesn't seem any worse than it would be with Sequell data available to the player in-game. This is especially true for new players, who have a ton of information to learn and absorb as it is.

I'm not sure that providing more information is really the solution to the already huge amount of info we expect new players to internalize. Instead, it seems like it'd be better to focus on creating more effective and comprehensive tutorials.

I think Sandman correctly points out that "memory demand" (good phrase) is made worse by the fact that players are not encouraged to efficiently learn from mistakes, and must learn and re-learn information based essentially on trial and error. This wouldn't be so bad except that, as Sandman points out, at some point in the game you're learning by trial and error after investing multiple hours into a character.

I think this is exaggerated; there are only so many kinds of mistakes that a player can make, after all, most of which you can encounter in the pre-Lair game. I'd argue that the average player will figure out things like "some monsters hit unusually hard" and "some monsters have attacks that don't target HP" or "some monsters can incapacitate me," as well as the in-game indicators that suggest that this stuff happens. This stuff probably isn't perfect, but,

I wouldn't want the equivalent of Sequell data popping up on the bottom of the screen during play. But having more data available on the xv menu regarding hidden monster mechanics, or explanations for how non-intuitive mechanics work (or hell, even as a sub-menu on the xv menu called "Detailed Information", for example) seems non-controversial to me.

The main reason I think it's "controversial," for lack of a better word, is that more in-game data isn't really all that necessary for the kinds of players interested in roguelikes.

I see two groups of players. The first group wants to figure everything out on their own, or at least broad swaths of the game. I've had more than one tiles player tell me they don't want to be spoiled, as crazy as that sounds to me. For these players, providing too much information would spoil the mystery of the game. The second group, which I think is typified by Sandman25, wants to know everything about the game. While it might be marginally more convenient if they could get all this information inside the game, these players are going to seek out information outside the game anyway; the game is never going to include things like .des files or other stuff you can get from the source, after all.

I'd argue that instead of begging the devs to add more info, people would have more success if they worked on tools like Chrome extensions or patches to tileschat in order to make querying the bots easier and more convenient. I'd also love more tools like LookupDB, which make it possible to get bot information without logging into IRC.

For this message the author archaeo has received thanks: 3
Lasty, Shard1697, WalkerBoh

Dungeon Master

Posts: 3618

Joined: Thursday, 23rd December 2010, 12:43

Post Friday, 4th September 2015, 16:46

Re: Learning from death

archeo is spot on, many thanks for that posting!

I think there are more types of players: for example, I don't really care about being spoiled, I just don't worry about numbers too much. To this day, I didn't know about the ogres 39 maxdamage, but I can deal with ogres using experience and feeling. Maybe that's a bad sign (i.e. I have played too much Crawl, so that I can now judge ogres intuitively, and I don't want others to take a shortcut by denying them numbers), but I could not get behind the Sil kind of transparency.

Brogue does it very well, as has been said time and again, but of course Brogue also has very few monsters. And means of attack. This makes it much easier to show the relevant numbers (by the way, Brogue hides health and always deals with percentages, which is a very clever move). Many will argue that Crawl has way too many monsters, and there's some truth to this, but then again, it's a baroque and big game, and that has a lot of appeal in its own (i.e. there are people who like that and play Crawl for it).

I am not sure if Sandman is aware how much walls of text would have to be displayed to get all of this information across. And you cannot make it automatic: whenever some monster effect is added, you have to maintain the info section. Also, monster health is randomised... should we get rid of this for transparency? I think instead of trying to explain all the details (which Crawl is not suited for), we should look hard for the most relevant piece of information. If it is max damage (might well be), then let's try that.

Vestibule Violator

Posts: 1508

Joined: Monday, 21st November 2011, 07:40

Post Friday, 4th September 2015, 16:53

Re: Learning from death

I would suggest max damage and actual max HP as the two most useful numbers currently hidden from monster stats.

I think the most useful number overall that's currently hidden is accuracy chance; how likely am I to hit this with my current weapon or spell?

Kudos for the Monster hex success rates by the way. I never used to use xv, as it just wasn't worth my time, and figured that putting more info there would be useless to me as I'd never see it. I now check chance to paralyze, banish, etc. maybe 10-20 times per game, and almost always make more informed decisions about tactics and strategy because of it. Which is awesome.
Usual account: pblur on kelbi
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Friday, 4th September 2015, 17:33

Re: Learning from death

byrel wrote:I would suggest max damage and actual max HP as the two most useful numbers currently hidden from monster stats.

Max damage might be helpful (though it would be cooler if it displayed it as a percentage of your character's current max HP), but a monster's max HP is a number that's kind of meaningless without knowing player damage. You could always implement it as a slider, something like "Durability ++....." or whatever.

I think the most useful number overall that's currently hidden is accuracy chance; how likely am I to hit this with my current weapon or spell?

It's only "hidden" in the same way that player damage is, i.e. there are enough popcorn fights in Crawl that you generally get a good handle on the relative accuracy/strength of your weapon in a bunch of relatively safe fights, imo.

Vestibule Violator

Posts: 1508

Joined: Monday, 21st November 2011, 07:40

Post Friday, 4th September 2015, 17:51

Re: Learning from death

archaeo wrote:
byrel wrote:I would suggest max damage and actual max HP as the two most useful numbers currently hidden from monster stats.

Max damage might be helpful (though it would be cooler if it displayed it as a percentage of your character's current max HP), but a monster's max HP is a number that's kind of meaningless without knowing player damage. You could always implement it as a slider, something like "Durability ++....." or whatever.

I think the most useful number overall that's currently hidden is accuracy chance; how likely am I to hit this with my current weapon or spell?

It's only "hidden" in the same way that player damage is, i.e. there are enough popcorn fights in Crawl that you generally get a good handle on the relative accuracy/strength of your weapon in a bunch of relatively safe fights, imo.


I'd be fine with a slider, but I think we already have the numbers to make some sense of max HP. Crawl's combat is swingy enough, it's not like you can predict the number of turns to kill with any precision anyhow. You know the base damage of your weapon, it's got some plusses, and presumably your skill and fighting and attributes affect that multiplicatively somehow. That's enough. The big point is boggarts have 25 HP, and caustic shrikes have 100. How much do wolf spiders have? About 55.

The point isn't to enable you to make accurate calculations about how long a fights going to last. The point is to give people some basis for estimating the tankiness of a monster we're meeting for the first time (Or the first few times really). We show AC and EV, but those are kind of meaningless without an HP pool. They're essentially multiplicative effective HP. Showing them DOES enable players to pick effective weapons (you shouldn't use rod of inaccuracy on mambas.)
Usual account: pblur on kelbi

Tomb Titivator

Posts: 853

Joined: Thursday, 29th August 2013, 18:39

Post Friday, 4th September 2015, 18:06

Re: Learning from death

It would actually provide a cool sense of progression if you filled in an (account-linked for online) bestiary as you played. You could gradually fill in more information on monsters as you fought them. Once you attack a crimson imp with fire, you figure out they are immune, and that appears on their monster page going forward. Obviously this is a huge amount of work and I'm not volunteering anyone else for it, but it would be a fun compromise between the two approaches as described by archaeo. (A sense of inter-character progress is kind of anathema in traditional roguelikes but I think current game devs have shown that people eat that sort of thing up.) Things like maximum damage would be hard to represent like that though - I don't have a way to address that.

I guess I'm just sort of okay with not knowing the numbers? It's hard to put myself in the perspective of a new player again - I know what fighting an ogre *feels* like, I don't need exact statistics. (Though ogres I actually have a heuristic for - in my head I don't melee an ogre until I have more than 40 hp, so I'm tickled that they do a max of 39 damage.)
User avatar

Ziggurat Zagger

Posts: 4478

Joined: Wednesday, 23rd October 2013, 07:56

Post Friday, 4th September 2015, 18:16

Re: Learning from death

johlstei wrote:t would actually provide a cool sense of progression if you filled in an (account-linked for online) bestiary as you played.

FWIW, ADoM has this kind of monster memory, but it is per game.
DCSS: 97:...MfCj}SpNeBaEEGrFE{HaAKTrCK}DsFESpHu{FoArNaBe}
FeEE{HOIEMiAE}GrGlHuWrGnWrNaAKBaFi{MiDeMfDe}{DrAKTrAMGhEnGnWz}
{PaBeDjFi}OgAKPaCAGnCjOgCKMfAEAtCKSpCjDEEE{HOSu
Bloat: 17: RaRoPrPh{GuStGnCa}{ArEtZoNb}KiPaAnDrBXDBQOApDaMeAGBiOCNKAsFnFlUs{RoBoNeWi

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 4th September 2015, 20:03

Re: Learning from death

Personally I don't want to see max damage numbers, nor would I want to see the exact max hps of a critter.

I would want to see an indication of whether a given critter has significantly more or less hps than others of it's type. I'd use "Robust' and "Fragile" for this purpose, aka:
"The robust orc priest comes into view"
"The fragile orc warrior hits you with the glaive!!"
That way I can derive that it's easier to kill a "fragile" ogre than it is to kill a "robust" one and it gives me more context to evaluate my damage, without throwing the exact statistics of every thing in my face, and it doesn't mislead me into thinking that two critters with different defenses, but equal hps are equally easy to kill.

For max damage I'd really just prefer if the 'dangerous' mechanism was better, max damage by itself without speed isn't a terribly useful mechanism (A black mamba is way more dangerous than it's max damage would imply on it's own) nor is dropping the number there without any context for things like GDR (and whether it applies to the 'max damage' stat).

What I really want to know is "is it possible for this thing to kill me" I'd make something "Incredibly dangerous" if it could kill me from full in one 10-aut action with no AC (with energy randomization) "very dangerous" if it could kill me from full in two 10 aut actions, etc." figuring a best-case roll on it's most damaging attack and using 0 AC no resistances and presuming it could get an extra attack from energy randomization if possible. The problem with representing this as a single abstract number is that "dangerousness" doesn't boil down to a single number the game's more complicated than that. I think tacking a max damage number on to critters will result in new players assuming two critters with the same "max damage" number will be equal in terms of dangerousness, which isn't really consistently true.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Vestibule Violator

Posts: 1601

Joined: Sunday, 14th July 2013, 16:36

Post Friday, 4th September 2015, 21:48

Re: Learning from death

archaeo wrote:I see two groups of players. The first group wants to figure everything out on their own, or at least broad swaths of the game. I've had more than one tiles player tell me they don't want to be spoiled, as crazy as that sounds to me. For these players, providing too much information would spoil the mystery of the game. The second group, which I think is typified by Sandman25, wants to know everything about the game. While it might be marginally more convenient if they could get all this information inside the game, these players are going to seek out information outside the game anyway; the game is never going to include things like .des files or other stuff you can get from the source, after all.

There's a wide area in-between. And the problem, IMO, is that crawl is way too combat-oriented and offers way too many options to really warrant being as opaque about combat as it is.

I would split the groups differently: those that don't want to figure things out, and those that do.
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Friday, 4th September 2015, 23:38

Re: Learning from death

I would split the groups differently: those that don't want to figure things out, and those that do.

But you still end up with the same kinds of player behaviors, wouldn't you? Those who don't want to figure things out don't, and those that do have access to literally all of the information being requested here and more. It just isn't right there, in the game window. If this stuff really is a problem, it's one involving ease of access, not an absence of information, and making it easier to access doesn't necessarily have to mean "plug it all into the game directly."

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 4th September 2015, 23:41

Re: Learning from death

archaeo wrote:
I would split the groups differently: those that don't want to figure things out, and those that do.

But you still end up with the same kinds of player behaviors, wouldn't you? Those who don't want to figure things out don't, and those that do have access to literally all of the information being requested here and more. It just isn't right there, in the game window. If this stuff really is a problem, it's one involving ease of access, not an absence of information, and making it easier to access doesn't necessarily have to mean "plug it all into the game directly."

I'm confused by this statement, isn't "plug it all into the game directly" the same as "in the game window" Or did you have a meaning I misunderstood?
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Zot Zealot

Posts: 1031

Joined: Friday, 26th April 2013, 19:52

Location: AZ, USA

Post Friday, 4th September 2015, 23:43

Re: Learning from death

Thanks for your response, archaeo. I don't disagree with you except for a couple points, which I"ll clarify.

archaeo wrote:
WalkerBoh wrote:I think Sandman correctly points out that "memory demand" (good phrase) is made worse by the fact that players are not encouraged to efficiently learn from mistakes, and must learn and re-learn information based essentially on trial and error. This wouldn't be so bad except that, as Sandman points out, at some point in the game you're learning by trial and error after investing multiple hours into a character.

I think this is exaggerated; there are only so many kinds of mistakes that a player can make, after all, most of which you can encounter in the pre-Lair game. I'd argue that the average player will figure out things like "some monsters hit unusually hard" and "some monsters have attacks that don't target HP" or "some monsters can incapacitate me," as well as the in-game indicators that suggest that this stuff happens. This stuff probably isn't perfect, but,

I may have exaggerated slightly, but you are oversimplifying I think. In this case, what I'm mainly referring to by "learning by trial and error" are monster mechanics. I don't think adding more numbers to the game is particularly helpful. Something like this is massive overkill to me:
...it will let you know that Ijyb has a wand of cold in inventory and the wand can do 1-37 damage and that you can get it to 50% by rC+, to 33% by rC++ and to 20% by rC+++

But every time you run into a monster with a new mechanic, you are learning by trial and error. Using constriction as an example, ball pythons teach you early on how the mechanic works, and enemies in Snake reinforce the concept later. You learn that constriction doesn't really matter much unless you're trying to escape - it's not imperative to break constriction as soon as you get the status. Then you make it to Depths for the first time after hours of gameplay and run into an octopode crusher. This is when you learn the hard way that when those guys constrict you, you can get thrown for massive amounts of damage. Whoops! Crawl has enough of these surprises and special monster abilities/tweaks that learning purely by trial and error (i.e., dying) can be very, very frustrating. And sometimes it leads you to draw conclusions that are quite counter-intuitive or just plain wrong, which you then unlearn again later with another frustrating death.

Now, I don't know offhand what the description for octopode crushers says - maybe it's very clear about the threat and differences, in which case that's great! Crawl has improved a lot in this regard over the last few versions (by showing monster spell sets, pan lord attributes, player ghost weapons, etc.). But to me, continuing to make "spoilery" information (for lack of a better word) available to players in-game is desirable.
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Friday, 4th September 2015, 23:58

Re: Learning from death

Siegurt wrote:I'm confused by this statement, isn't "plug it all into the game directly" the same as "in the game window" Or did you have a meaning I misunderstood?

Nope, you got it.

WalkerBoh wrote:Whoops! Crawl has enough of these surprises and special monster abilities/tweaks that learning purely by trial and error (i.e., dying) can be very, very frustrating. And sometimes it leads you to draw conclusions that are quite counter-intuitive or just plain wrong, which you then unlearn again later with another frustrating death.

Except that you've reduced Crawl's feedback mechanism to a binary live-or-die system, which isn't the case at all. Sure, in the early game, the time between a mistake and death is really short, but that length gets extended as the game goes on, especially for the combos we typically recommend to new players. By the time the game expects you to synthesize the information it taught you in the early game, you have way better options for responding to a new surprise.

And of course, if dying repeatedly to learn things isn't your cup of tea, there's a whole universe of information and advice accessible via your web browser, which players will be delighted to find out they already have access to by dint of being able to download Crawl in the first place. In my experience, half the fun of a new roguelike is exploring the spoilers; I like NetHackWiki more than I like NetHack most of the time.

Crawl has improved a lot in this regard over the last few versions (by showing monster spell sets, pan lord attributes, player ghost weapons, etc.). But to me, continuing to make "spoilery" information (for lack of a better word) available to players in-game is desirable.

I'm not saying I don't want Crawl to include any more information; I agree that the stuff added in the last few versions has been good and helpful, and more of that would be great. Just that the wholesale inclusion of info that more than a dozen people "thanked" in the OP would almost certainly be taking things way too far in the other direction.

Spider Stomper

Posts: 220

Joined: Sunday, 26th July 2015, 15:38

Post Saturday, 5th September 2015, 00:43

Re: Learning from death

In the same vein:

Good roguelike has transparent, easy to remember formulas telling you exactly what you get for investing in skills and spells.

Bad roguelikes require several page long wiki articles to describe the basics of a single skill or even aspect of one part of a skill like spell power, as well as unintuitive rediculous concepts like weapon skills essentially becoming worthless past a magic number but allow you to train past anyways, and spellpower being required to not starve to death.

For this message the author greedo has received thanks: 2
all before, Sandman25

Vestibule Violator

Posts: 1601

Joined: Sunday, 14th July 2013, 16:36

Post Saturday, 5th September 2015, 01:04

Re: Learning from death

archaeo wrote:
I would split the groups differently: those that don't want to figure things out, and those that do.

But you still end up with the same kinds of player behaviors, wouldn't you? Those who don't want to figure things out don't, and those that do have access to literally all of the information being requested here and more. It just isn't right there, in the game window. If this stuff really is a problem, it's one involving ease of access, not an absence of information, and making it easier to access doesn't necessarily have to mean "plug it all into the game directly."

Not necessarily.

I assert the majority of the people who want things like raw numbers do so not because they actually care about the raw numbers, but because they feel the need to know something that the game doesn't tell them, and knowing the raw numbers and mechanics is the only way they have to obtain that information with any accuracy.

e.g. when wielding a +2 broad axe, a character comes across a +0 hand axe of electrocution. The first category just does something arbitrary; they're content to go with whatever choice suits their whims or superstition. The second category actually wants to consider which weapon they can use. They generally don't want "all the information and more" -- they just want something that tells them how effective each weapon is, or barring that enough feedback to see which is effective. For the most part, looking up "all the information and more" is done only because the only way to obtain the information that actually matters is to work it out from the raw data and mechanics. (fsim works for most things, but I think most people don't know about it, and even for those that do it is rather inconvenient to use and difficult to find any documentation for it)

Hopefully, progress can be made on this particular front in the wake of the double damage bug, now that the myth that "if the difference was relevant, players could tell" has been mostly dispelled.

For this message the author Hurkyl has received thanks:
Sandman25
User avatar

Barkeep

Posts: 4435

Joined: Tuesday, 11th January 2011, 12:28

Post Saturday, 5th September 2015, 01:06

Re: Learning from death

Siegurt wrote:I would want to see an indication of whether a given critter has significantly more or less hps than others of it's type. I'd use "Robust' and "Fragile" for this purpose

I would much rather that monsters of a type all had the same number of HP, even if it was always the max as opposed to the average.
I am not a very good player. My mouth is a foul pit of LIES. KNOW THIS.

For this message the author njvack has received thanks:
Sandman25
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Saturday, 5th September 2015, 02:11

Re: Learning from death

Hurkyl wrote:They generally don't want "all the information and more" -- they just want something that tells them how effective each weapon is, or barring that enough feedback to see which is effective. For the most part, looking up "all the information and more" is done only because the only way to obtain the information that actually matters is to work it out from the raw data and mechanics. (fsim works for most things, but I think most people don't know about it, and even for those that do it is rather inconvenient to use and difficult to find any documentation for it)

I wouldn't mind the line about a weapon's brand in xv being a little more specific, but generally, weapons already give you all the pertinent numbers. Usually, if you're interested in seeing what weapon works better, you can always just bash some monsters with it, but other than the trickiest details about how different brands work, it's usually pretty intuitive as to what's going to happen when you use a weapon brand.

Let me put it this way: I'm not at all opposed to the game becoming clearer and more intuitive. I just think there are way better ways to do that than dumping all the game's numbers on the player and expecting them to figure it out.

Hopefully, progress can be made on this particular front in the wake of the double damage bug, now that the myth that "if the difference was relevant, players could tell" has been mostly dispelled.

Can we retire this meme? There was a bug in the game that wasn't easy to see, the players and devs found it (during the tournament they hold in part to shake out the last bugs), and then they fixed it. It isn't some kind of mythic event that encapsulates Crawl's woeful and obscure complexity.

Vestibule Violator

Posts: 1601

Joined: Sunday, 14th July 2013, 16:36

Post Saturday, 5th September 2015, 02:50

Re: Learning from death

archaeo wrote:
Hurkyl wrote:They generally don't want "all the information and more" -- they just want something that tells them how effective each weapon is, or barring that enough feedback to see which is effective. For the most part, looking up "all the information and more" is done only because the only way to obtain the information that actually matters is to work it out from the raw data and mechanics. (fsim works for most things, but I think most people don't know about it, and even for those that do it is rather inconvenient to use and difficult to find any documentation for it)

I wouldn't mind the line about a weapon's brand in xv being a little more specific, but generally, weapons already give you all the pertinent numbers.

By far the most pertinent number is AvEffDam, something which currently does not appear anywhere except fsim.


Hopefully, progress can be made on this particular front in the wake of the double damage bug, now that the myth that "if the difference was relevant, players could tell" has been mostly dispelled.

Can we retire this meme? There was a bug in the game that wasn't easy to see, the players and devs found it (during the tournament they hold in part to shake out the last bugs), and then they fixed it. It isn't some kind of mythic event that encapsulates Crawl's woeful and obscure complexity.

Clearly, it is too early to stop bringing this up, given the claim
Usually, if you're interested in seeing what weapon works better, you can always just bash some monsters with it,

For this message the author Hurkyl has received thanks: 3
duvessa, rockygargoyle, Sandman25

Abyss Ambulator

Posts: 1217

Joined: Sunday, 14th April 2013, 04:01

Post Wednesday, 9th September 2015, 13:28

Re: Learning from death

Firstly, sorry to hear you are leaving :(

Secondly, Crawl development HAS been trending towards more information- you can see MR and speed by viewing monsters, Hexes have the hit chance.

Thirdly, Crawl's primary form of learning at least the first time is "mess with this" or "don't mess with this" and I actually kinda like that. I don't need to know my chances of messing with a caustic strike the second time I meet them because I've learned to gtfo. Over time, I might try to fight, but I'd almost rather not know than go "I have a 75% chance of winning" and then dying.
Three wins: Gargoyle Earth Elementalist of Ash, Ogre Fighter of Ru, Deep Dwarf Fighter of Makhleb (0.16 bugbuild :( )

For this message the author TeshiAlair has received thanks:
Sandman25

Mines Malingerer

Posts: 40

Joined: Thursday, 1st May 2014, 20:49

Post Thursday, 10th September 2015, 02:55

Re: Learning from death

There was some old maxim rom a forum I used to visit that stated something along the lines of "anyone who cares enough to post a "I'm never coming back message" will inevitably return.

bel

Cocytus Succeeder

Posts: 2184

Joined: Tuesday, 3rd February 2015, 22:05

Post Thursday, 10th September 2015, 11:44

Re: Learning from death

Hurkyl wrote:Clearly, it is too early to stop bringing this up, given the claim
Usually, if you're interested in seeing what weapon works better, you can always just bash some monsters with it,

This has been argued against so many times that it is rather silly to keep saying this. The double damage bug was between different versions. The comparison of weapons is within a single game. I am pretty sure an unspoiled player can figure out that an eveningstar does much more damage than a dagger in melee in the same game. There is no point at all in comparing apples and elephants.

Really, less opacity and more sanity in crawl's damage formulae would help, but silly arguments only weaken the case.

Vestibule Violator

Posts: 1601

Joined: Sunday, 14th July 2013, 16:36

Post Thursday, 10th September 2015, 12:37

Re: Learning from death

bel wrote:I am pretty sure an unspoiled player can figure out that an eveningstar does much more damage than a dagger in melee in the same game.

That's mostly because the dagger gets crippled by AC. (also, this judgement is aided by confirmation bias (daggers sound weak) and the illusion caused by spreading the damage across more attacks) (I assume you don't mean for the dagger to be electrocution or pain, in which case it can actually outdamage the eveningstar in reasonable circumstances)

bel

Cocytus Succeeder

Posts: 2184

Joined: Tuesday, 3rd February 2015, 22:05

Post Thursday, 10th September 2015, 15:40

Re: Learning from death

Hurkyl wrote:That's mostly because the dagger gets crippled by AC. (also, this judgement is aided by confirmation bias (daggers sound weak) and the illusion caused by spreading the damage across more attacks) (I assume you don't mean for the dagger to be electrocution or pain, in which case it can actually outdamage the eveningstar in reasonable circumstances)

I am indeed talking about plain, unbranded, 0 ench weapons.

I am pretty sure one can do this for mace vs eveningstar (a HuFi with 15ish M&F/fighting does twice the damage against an ogre for the latter). So I don't think that confirmation bias is operative here. I don't think an eveningstar even exists in reality: it's just an joke on morning star, which does.

Halls Hopper

Posts: 79

Joined: Thursday, 7th July 2011, 14:54

Post Thursday, 10th September 2015, 16:23

Re: Learning from death

I've been playing this game for years and still have no feel for the answer to "is a +6 flail better than a +9 mace or a +1 morningstar?" Does the answer to that question changes at different skill levels?

"Swing the weapon a few times at enemies and see" is not an adequate answer to this issue. In practice there are two situations that occur. 1) I'm swinging at an enemy that poses no threat and won't provide an adequate gauge of the weapons power 2) I'm swinging at an enemy that is potentially threatening and i don't have time to experiment. Different swing speeds, different max damage, no reported information on DPS, only reported single swing reduction in enemy health bar and auts taken. Conclusion is that there is no way to currently gauge which weapon is best.

For this message the author plantaspoon has received thanks: 4
Hurkyl, rockygargoyle, Sandman25, Shard1697

bel

Cocytus Succeeder

Posts: 2184

Joined: Tuesday, 3rd February 2015, 22:05

Post Thursday, 10th September 2015, 17:08

Re: Learning from death

You have misunderstood the discussion. The question is not which one is best, but whether the two weapons are hugely different in quality. Your example (+9 mace vs +1 morningstar), the difference is about 10% (at 15ish M&F/Fighting).

If the difference was 100%, as the misleading references to double damage bug imply, that wouldn't happen. You would very easily know which was better, even if one accepts all your caveats.

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Thursday, 10th September 2015, 17:12

Re: Learning from death

plantaspoon wrote:I've been playing this game for years and still have no feel for the answer to "is a +6 flail better than a +9 mace or a +1 morningstar?" Does the answer to that question changes at different skill levels?

"Swing the weapon a few times at enemies and see" is not an adequate answer to this issue. In practice there are two situations that occur. 1) I'm swinging at an enemy that poses no threat and won't provide an adequate gauge of the weapons power 2) I'm swinging at an enemy that is potentially threatening and i don't have time to experiment. Different swing speeds, different max damage, no reported information on DPS, only reported single swing reduction in enemy health bar and auts taken. Conclusion is that there is no way to currently gauge which weapon is best.

Ah but when you can't tell common logic is "they are close enough that it just doesn't matter, pick the faster one early or the heavier hitting one later when swinging against high ac" which is often close enough to true that other considerations than weapon damage should be forefront on your mind.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Halls Hopper

Posts: 79

Joined: Thursday, 7th July 2011, 14:54

Post Thursday, 10th September 2015, 18:05

Re: Learning from death

Siegurt wrote:
plantaspoon wrote:I've been playing this game for years and still have no feel for the answer to "is a +6 flail better than a +9 mace or a +1 morningstar?" Does the answer to that question changes at different skill levels?

"Swing the weapon a few times at enemies and see" is not an adequate answer to this issue. In practice there are two situations that occur. 1) I'm swinging at an enemy that poses no threat and won't provide an adequate gauge of the weapons power 2) I'm swinging at an enemy that is potentially threatening and i don't have time to experiment. Different swing speeds, different max damage, no reported information on DPS, only reported single swing reduction in enemy health bar and auts taken. Conclusion is that there is no way to currently gauge which weapon is best.

Ah but when you can't tell common logic is "they are close enough that it just doesn't matter, pick the faster one early or the heavier hitting one later when swinging against high ac" which is often close enough to true that other considerations than weapon damage should be forefront on your mind.


Let put this another way. Assume we are dealing with type 2 fights, these are fights which I have a reasonable chance at success, and if it is not going my way I can escape and reset the fight. Weapon "A" is 10% worse than weapon "B", but there is no way for me to tell that. A series of type 2 fights using weapon A has a greater chance of leading to an uncomfortable situation for me and has a long term tax on my consumables. I'm being punished by bad UI, not bad decisions.

Dungeon Master

Posts: 3160

Joined: Sunday, 5th August 2012, 14:52

Post Thursday, 10th September 2015, 18:21

Re: Learning from death

1) If two weapons are similar enough that you can't tell them apart in melee, the difference in their average damage output is small.
2) In Crawl, output of weapon attacks is quite random.
3) The degree of randomness in the outcomes of weapon attacks overshadows the difference between two similar-seeming weapons.
4) If you are in a situation where using the lesser of two weapons that have a 10% difference in average damage over time significantly decreases your life expectancy, then you are also in a situation where simply getting unlucky could kill you regardless of the weapon you are wielding.
5) If you are in a situation where getting unlucky could kill you, you should do what you can to not be in that situation or significantly alter the odds of a fatal streak of bad luck happening.
6) Knowing which of two very similar weapons is better has very little impact on your survival if you can't tell the difference between those two weapons.

For this message the author Lasty has received thanks:
nago

Halls Hopper

Posts: 79

Joined: Thursday, 7th July 2011, 14:54

Post Thursday, 10th September 2015, 18:45

Re: Learning from death

Lasty wrote:1) If two weapons are similar enough that you can't tell them apart in melee, the difference in their average damage output is small.
2) In Crawl, output of weapon attacks is quite random.
3) The degree of randomness in the outcomes of weapon attacks overshadows the difference between two similar-seeming weapons.
4) If you are in a situation where using the lesser of two weapons that have a 10% difference in average damage over time significantly decreases your life expectancy, then you are also in a situation where simply getting unlucky could kill you regardless of the weapon you are wielding.
5) If you are in a situation where getting unlucky could kill you, you should do what you can to not be in that situation or significantly alter the odds of a fatal streak of bad luck happening.
6) Knowing which of two very similar weapons is better has very little impact on your survival if you can't tell the difference between those two weapons.


You've made my points exactly and drawn a different conclusion. just take it one tiny step further and we might agree.
I'm with you up until 4)
re 4) getting unlucky means using a consumable
re 5) Consumables significantly alter the battle in your favor
re 6) Wielding a worse weapon is a tax on my consumables which does have long term impact on my survivability, but not one that is detectable via experimentation due to bad UI described in points 1 thru 3.

For this message the author plantaspoon has received thanks: 2
Hurkyl, Sandman25
User avatar

Barkeep

Posts: 1788

Joined: Saturday, 29th June 2013, 16:52

Post Thursday, 10th September 2015, 18:54

Re: Learning from death

Is it really a "bad UI" problem, though? It seems like the answer to "It's hard to tell weapons apart" is to have fewer indistinguishable weapons, not to crowd the UI with more numbers.

For this message the author archaeo has received thanks: 2
Hurkyl, plantaspoon

Halls Hopper

Posts: 79

Joined: Thursday, 7th July 2011, 14:54

Post Thursday, 10th September 2015, 19:26

Re: Learning from death

archaeo wrote:Is it really a "bad UI" problem, though? It seems like the answer to "It's hard to tell weapons apart" is to have fewer indistinguishable weapons, not to crowd the UI with more numbers.


It doesn't have to be a TON of numbers, just modifying the weapon description to say "with your current skill it does an average of x damage against an unarmored target and has an average swing speed of x"

then add a line in the XV menu that says with your current weapon you have X chance to hit this monster.

No additional messages needed in the combat log, only slight modifications to the information that is already presented to the player.

For this message the author plantaspoon has received thanks:
Sandman25
Next

Return to Game Design Discussion

Who is online

Users browsing this forum: No registered users and 25 guests

cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.