00:45:23 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-295-g4b1f6b956e 03:55:36 Valeaz (L24 GnEE) Crash caused by signal #6: Aborted (Crypt:1) 03:55:40 Valeaz (L24 GnEE) Crash caused by signal #6: Aborted (Crypt:1) 03:55:45 Valeaz (L24 GnEE) Crash caused by signal #6: Aborted (Crypt:1) 04:33:54 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5249-g4a8afe7061 06:04:59 <06wensley> I have a tentative design document for a MVC architecture which would make headless mode the default, but don't hold your breath that I ever get around to actually implementing it unless people are really enthusiastic about it 😛 12:08:02 <06wensley> I graphed the winrates vs. games played for all species in 0.32. only counting non-boring games, which excludes 25% of all felid games and 13% of all demonspawn games. Oni is crushing it right now with a 2.8% winrate, meanwhile people love octopodes and demonspawn despite their atrocious winrates 12:08:03 <06wensley> https://cdn.discordapp.com/attachments/747522859361894521/1302348515397079151/image.png?ex=6727ca12&is=67267892&hm=8919031eeda4498a5490de33c214ef4f87acdbd83a0556ca7bf38f8302355184& 12:12:57 <06wensley> big cluster in the lower-left octant indicates species that could probably use some love: formicid, spriggan, human, mummy, tengu, kobold, vine stalker 12:22:22 -!- wantyapps_ is now known as wantyapps 12:24:31 <06pleasingfungus> y axis crimes alert! 🚨 12:25:26 <04dracoomega> While this is a fun graph to look at, I'd be wary of designing as though there's a particular target power level or popularity for species as a whole. Like, mummy is explicitly a chalenge species, so both low winrate and popularity are expected across the entire playerbase (and, in fact, they are iconic and popular to talk about for what they are). Formicid, despite being in the lower-left octant, are more popular than over half the 12:25:27 species in the game, which is pretty good for another 'challenge' species, suggesting people actually find their power fantasy fun. In a game with so many different species, their popularity is always going to be kind of uneven. Or rather, I think it's inevitable that a small number will always dwarf most of the rest of them, and this is fine so long as individual species feel like they have an acceptable niche and at least some decent number of 12:25:27 people who enjoy them. 12:26:32 <04dracoomega> This is not to say that everything over there is sacrosanct, but I'd want a more specific reason to do something to one of them than just 'relatively low power and relatively low popularity' 12:26:35 <06pleasingfungus> seems like one could also argue the species in the bottom right need nerfs 12:26:42 <06pleasingfungus> since they’re too easy to be fun 12:26:51 <06pleasingfungus> or something 12:27:01 <04dracoomega> I think felid has niche popularity which inflates its winrate a bit 12:27:13 <12ge0ff> According to the logs, people ask for libcrawl at least 2-4 times per year. Does this count? 🙂 Also, where can we read this design document? 12:27:55 <04dracoomega> (Though I felt like I remember seeing some recent stats that had oni's dramatic winrate lead slip a lot, and I wonder now exactly what I was looking at) 12:28:20 <06wensley> I think a lot of felid players quit out when they lose a lot of lives in short succession, their winrate goes down significantly if you don't exclude boring games 12:28:30 <04dracoomega> Oh, huh 12:29:02 Kind of wants "boring, but only if you quit with 0 points" 12:29:06 <06wensley> as mentioned, 25% of ALL felid games are boring 12:29:12 <04dracoomega> I do think felid is probably a somewhat hard species to gently nerf without making the experience of playing them miserable, but I have never been a felid fan myself, so I may lack perspective 12:29:13 <06wensley> which is a huge outlier 12:32:09 <04dracoomega> Oh, really, was it just 'trunk', I wonder? 12:32:14 <04dracoomega> !lg * !boring cv>0.32 bysp / won 12:32:14 <04Cerebot> 1859/146784 games for * (!boring cv>0.32 bysp): 163/11089x Mi [1.47%], 153/14563x Op [1.05%], 148/19564x Ds [0.76%], 130/6287x Dr [2.07%], 113/11782x Co [0.96%], 94/7490x On [1.26%], 92/7428x Gr [1.24%], 89/7982x MD [1.12%], 81/3823x Dj [2.12%], 71/7695x DE [0.92%], 67/5251x Tr [1.28%], 56/1068x Fe [5.24%], 53/5455x Gn [0.97%], 51/2688x Mu [1.90%], 49/3405x Mf [1.44%], 49/5319x Fo [0.92%], 47/352... 12:33:43 <04dracoomega> Where oni seem perfectly back in the pack of 'pretty good' and not 'wildly good'. (Whether there's any other statistical confusion going on there which I'm overlooking, I do wonder if there's a certain degree of 'new, but not wildly popular, species get disproportionally played by veterans for a while, which inflates their winrate also'. Not sure. 12:34:07 <04dracoomega> (Coglins and mountain dwarves were, of course, extremely widely popular from their debut) 12:35:54 <06wensley> yeah I wouldn't worry about any species being too popular, more just trying to make sure that unpopular species are unpopular for interesting reasons 12:36:19 <04dracoomega> Really, ghoul was slated for some touch-ups because it was a combination of both being literally the least popular and also being considered fairly boring/limited within that 12:36:41 <04dracoomega> Like, felids may be unpopular, but clearly have superfans. I've never seen anyone profess love of ghouls 12:38:06 I've a soft spot for ghouls but then I crave human flesh^W^W^W kind of accidentally 15-runed one 12:42:33 <04dracoomega> Could you crave human souls instead? 12:42:45 <04dracoomega> Ghoul replacement will still care about such things 😛 12:44:18 <04dracoomega> (Really, the idea was that revenant remained fairly good in the ways that ghoul was good, while ideally being a little more interesting/flexible in other ways. And a new aesthetic coat of paint.) 12:44:26 <04dracoomega> Spooky skeleton on fire! 12:44:54 <06wensley> I also think that ghoul and mummy never really recovered from food removal, like mummy wasn't supposed to be a challenge character, ignoring food was such a powerful ability that the loss of potion drinking was considered largely balanced 12:46:22 (as someone who's still playing with food) how extraordinary 12:48:07 <04dracoomega> I... I think mummy was kind of always a challenge character, tbh. Their food perk let them do a very limited number of things better than other people did, but they were still pretty bad in general. 12:49:19 <04dracoomega> (Like, it's true that they at least felt like they had A unique upside in a way they might not at the moment, but largely I think they were fairly bad. Certain mummy combos were legendarily bad since very old days) 13:02:04 <06wensley> I mean, MuBe and MuPr were legendarily bad, but only in the sense that we at least let you play them where we didn't let you play DgBe and DgPr 😛 13:02:49 <06wensley> (I could be misremembering whether we let you actually select MuPr, maybe that was a joke) 13:09:34 <06pleasingfungus> !hs * mupr 13:09:36 <04Cerebot> 1167. greensnark the Destroyer (L27 MuPr of Yredelemnul), escaped with the Orb and 3 runes on 2010-03-14 21:13:08, with 1146860 points after 185906 turns and 13:36:33. 13:09:38 <06wensley> the design document is a notebook that I write in when I'm bored in meetings 😛 the first step would be to define a stable interface for frontends to speak to the model. this would be relatively straightforward I think, since there's a not-overwhelming number of commands that players can send to the game. the harder part is having an interface for the model to communicate with the frontends (communicating all map knowledge, messages, 13:09:38 menus, etc.). but there's a TON of potential benefits to decoupling in this way. but also it's a lot of work so 13:10:14 <06wensley> oh I guess mupr just got yred like how hopr got beogh? that would be a good combo actually then 13:13:43 <06pleasingfungus> !lg * mupr s=god 13:13:44 <04Cerebot> 1167 games for * (mupr): 1138x Yredelemnul, 24x, 3x Okawaru, Nemelex Xobeh, Sif Muna 13:13:49 <06pleasingfungus> looks like it 13:14:02 <06pleasingfungus> !lg * pr yred s=crace 13:14:04 <04Cerebot> 5298 games for * (pr yred): 1349x Demonspawn, 1138x Mummy, 328x Kobold, 294x Human, 284x Draconian, 234x Vampire, 212x Mountain Dwarf, 206x Minotaur, 175x High Elf, 157x Sludge Elf, 147x Deep Elf, 110x Ghoul, 109x Ogre, 69x Naga, 67x Troll, 64x Merfolk, 60x Grey Elf, 56x Hill Orc, 55x Centaur, 40x Halfling, 37x Spriggan, 37x Tengu, 29x Deep Dwarf, 16x Felid, 9x Hill Dwarf, 9x Elf, 6x Ogre-Mage, G... 13:14:35 <03wheals> poor ghoul 13:16:20 <06pleasingfungus> !hs * ghpr 13:16:23 <04Cerebot> 111. rob the Middleweight Champion (L27 GhPr of Yredelemnul), escaped with the Orb and 3 runes on 2009-08-08 16:50:12, with 1543959 points after 80725 turns and 7:49:14. 14:21:51 <09gammafunk> hard to beat rob, he was good (probably still is, but I'm not sure he plays crawl any more) 14:24:12 <09gammafunk> also, aside from the usual caveats of trying to design species around their "current" winrate and all the challenges that faces, I think in some of those cases the winrate are just extremely far from the species' actual power level 14:24:37 <09gammafunk> like spriggan as "the 4th weakest" species is pretty laughable 14:26:54 <09gammafunk> in fact: 14:26:56 <09gammafunk> !lg * 0.32 sp !boring s=name 14:26:58 <04Cerebot> 5598 games for * (0.32 sp !boring): 1326x diceflayer, 520x tdog, 222x Denis, 160x SoulStone, 113x isplatspriggans, 106x thisguise, 99x wolf37, 72x Blackmore, 65x Selgeron, 63x Skittlebrew, 62x Dumalacath, 57x 7tar4e, 53x xccam, 47x Fyre, 46x MintTicTac, 45x Skryme, 38x KarmaDistortion, 36x pichaelthompson, 35x humterzook, 33x whitephoenix, 32x Doucebaguette, 32x monsoonserge, 31x silberrucken, 30... 14:27:16 <09gammafunk> !calc 1326.0 / 5598 * 100 14:27:17 <04Cerebot> 23.69 14:27:32 <09gammafunk> one player is responsible for almost 24% of all non-boring 0.32 sp games, for example 14:28:19 <09gammafunk> and for the other less played species there will be other instances of this, possibly more extreme, that can shift winrate up or down significantly based on what those dominant players are doing 14:28:39 <09gammafunk> now it'll be funny if diceflayer is actually driving up sp winrate, so maybe I should check that 14:28:52 <09gammafunk> !lg !diceflayer 0.32 sp !boring / won 14:28:54 <04Cerebot> 38/4272 games for diceflayer (0.32 sp !boring): N=38/4272 (0.89%) 14:29:14 <09gammafunk> yeah that's a pretty dramatic change in winrate compared to the graph 14:29:22 <09gammafunk> still below 1% overall but you get the idea 14:30:07 <09gammafunk> it's pretty tricky to normalize for all of this stuff 14:42:03 <04dracoomega> I do think spriggan's move speed is meaningfully less strong in modern times than it was in older ones, for a couple reasons, and they obviously can be very squishy. So I'm not actually sure that they're nearly as strong in relative terms as they once were. But regardless of debate about 'true power level' (whatever this means), species varying substantially in this is something Crawl has considered deliberate design for a long 14:42:04 tme 14:43:52 <09gammafunk> qw has won a spriggan but has never won an oni, so I think that proves pretty decisively that spriggan is in fact stronger than oni 14:43:59 <09gammafunk> (joke) 14:47:57 <09gammafunk> regarding modern spriggan move speed advantage, I feel that for a Theoretical Optimal Player (a TOP), the pre-lair move speed advantage is still extremely strong, and these areas are a dominant factor for winrate. Obviously the early game has gotten a bit easier compared to older versions in terms of what spawns, so move speed advantage might be helpful less often, but it's still extremely good 14:48:22 <09gammafunk> If we're talking winrate for a more casual player, maybe it's different now 14:51:41 <04dracoomega> I think attacks of opportunity make fast move less of a gauranteed escape from a lot of things that it used to be 14:51:54 <04dracoomega> And that adds up quite a bit in average play 14:52:18 <04dracoomega> Used to be many situations where you knew that if the dice rolls went badly that you could just step away 100% of the time. And now that isn't true. 14:53:09 <04dracoomega> (I am not saying that fast move isn't very strong, since of course it is. But I do think it's meaningfully less strong in a bunch of common scenarios. And spriggans were always very fragile outside of that) 16:58:23 03regret-index02 07* 0.33-a0-296-g381ce9d8e6: Don't display doubled vigour fading messages out-of-los (Implojin) 10(13 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/381ce9d8e681 20:38:31 -!- The topic of #crawl-dev is: Crawl Development | https://github.com/crawl/crawl | Logs: http://s-z.org/crawl-dev/, temporarily http://crawl.akrasiac.org/logs/cheibriados/ | People with +v have commit access, devs on bridged discord as well | General Crawl-related chat to #crawl | Long stuff to a pastebin service, please 20:38:31 -!- The topic of #crawl is: Play Dungeon Crawl Stone Soup online now! Type ??online for instructions, ??lg / !lg for play stats | PM Sequell for long queries | http://crawl.develz.org | FooTV game replays: ??footv for instructions | #crawl-dev for dev discussion, #crawl-offtopic for offtopic 23:35:30 Unstable branch on crawl.develz.org updated to: 0.33-a0-296-g381ce9d8e6 (34) 23:59:44 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-296-g381ce9d8e6