@dev team: if you had a time machine


If it doesn't fit anywhere else, it belongs here. Also, come here if you just need to get hammered.

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Wednesday, 22nd February 2017, 07:32

@dev team: if you had a time machine

Hello,

I'm starting to develop a new open-source roguelike from scratch and DCSS is a big inspiration as when I was thinking through what a good game would be like I came up with guidelines that are very similar to your design philosophy.

Is there some non-obvious advice you could give to a newbie roguelike developer? (I'm a professional non-game developer in my non-spare time)

If you could change anything in DCSS that's currently not possible due to scope or other reasons, what would it be?

Thanks

(Mod edit: moved out of GDD, since the discussion isn't Crawl-specific).

Dungeon Master

Posts: 3618

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

Post Wednesday, 22nd February 2017, 11:06

Re: @dev team: if you had a time machine

I wrote the philosophy section for Crawl. Back then, this seems to have been something of a novelty because people still mention it. It's one of my main contributions to Crawl.

I don't post on the tavern anymore, but I make an exception for you.

If I'd develop a roguelike from scratch, here are some staples I would avoid at all costs:

  1. No resting
    Regaining HP (or MP or anything else) over time, i.e. "resting" is a very clumsy mechanic. I talked to Jeff Lait (of POWDER and a million other good roguelikes) and we agreed on that one right away. There's a whole bunch of alternatives, and many newer roguelikes explore them: healing from exploration; healing from consumables only; HP=1 i.e. one hit and you're dead.
  2. A working clock
    Crawl does not really have one, and it's noticeable. For a really good game with a tight food clock, check Brogue. (This game is a gem in general: it's really, really good about forcing the player to make meaningful decisions. It's history is interesting, too, because Pender started with experience and levels, and then scrapped them. Brogue also favours terrain and traps more than monsters/combat, at least when compared to normal roguelikes.)
  3. Tactics vs strategy
    This only applies if your game is big enough in scope; Crawl qualifies. The problem I have in mind is that we use the same map (with 1x1 tiles) for tactical representation (combat, exploration) and for travel. This is one underlying reason that leads to luring/kiting (where you extend the tactical space to a level where tedium creeps in). For a game that somehow addresses it, check 100 Rogues by Keith Burgun. There, all combat is restricted to rooms. (It doesn't do anything about boring travel, though.)
    Crawl is a dinosaur, and mostly papers over the issue using the travel and search commands. It's good enough, but it could be much better. :)
    edit: By the way, resting over time is similar to this: it comes from using the same (smallest!) unit for battle (the turn) as elsewhere. Compare with a book or a movie: you'd zoom in (high resolution, spatially and for time, e.g. slow motion) for action scenes, but you'd zoom out for everything else. A game like Crawl only has one scale (the tile, the turn/aut), and it's too big for this, in my opinion.
There's a bunch of minor stuff I regret, like trying to fix Nemelex back in the day (around 0.4 or so). I did not even consider deicide (no idea if that idea would've been accepted) -- this issue is mostly solved by making decks Nemelex-only and much slimmer. I still wouldn't mourn if Nemelex had an accident but so many people put work into the god that I couldn't propose it myself.

If you have more questions, please send me a private message. Good luck with your game!

edit a few days later: It's not something as close to me as the above issues, but the dungeon structure is problematic, too. Among traditional roguelikes, you get the cleanest design like this: unbranched (linear dungeon) without backtracking -- incidentally (and probably accidentally, for technical reasons), the original Rogue did it like this.
Last edited by dpeg on Thursday, 23rd February 2017, 21:08, edited 2 times in total.

For this message the author dpeg has received thanks: 13
Brannock, Doesnt, MainiacJoe, neverEnough, njvack, ohmi, PleasingFungus, Sar, Shard1697, Shtopit and 3 more users

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Wednesday, 22nd February 2017, 18:38

Re: @dev team: if you had a time machine

Thanks dpeg!

That's really good food for thought. Hopefully I am able to come up with interesting solutions.

Ziggurat Zagger

Posts: 5382

Joined: Friday, 25th November 2011, 07:36

Post Wednesday, 22nd February 2017, 19:50

Re: @dev team: if you had a time machine

Honestly while I understand the arguments against them, resting and the lack of a clock never really bothered me all that much. If healing is overall finite then you have to treat every encounter very carefully, which seems like it would get tedious, although I suppose you could solve this by having fewer encounters/a shorter game.

Anyways: my recommendation: Transparent combat/growth mechanics. Make it clear what weapons do; display values that players need to figure out what's going on. Avoid "hinting" at values like when crawl tells you that you did "!!!" damage rather than just saying "46". When a player gets stronger, make it clear how they got stronger and what they can do with their new powers (ie, you gained 3 dex, this gives you +5% chance to hit and 3% chance to dodge".

It'll drastically lower the learning curve and help people get into your game easier.

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Wednesday, 22nd February 2017, 19:53

Re: @dev team: if you had a time machine

Please welcome tasonir, the devteam's newest member!

For this message the author duvessa has received thanks:
tasonir

Tartarus Sorceror

Posts: 1822

Joined: Thursday, 31st May 2012, 15:45

Post Wednesday, 22nd February 2017, 21:13

Re: @dev team: if you had a time machine

I've seen this sort of hard-number feedback in some people's games when spectating online, and would like to know what sort of rcfile settings they're using to get it. Probably a lot of lua.
Won (52). Remaining (15): 5 species: Ba, Fe, Mu, Na, Op; 5 Backgrounds: AM, Wr, Su, AE, Ar; 5 gods: Jiyv, newNem, WJC, newSif, newFedh

Shoals Surfer

Posts: 276

Joined: Sunday, 6th November 2016, 19:19

Post Wednesday, 22nd February 2017, 22:01

Re: @dev team: if you had a time machine

MainiacJoe wrote:I've seen this sort of hard-number feedback in some people's games when spectating online, and would like to know what sort of rcfile settings they're using to get it. Probably a lot of lua.


You mean the 'you take x damage' prompts?
if so, the rcfile settings are:
  Code:
<
  function ready()
    AnnounceDamage()
    --SpoilerAlert()
    OpenSkills()
  end
>

####################
# Included RcFiles #
####################

# Just Message colors
# include += HDAMessageColors.rc

# Just Item colors
# include += HDAItemColors.rc

# Monster warning messages -- disabled by default
# Uncomment the next line and the call to SpoilerAlert() in the ready function to use it
include += SpoilerAlert.rc

# The damage announcement messages
include += HDamage.rc

# Force More prompts
#  include += HDAForceMore.rc

# Message and Item colors
include += HDAColors.rc


I commented out the force more prompts, because it's annoying to play with uskayaw and deal with a billion force mores everytime your piety drops or increases.
16/26 on the way to GreaterOctopode (Win all backgrounds as an Octopode)

Progress so far : OpFi, OpGl, OpWn, OpAr, OpCK, OpMo, OpBe, OpHu, OpVM, OpAM, OpWr, OpFE, OpEE, OpNe, OpTm, OpSk

For this message the author Vajrapani has received thanks:
MainiacJoe

Tartarus Sorceror

Posts: 1822

Joined: Thursday, 31st May 2012, 15:45

Post Wednesday, 22nd February 2017, 22:05

Re: @dev team: if you had a time machine

I don't have my offline build with me at work to check: are these extra .rc files part of the tiles zip that I download from the Crawl homepage, or do I need to find them somewhere?
Won (52). Remaining (15): 5 species: Ba, Fe, Mu, Na, Op; 5 Backgrounds: AM, Wr, Su, AE, Ar; 5 gods: Jiyv, newNem, WJC, newSif, newFedh

Shoals Surfer

Posts: 276

Joined: Sunday, 6th November 2016, 19:19

Post Wednesday, 22nd February 2017, 22:10

Re: @dev team: if you had a time machine

I only play online so I'm not too sure, but those rc files are from a guy called HilariousDeathArtist, here's his github page https://github.com/HilariousDeathArtist/DCSSConfigFile
16/26 on the way to GreaterOctopode (Win all backgrounds as an Octopode)

Progress so far : OpFi, OpGl, OpWn, OpAr, OpCK, OpMo, OpBe, OpHu, OpVM, OpAM, OpWr, OpFE, OpEE, OpNe, OpTm, OpSk

Spider Stomper

Posts: 236

Joined: Saturday, 2nd July 2016, 13:16

Post Wednesday, 22nd February 2017, 22:24

Re: @dev team: if you had a time machine

MainiacJoe wrote:I've seen this sort of hard-number feedback in some people's games when spectating online, and would like to know what sort of rcfile settings they're using to get it. Probably a lot of lua.
Maybe just hellcrawl?

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Wednesday, 22nd February 2017, 23:04

Re: @dev team: if you had a time machine

dpeg wrote:If you have more questions, please send me a private message. Good luck with your game!

I need to get more posts before I am able to PM, so I'll be replying here for now if you don't mind

dpeg wrote:No resting
Regaining HP (or MP or anything else) over time, i.e. "resting" is a very clumsy mechanic. I talked to Jeff Lait (of POWDER and a million other good roguelikes) and we agreed on that one right away. There's a whole bunch of alternatives, and many newer roguelikes explore them: healing from exploration; healing from consumables only; HP=1 i.e. one hit and you're dead.

I believe that having varying Max HP/MP between characters and several ways of HP and MP gain (with varying speeds and costs) adds tactical complexity and potentially meaningful choices. To avoid tedium the slowest method should be something that a player would do anyway during normal play. Some candidates:

  1. Eating: in practice wouldn't be too different than resting and could cause to unintuitively avoid eating when hungry but at full HP.
  2. Exploration: while obviously rewards the player to take on risks it can force one to leave safe parts unexplored and backtrack to be safe
  3. XP gain: if xp is also gained from exploration it has the same benefits, but with many other XP sources I believe it would be less likely to encourage unintuitive behavior. To avoid farming all XP sources will max out after some point.
dpeg wrote:A working clock
Crawl does not really have one, and it's noticeable. For a really good game with a tight food clock, check Brogue. (This game is a gem in general: it's really, really good about forcing the player to make meaningful decisions. It's history is interesting, too, because Pender started with experience and levels, and then scrapped them. Brogue also favours terrain and traps more than monsters/combat, at least when compared to normal roguelikes.)

I'm familiar with Brogue and while it's very cool and innovative it doesn't represent the gameplay I currently envision. While having a clock that encourages optimal gameplay is good I don't think it should be too punishing to completionists or players that enjoy messing around (this is probably a backlash against all the survival games that have been coming out recently).

I'm actually considering removing hunger to reduce micro-managment and only keeping food for intrinsics and other special properties.

The main clock will be the score as it will decrease for every action the character takes. XP and items will not influence the score.
The goal of the game will be to kill or disable 'The Endboss'. Initially TEB will be pretty much invincible, but for every 'Rune' a player collects TEB will have a predetermined penalty applied (e.g. fire immunity removed or armor decreased). The catch is that for every rune collected the score will be further decreased (in addition to the turns penalty).
The only way of improving the score is by maintaining conducts.

dpeg wrote:Tactics vs strategy
This only applies if your game is big enough in scope; Crawl qualifies. The problem I have in mind is that we use the same map (with 1x1 tiles) for tactical representation (combat, exploration) and for travel. This is one underlying reason that leads to luring/kiting (where you extend the tactical space to a level where tedium creeps in). For a game that somehow addresses it, check 100 Rogues by Keith Burgun. There, all combat is restricted to rooms. (It doesn't do anything about boring travel, though.)
Crawl is a dinosaur, and mostly papers over the issue using the travel and search commands. It's good enough, but it could be much better. :)

Do you mean that in 100 Rogues the player can't travel to the previous room?
IMHO the commands are a good way to deal with the tedium while still giving the options to freely travel and backtrack. The only annoyance I have is stair dancing, perhaps it could be fixed by allowing monsters to attack the player from other level or allow them to push through if strong/agile enough.

dpeg wrote:There's a bunch of minor stuff I regret, like trying to fix Nemelex back in the day (around 0.4 or so). I did not even consider deicide (no idea if that idea would've been accepted) -- this issue is mostly solved by making decks Nemelex-only and much slimmer. I still wouldn't mourn if Nemelex had an accident but so many people put work into the god that I couldn't propose it myself.

tasonir wrote:Anyways: my recommendation: Transparent combat/growth mechanics. Make it clear what weapons do; display values that players need to figure out what's going on. Avoid "hinting" at values like when crawl tells you that you did "!!!" damage rather than just saying "46". When a player gets stronger, make it clear how they got stronger and what they can do with their new powers (ie, you gained 3 dex, this gives you +5% chance to hit and 3% chance to dodge".

It'll drastically lower the learning curve and help people get into your game easier.

Yes, I find this type of randomness/obscurity annoying. I plan on having fixed damage values, non-random effects and no spell cast failure chance. I'm still thinking about how to handle hit/miss.
Also the interface will have easy access to information that you would normally have to remember or search on wiki.

Vaults Vanquisher

Posts: 463

Joined: Monday, 20th July 2015, 04:01

Post Thursday, 23rd February 2017, 06:09

Re: @dev team: if you had a time machine

tantanoid wrote:I believe that having varying Max HP/MP between characters and several ways of HP and MP gain (with varying speeds and costs) adds tactical complexity and potentially meaningful choices. To avoid tedium the slowest method should be something that a player would do anyway during normal play. Some candidates:

  1. Eating: in practice wouldn't be too different than resting and could cause to unintuitively avoid eating when hungry but at full HP.
  2. Exploration: while obviously rewards the player to take on risks it can force one to leave safe parts unexplored and backtrack to be safe
  3. XP gain: if xp is also gained from exploration it has the same benefits, but with many other XP sources I believe it would be less likely to encourage unintuitive behavior. To avoid farming all XP sources will max out after some point.


1. This can work if food is limited (which I assume it is lol), but there's something I don't like about it. It just feels like an annoyance to have to make that decision. It's not a meaningful choice that adds enjoyment to the game, if you get what I'm saying. It's stressful.
2. An interesting thought but this introduces tedium as well.
3. HP healing by XP gain be exciting, but it can also be very agitating to the player. I think most people play crawl to relax; if you use this it should be in a game where people actually want to get excited, aka something that isn't turn-based.

I'm familiar with Brogue and while it's very cool and innovative it doesn't represent the gameplay I currently envision. While having a clock that encourages optimal gameplay is good I don't think it should be too punishing to completionists or players that enjoy messing around (this is probably a backlash against all the survival games that have been coming out recently).


The clock can be used as an answer to the HP/MP regain thing, if combined with resting, although that introduces tedium and complicated, unfun thought processes involving how much the player would want to rest.

Alternatively, you can combine it with the "gain XP for HP" thing to amplify the effect.

I'm actually considering removing hunger to reduce micro-managment and only keeping food for intrinsics and other special properties.

The main clock will be the score as it will decrease for every action the character takes. XP and items will not influence the score.
The goal of the game will be to kill or disable 'The Endboss'. Initially TEB will be pretty much invincible, but for every 'Rune' a player collects TEB will have a predetermined penalty applied (e.g. fire immunity removed or armor decreased). The catch is that for every rune collected the score will be further decreased (in addition to the turns penalty).
The only way of improving the score is by maintaining conducts.


I really like both of these. Hey dpeg can the number of runes affect Zot somehow?

tasonir wrote:Anyways: my recommendation: Transparent combat/growth mechanics. Make it clear what weapons do; display values that players need to figure out what's going on. Avoid "hinting" at values like when crawl tells you that you did "!!!" damage rather than just saying "46". When a player gets stronger, make it clear how they got stronger and what they can do with their new powers (ie, you gained 3 dex, this gives you +5% chance to hit and 3% chance to dodge".

It'll drastically lower the learning curve and help people get into your game easier.

Yes, I find this type of randomness/obscurity annoying. I plan on having fixed damage values, non-random effects and no spell cast failure chance. I'm still thinking about how to handle hit/miss.
Also the interface will have easy access to information that you would normally have to remember or search on wiki.


It's there for a reason. It does feel less of a drain the less numbers that are displayed. There was a lot of thought put into the concepts that enemy AC, EV, and MR be obscure. Only recently has max HP been revealed, and that's honestly just there to give the player the ability to better analyze the enemy (ie it tells you that slimes and phantoms have a crapton of health). All that extra mumbo jumbo makes it tempting and subsequently very annoying to calculate stuff. This sort of thing works in Fire Emblem because calculating that stuff is simple addition/subtraction math, but not here.

That being said, I wouldn't mind seeing some accuracy numbers, nor some damage numbers for the player, if only to see the difference between archmagi and no archmagi.

If the player really wants extremely descriptive values for leveling up skills and xp level, you should instead include them as part of a command. Like pressing '@' would display what you would get from your next level.

Random effects, including variable damage, introduce 'complications' to an otherwise completely predictable situation. They make the game interesting and I don't think blanket removal is a good answer here. Getting rid of spell cast failure chance I can get behind though. It's a case-by-case basis.

For hit/miss, you ideally want damage between player turns to fall on a bell-curve. This is done in crawl by having multiple monsters attack at once (albeit somewhat lopsidedly) and having rolls for damage. If you don't like that single attacks are 100% hit or miss, you could do something similar to Pillars of Eternity and have more than one accuracy category. PoE has 4: Miss, Graze, Hit, and Crit, which multiply effects by 0%, 50%, 100%, and some number > 100%, respectively. I use the word effects because it's not just damage that gets multiplied. Status effect duration and magnitude are also affected by the accuracy categories, with a critical bleeding status effect being way worse than a grazed bleeding status effect.

@dpeg: multiple accuracy categories might actually make accuracy in crawl worth thinking about, although if done badly it could add another complicated layer.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Thursday, 23rd February 2017, 06:33

Re: @dev team: if you had a time machine

If you want to see how heal by XP feels, you can try Deep Dwarf of Makhleb. Multiple travels between Tomb and other branches, praying to RNG to meet popcorn instead of dangerous monsters in Pan while wounded, "meaningful" decisions to explore Pan while being below 100 HP etc.
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

For this message the author VeryAngryFelid has received thanks:
Rast

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Thursday, 23rd February 2017, 06:41

Re: @dev team: if you had a time machine

PowerOfKaishin wrote:3. HP healing by XP gain be exciting, but it can also be very agitating to the player. I think most people play crawl to relax; if you use this it should be in a game where people actually want to get excited, aka something that isn't turn-based.

As a fallback there can be diablo-style shrines that replenish HP/MP, but don't recharge after being used.

PowerOfKaishin wrote:It's there for a reason. It does feel less of a drain the less numbers that are displayed. There was a lot of thought put into the concepts that enemy AC, EV, and MR be obscure. Only recently has max HP been revealed, and that's honestly just there to give the player the ability to better analyze the enemy (ie it tells you that slimes and phantoms have a crapton of health). All that extra mumbo jumbo makes it tempting and subsequently very annoying to calculate stuff. This sort of thing works in Fire Emblem because calculating that stuff is simple addition/subtraction math, but not here.

That being said, I wouldn't mind seeing some accuracy numbers, nor some damage numbers for the player, if only to see the difference between archmagi and no archmagi.

I was thinking about presenting the information in an easily digestible form, something like brogue:
  Code:
The goblin typically can kill you in 10 turns, at worst it can kill you in 5 turns.
Typically you can kill the goblin in 3 turns, at best you can kill it in 2 turns.


PowerOfKaishin wrote:Random effects, including variable damage, introduce 'complications' to an otherwise completely predictable situation. They make the game interesting and I don't think blanket removal is a good answer here. Getting rid of spell cast failure chance I can get behind though. It's a case-by-case basis.

I feel that the unpredictability of your actions can feel like you are not in control and can be frustrating. And while the damage or effect of a specific enemy attack is predictable you can't reliably predict what action the enemy will perform next and this will get progressively worse as enemies get stronger and smarter.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Thursday, 23rd February 2017, 07:11

Re: @dev team: if you had a time machine

tantanoid wrote:I feel that the unpredictability of your actions can feel like you are not in control and can be frustrating. And while the damage or effect of a specific enemy attack is predictable you can't reliably predict what action the enemy will perform next and this will get progressively worse as enemies get stronger and smarter.


Have you tried demon game from http://demon.ferretdev.org/?
Monsters always pick the best action while damage is still somewhat random (80-120%), misses and failures to apply negative effect (slow enemy, for example) add fun IMHO, also resistance/vulnerability makes negative effects weaker/stronger vs different monsters. Of course there are irresistible attacks and effects but they are weaker than resistable. Dev did really great work balancing different actions IMHO.
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25
User avatar

Abyss Ambulator

Posts: 1194

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

Post Thursday, 23rd February 2017, 14:55

Re: @dev team: if you had a time machine

I don't think resting to heal is that bad now. I used to think it was but in Sil you just regen really fast so you can keep walking around without having to worry about your health being too low to safely explore in most cases. It's like being a Vine Stalker or Troll basically. As long as the player heals fast enough it doesn't really interrupt game flow. I find this preferable to new DD, for example.
remove food

For this message the author tabstorm has received thanks:
Rast

tch

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 00:16

Post Thursday, 23rd February 2017, 17:06

Re: @dev team: if you had a time machine

I am not in the dev team and I would consider myself as newbie on crawl (even after hundres of hours of playing;). However, if you can realize this two points I will definitely play your game:

1. The actual reason why I started to play crawl was that the skills you train depend on the actions you take. This reminded my of my all time favourite Dungeon Master and if I would create a game it would definitely rely on this. I find it a bit sad, that this concept was softened (or removed at all) from crawl.
2. In general I like games where the enemy ai plays on the same rules as the player. Of course it is not always possible to realize this, but it makes the player feel part of the world instead of being an overpowered hero that is like nothing else in the game. To some extend this is true for crawl, for example when I meet an ogre with a huge club then I think that could be me after turning to the other side and protecting the orb instead of trying to get it.

Good luck!

Swamp Slogger

Posts: 135

Joined: Saturday, 21st February 2015, 07:40

Post Thursday, 23rd February 2017, 17:36

Re: @dev team: if you had a time machine

tch wrote:
1. The actual reason why I started to play crawl was that the skills you train depend on the actions you take. This reminded my of my all time favourite Dungeon Master and if I would create a game it would definitely rely on this. I find it a bit sad, that this concept was softened (or removed at all) from crawl.

Good luck!


As a heads up, while automatic skill training is on by default, it is not a very good idea to leave it on and most experienced players manually allocate which skills are being trained, so this aspect of the game isn't actually very prominent or supported (leaving it on doesn't usually result in very good skill distributions, and wastes points in skills that should either be taken to breakpoints or not invested in at all)

Edit:nvm misread your post,l. That said victory dancing wasn't very interesting so good riddance that's gone, and crawl is poorly suited to automatic skilling in the first place because of all the breakpoint skills and the very nonlinear skill cost growth

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Thursday, 23rd February 2017, 17:58

Re: @dev team: if you had a time machine

Thread summary:
"@dev team: if you had a time machine, what would you do?"

"I'm not a dev, but..." x100

For this message the author duvessa has received thanks:
neverEnough

Tomb Titivator

Posts: 900

Joined: Sunday, 30th December 2012, 05:26

Post Thursday, 23rd February 2017, 18:16

Re: @dev team: if you had a time machine

You hear "I'm not a dev, but...!" x44; You hear a dev comment! You hear "I'm not a dev, but...!" x56

For this message the author ZipZipskins has received thanks: 3
dracos369, PleasingFungus, VeryAngryFelid

Ziggurat Zagger

Posts: 5382

Joined: Friday, 25th November 2011, 07:36

Post Thursday, 23rd February 2017, 19:40

Re: @dev team: if you had a time machine

If you really wanted dev answers I'd recommend using the mailing list, if you post it on the forum, people will chime in. I assume tch is capable of telling who are devs and who aren't: devs have orange names, and usually don't respond on the forum, so if we really took @dev team to be exclusionary, then you'd likely end up getting no responses. The one dev who has commented started out by saying "I don't post on the tavern anymore, but I make an exception for you." I'm sorry we ruined what would have been a very dead thread with our thoughts on the topic :)

For this message the author tasonir has received thanks: 2
tantanoid, VeryAngryFelid

Tomb Titivator

Posts: 900

Joined: Sunday, 30th December 2012, 05:26

Post Thursday, 23rd February 2017, 19:42

Re: @dev team: if you had a time machine

you should be

in fact be sorrier

For this message the author ZipZipskins has received thanks: 2
chequers, duvessa

Ziggurat Zagger

Posts: 5382

Joined: Friday, 25th November 2011, 07:36

Post Thursday, 23rd February 2017, 19:52

Re: @dev team: if you had a time machine

ZipZipskins wrote:you should be

in fact be sorrier

I promise to self-flagellate by continuing to play the majority of my crawl games with Cheibriados.

For this message the author tasonir has received thanks:
ZipZipskins

Tomb Titivator

Posts: 900

Joined: Sunday, 30th December 2012, 05:26

Post Thursday, 23rd February 2017, 19:56

Re: @dev team: if you had a time machine

you're sick and you need help

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Thursday, 23rd February 2017, 20:43

Re: @dev team: if you had a time machine

tasonir wrote:If you really wanted dev answers I'd recommend using the mailing list, if you post it on the forum, people will chime in.


Thanks, I didn't even consider that the devs were still on the mailing list.

Dungeon Master

Posts: 3160

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

Post Thursday, 23rd February 2017, 23:27

Re: @dev team: if you had a time machine

I don't have the historical lens that dpeg does, but as a relatively new dev to the project, I'd say that crawl's biggest problem is a messy, disorganized codebase. Write a clean, organized one instead. :p

Beyond that, what I wish had been done differently in crawl is pretty much what my plans are now:
1) Replace food with a meaningful clock,
2) Differentiate ranged combat (see my ranged reform project that's perpetually in progress),
3) Shorten the core game,
4) Discourage luring,
5) Make curses more relevant.

If some of those things had been done up front, we'd probably have fairly polished versions of them by now.

But really, your question isn't what do we wish we'd done, but what should you do, and that depends heavily on what sort of game you're trying to build. Crawl has been evolving from a more tactical but still quite baroque Nethack descendant by stripping away the baroque features and focusing on the tactical gameplay. We've stripped enough of the baroque stuff (item weight, item destruction, etc.) that now the principle remaining issues are the lack of a meaningful clock and the power of luring, but we've chosen to keep the exploration feel with our fully backtrackable dungeon. However, if you want to make a more baroque roguelike, or a fully tactical / no exploration roguelike, or any other direction, you might find entirely different choices and challenges between you and that goal.

So what I think you need to do is decide what your game should feel like and then try to identify which design directions will amplify that feeling without creating headaches for future development. Sometimes easier said than done.

For this message the author Lasty has received thanks: 4
Brannock, Rast, tantanoid, VeryAngryFelid

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Thursday, 23rd February 2017, 23:49

Re: @dev team: if you had a time machine

Lasty wrote:But really, your question isn't what do we wish we'd done, but what should you do, and that depends heavily on what sort of game you're trying to build. Crawl has been evolving from a more tactical but still quite baroque Nethack descendant by stripping away the baroque features and focusing on the tactical gameplay. We've stripped enough of the baroque stuff (item weight, item destruction, etc.) that now the principle remaining issues are the lack of a meaningful clock and the power of luring, but we've chosen to keep the exploration feel with our fully backtrackable dungeon. However, if you want to make a more baroque roguelike, or a fully tactical / no exploration roguelike, or any other direction, you might find entirely different choices and challenges between you and that goal.

So what I think you need to do is decide what your game should feel like and then try to identify which design directions will amplify that feeling without creating headaches for future development. Sometimes easier said than done.


The reason I am asking here is because DCSS is the closest RL to what I am trying to make. Really I could've just started a fork, I am just choosing to write it from scratch to be able to use a different underlying technology and a different license.

When you talk about luring do you mean just separating a monster from the group to kill it more easily or do you also include kiting and stair dancing as part of the issue?

bel

Cocytus Succeeder

Posts: 2184

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

Post Friday, 24th February 2017, 05:45

Re: @dev team: if you had a time machine

tantanoid wrote:When you talk about luring do you mean just separating a monster from the group to kill it more easily or do you also include kiting and stair dancing as part of the issue?

I don't know what Lasty means by luring (people's definitions differ a bit), but to me, luring means trying to fight as few monsters as possible in as controlled an environment as possible. Some of this is inherent to the way crawl works and is fine; but it is bad when taken to excess because it kills or mitigates much of the point of various things in the game.

This definition would include some of the stuff you mentioned (some kinds of kiting wouldn't qualify, perhaps), as well as luring single monsters to an already cleared area (or upstairs).

You can search for various threads on luring using the "search" function. It's a perennial topic of discussion here.

Shoals Surfer

Posts: 300

Joined: Thursday, 1st May 2014, 13:13

Post Friday, 24th February 2017, 07:19

Re: @dev team: if you had a time machine

not a dev, but:
1. Persistent floors are a bad idea.
2. Roguelikes should be linear. (Branching paths are ok, but not going back and completing the initially unchosen path.)
3. Roguelikes can't both be long and have strategic (rather than tactical) depth.
4. You should avoid having multiple systems that take up the same design space. For example, DCSS probably shouldn't have both magic and gods.

Ziggurat Zagger

Posts: 4432

Joined: Friday, 8th May 2015, 17:51

Post Friday, 24th February 2017, 10:22

Re: @dev team: if you had a time machine

all before wrote:3. Roguelikes can't both be long and have strategic (rather than tactical) depth.


Why not? I lost a game in demon where I had been enjoying playing all the way until I had reached a point where I quickly realized I made some strategic mistakes, I died pretty quickly after that. It didn't require spoilers to win, I suspected I was doing something wrong (similar to ignoring fighting and training just fire on a caster in dcss).
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Dungeon Master

Posts: 3160

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

Post Friday, 24th February 2017, 14:18

Re: @dev team: if you had a time machine

tantanoid wrote:When you talk about luring do you mean just separating a monster from the group to kill it more easily or do you also include kiting and stair dancing as part of the issue?

What I mean is that by luring is activities that exchange non-trivial time/keypresses for tactical advantage, and activities that separate packs of monsters without any meaningful cost except non-trivial time/keypresses. I don't believe that luring should be removed, but I do believe that it needs to have a non-trivial in-game cost to discourage people from thinking that the correct tactic will always be to lure all monsters back to (a killhole|a staircase|some rare good terrain) before fighting.

In brief, there should not be one almost-always-correct tactical choice to minimize overall danger, and the best choices should not involve a meaningful quantities of extra keypresses/real-world time to execute as compared with less-good choices.
User avatar

Slime Squisher

Posts: 419

Joined: Wednesday, 21st September 2011, 09:45

Post Friday, 24th February 2017, 16:12

Re: @dev team: if you had a time machine

Some kind of claustrophobia penalty would have been good to fuck up kill hole strategy. Basically if you have 7 tiles of walls next to you and only one tile you can see, you'll get penalties. I kind of liked disint wand chopping one piece of wall no matter if I was making a kill hole or not.

Just throwing this on the air since this is CYC :)

Temple Termagant

Posts: 8

Joined: Wednesday, 22nd February 2017, 07:23

Post Friday, 24th February 2017, 19:46

Re: @dev team: if you had a time machine

Lasty wrote:What I mean is that by luring is activities that exchange non-trivial time/keypresses for tactical advantage, and activities that separate packs of monsters without any meaningful cost except non-trivial time/keypresses. I don't believe that luring should be removed, but I do believe that it needs to have a non-trivial in-game cost to discourage people from thinking that the correct tactic will always be to lure all monsters back to (a killhole|a staircase|some rare good terrain) before fighting.

In brief, there should not be one almost-always-correct tactical choice to minimize overall danger, and the best choices should not involve a meaningful quantities of extra keypresses/real-world time to execute as compared with less-good choices.


I see. Well there are a couple of options that occur to me:
  • Smart monsters could detect that they have been following the player for some time or that they are in disadvantaged terrain and switch to a different strategy like ranged attacks or using some ability that changes the terrain or retreating to a place where they can ambush the player instead.
  • For monsters that are balanced as a group the AI could always keep them grouped, so either they all chase the player or none do. And if they are funneled through a tight spot they should be able to swap out the ones low on hp to the back of the group as the player can do with their followers.
  • A radical way to discourage kiting would be to add directionality to actors and make turning around take time. Luring will become more difficult as well if the FOV prevents seeing whatever is behind. This could also add a lot more tactical depth in general, I'll experiment in this area.

Ziggurat Zagger

Posts: 5382

Joined: Friday, 25th November 2011, 07:36

Post Friday, 10th March 2017, 23:25

Re: @dev team: if you had a time machine

all before wrote:not a dev, but:
1. Persistent floors are a bad idea.
2. Roguelikes should be linear. (Branching paths are ok, but not going back and completing the initially unchosen path.)

I really like these ideas, and they're ones I hadn't thought of first. You can only move forward, and when you have a choice between two areas, can't go back to the one you skipped. If we're going to commit to this then you probably want to have some sort of guarantee that entrances to floors won't be completely surrounded with swarms of high-threat monsters, since the player can't go back up the stairs. But aside from that, it would really streamline play and discourage lots of backtracking/stashing/being overleveled for the area you're in.

Ziggurat Zagger

Posts: 5382

Joined: Friday, 25th November 2011, 07:36

Post Friday, 10th March 2017, 23:29

Re: @dev team: if you had a time machine

tabstorm wrote:I don't think resting to heal is that bad now. I used to think it was but in Sil you just regen really fast so you can keep walking around without having to worry about your health being too low to safely explore in most cases. It's like being a Vine Stalker or Troll basically. As long as the player heals fast enough it doesn't really interrupt game flow. I find this preferable to new DD, for example.

One of the things I started to get into the habit of after 20-30 wins was to (almost) never sit and rest to full. Technically it's non-optimal, but I also have a habit of building characters with high regeneration. I have a lot of troll and VS wins, and get the regeneration spell on ALL of my characters if I find the spell. Most of my wins are around 60-75k turns for 3-5 runes, and around 70-90k turns for 15 runes. These aren't really speedruns, in that I'm not "diving" and still explore the full levels, I just move quickly through them. When I worked on my "actual" speedrun, I ran with VSMo, and also picked up the regeneration spell.

For this message the author tasonir has received thanks:
stoneychips

Return to Crazy Yiuf's Corner

Who is online

Users browsing this forum: No registered users and 18 guests

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