00:01:04 although we were pretty much 5/6 00:01:31 compared to what implojin can normally do, you were at most 4/6 00:02:04 hah fair enough, it was his hyper account too! 00:02:45 this isn't really -dev so i did want to mention one thing 00:03:06 i think the scoring for real time vs. turn count is fairly imbalanced 00:04:08 yermak's low turn win got a lot less points than the fastest rt win, even though low turncount wins are a lot rarer than low rt wins 00:04:24 !lg !bot won turns>15000 s=name 00:04:33 57872 games for bot (won turns>15000): 666x comborobin, 464x 78291, 364x bmfx, 357x Yermak, 348x Sharkman1231, 334x elliptic, 302x johnnyzero, 290x nago, 280x MrPlanck, 264x Manman, 249x Dynast, 242x WalkerBoh, 231x jeanjacques, 227x araganzar, 216x minmay, 215x stickyfingers, 208x Basil, 208x HilariousDeathArtist, 207x Zooty, 207x Lasty, 205x MorganLeah, 201x hyperbolic, 200x Rubinko, 199x Bart, ... 00:04:37 oops 00:04:43 !lg !bot won turns<15000 s=name 00:04:48 19 games for bot (won turns<15000): 5x jeanjacques, 4x hyperbolic, 2x Sapher, splat, Charly, Yermak, Bloax, titmouse420, qwqw, Doesnt, damdam 00:05:04 nikheizen: yeah the constants there should be renormalized 00:05:06 !lg !bot won dur<45:00 s=name 00:05:11 15 games for bot (won dur<45:00): 4x DemiseSR, 3x YTcomUltraviolent4, 2x Ultraviolent4, 2x Shard1697, Mnoleg, SpeedrunTech, Demise, AllRandomAllDemise 00:05:50 when I originally chose the speedrun constants the number of points you could get for realtime/turncount with a very good run were pretty similar 00:06:11 but people have gotten a lot better at realtime (and various interface improvements have helped) 00:06:22 50:00 is probably a better number 00:07:03 !lg !bot won dur<50:00 s=name 00:07:08 22 games for bot (won dur<50:00): 5x DemiseSR, 3x Ultraviolent4, 3x YTcomUltraviolent4, 2x tabstorm, 2x SpeedrunTech, 2x Shard1697, AllRandomAllDemise, Mnoleg, Demise, manman, simm 00:07:55 while we're at it, it would be nice if streak scoring was weighted based on quality of the streak (perhaps defined based on the r/c/g pts for the won combo) 00:08:35 I think at the time (a long time ago) it felt like a 20k turn win and a 90 minute win were roughly similar difficulty 00:09:19 i can definitely do a sub90 win, but a sub20k seems a lot harder 00:09:39 nikheizen: actually, I think you misread something? yermak's turncount win was 480 points, fastest rt win was 544 00:09:44 not that far off 00:10:35 that's about half of the difference between manman and yermak so i think it's fairly substantial since yermak's win should (arguably) be worth more 00:11:01 I'd still vote for reducing rt win points by 25% or some such, yeah 00:11:11 but I don't think it is that ridiculous as is 00:11:48 WalkerBoh: sounds like a bad idea to me, would be very complicated and most r/c/g are very similar streak difficulty 00:12:24 elliptic: what if it were something simple like a streak win doubles your r/c/g pts for the combo? 00:12:32 i disagree that r/c/g is irrelevant for streaks 00:12:59 it just seems like a shame that you are incentivized to "easy streak" power combos 00:13:24 well you're also incentivized to choose combos that give you points other than for streak 00:13:26 WalkerBoh: even something simple like that would reward doing things like -> 00:13:32 if you didn't somehow prevent that 00:14:03 gammafunk: not really though, because those pts get so heavily compressed compared to the streak pts 00:14:05 WalkerBoh: you are still heavily rewarded for streaking nchoice or other high value combos 00:14:30 elliptic: i disagree, the difference there is almost irrelevant 00:14:59 12-wins with easy combos is worth more than 8 wins with hard ones 00:15:11 you're totally wrong there 00:15:27 WalkerBoh: it depends a bit on how many wins you get in total 00:15:41 tourney rules aren't really designed for the modern age of everyone getting 20+ wins :P 00:15:42 it's pretty tricky though, the point compression is specifically to downplay the effect of just getting lots of wins 00:15:46 of course, this is only relevant for the few players who can do that 00:16:00 seems like you're essentially arguing the point compression shouldn't exist for streak wins 00:16:10 gammafunk: i don't think the compression works well anyways 00:16:21 right, but that would be the thing to address, in that case 00:17:09 I wonder if a lot of this wouldn't be better if there were just separate competitions 00:17:11 fwiw I don't actually think a 12-streak with "easy combos" that still don't repeat race/class is easier than streaking 8 nchoices 00:17:33 so it doesn't bother me if in practice it is worth more points 00:17:41 elliptic: i suppose it depends somewhat on what the nchoice characters are 00:17:50 I think it would depend on how bad the nchoice combo in question is 00:17:53 but in general they have a much higher chance of dying early on 00:17:57 eh 00:18:29 maybe not for you 00:18:59 but I would feel a lot better about trying HOGl or some such than TeWr 00:19:41 anyways, I agree gammafunk that the tournament rules are hard because it has to have broad appeal 00:19:41 there were a lot of good nchoice combos this time 00:20:11 yeah WalkerBoh and you could have separate competition categories that only counted points for the think you're trying to rank, but it might make the tournament too fragmented 00:20:44 I mean, it is certainly a defensible option to just remove tourney points 00:21:13 the one thing I wish the tournament did a better job of is collaboration with your teammates 00:21:14 and maybe expand a bunch of the stuff on the tourney overview page to show top 5 instead of top 3 and stuff like that 00:21:45 so players can go for those or for banners 00:21:48 as they wish 00:21:51 I would prefer the tournament to really reward extraordinary feats above "no-lifing" 00:22:00 yeah, I could see that 00:22:24 I mean, no-lifing crawl is maybe not extraordinary but it's certainly hard to do :) 00:22:45 the clan/top player competition was really epic this year 00:23:09 that's true, and in this case a lot of players at the top were also doing pretty incredible things 00:23:13 I'm sure it has been as well in past tournaments where I've not payed as much attention or participated, but certainly some good competition stories 00:23:43 Yermak was driving manman insane 00:23:55 I think he was having nightmares about him jumping out ahead 00:24:01 when he put up about 700 pts and closed the gap in one day? 00:24:15 actually several times throughout the tournament! 00:24:28 when he broke out with that insane 10k low turncount win 00:24:30 what about a 0.1% score penalty for every hour spent playing 00:24:34 that we had no idea he could do 00:24:41 !lg qw t x=sum(dur) 00:24:52 1885 games for qw (t): sum(dur)=14d+6:08:33 00:24:52 gammafunk: that's what I mean, he put up that and the 55M game in one day 00:25:18 oh, yeah, I thought you were talking about his final push after losing the Gh; yeah, that wasn't a 700pt gain, so I should have realized 00:25:45 yeah 00:25:54 yeah those two games were probably the biggest moments; ManMan is not really a turncount player, more of a realtime player (and streak player, as it turns out) 00:26:09 manman still put up a pretty quick win though 00:26:11 17k? 00:26:16 yeah 17k turns is very good 00:26:30 he didn't have a fast allruner though 00:26:37 I put up 20.1k midway through and thought it would stand in the top 3, but it barely made top 5 00:27:21 right, he was doing what he could to pick up a turncount 3-rune win and pulled it off nicely, but I think that was specifically in response to yermak 00:27:26 from my point of view in the tournament right now though, it's very hard to do much if you break a streak around the 5-6 game mark 00:27:29 that's what killed Yermak in the end 00:28:03 late in the tournament there just isn't an effective way to pick up significant points if you're not on a streak 00:28:03 yeah, manman's streak almost matched theglow; I really thought he was going to do it 00:28:35 I mean, Yermak won last tournament by over 500 points while only having a 10-streak 00:28:40 "only" 00:28:43 so he would have won that one with only a 5-streak 00:29:40 uv's streak ending with him becoming a snack for the serpent of hell in dis was pretty great 00:29:43 so "hard to do much" seems like an exaggeration 00:29:44 doy would be happy about that 00:30:06 elliptic: Yermak is quite an outlier though 00:30:39 sure but I still think you are exaggerating... also not every tourney has a 15-streak in it 00:31:14 that's true too, the winning streak is consistently 10-15 though 00:31:36 I do think that it is a bit unfortunate that players can find themselves with very few ways to gain significant numbers of points in the last few days though 00:32:23 I can only say from my own perspective that it's very demoralizing to drop a 5-streak about 4 days in and know that realistically that means a top 5 finish is off the board 00:32:29 but I'm not sure how to fix that without making it even better to get 30 wins 00:33:07 the only reason people are winning 30 games is because win quantity is more important than win quality 00:33:19 no matter how low rune and r/c/g points go 00:33:25 Hmmm, a 0.1% total score penalty per 30 minutes spent playing during the tourney, would reduce qw's score by 6.84%, yermak's by 4.8%, and manman's by 2.88% 00:33:45 no, the reason is that people are playing for too many hours and are playing too quickly :P 00:34:18 if winning the tournament realistically meant that you had to collect a majority of tier III banners, and the emphasis on that was much higher than just winning all combos... 00:34:26 WalkerBoh: also, Beargit was top 5 this tourney with only a 4-streak 00:34:56 hm yeah, where did his points come from 00:35:47 I guess it was sheer win volume 00:36:05 a lot of high value wins too, I'm looking at that 334 point Op win 00:36:47 but yeah, 24 wins is a lot 00:37:35 this is all honestly nitpicking, because the tournament is a pretty good format 00:37:45 I'm dubious about placing too high a value on banners because most of them are not nearly as fun to get the second time 00:38:03 possibly that is a problem with the banners, but coming up with good banners is hard 00:38:35 reading banners from some early tournaments was pretty amusing, some of them have come a long ways 00:38:43 which ones would you say aren't fun multiple times? 00:39:04 heretic 00:39:06 avarice 3, vow of courage 3, and orbrun tomb all seem good 00:39:26 sure there are a few 00:39:28 haha heretic was a bit overdone this time around 00:39:55 I liked the version of heretic where you had to win while abandoning multiple gods in the same game 00:40:09 well it's unchanged from 0.19 00:40:13 spiteful 00:40:28 ascetic 00:40:31 but yeah I don't think we felt it was the perfect banner, just one that was reletively fitting and reasonable 00:40:51 gammafunk: it was fine since it was optional 00:41:16 i agree that points should only be given for banners that can be done in a single game 00:41:20 and a bunch of the others are sort of similar 00:41:29 or otherwise sort of stupid like the invo title one 00:41:43 my point was just that I would personally rather have those weighted much higher 00:42:07 increasing banner points a bit sounds reasonable to me since scores are so much higher now 00:43:50 100 seemed reasonable to me - that would motivate me to go for them late in the tournament 00:43:52 and also as a way to give people more things to do late in the tourney 00:44:00 exactly 00:45:37 I'd probably suggest raising them all to 75 00:48:55 reduce real time points or increase turncount points? 00:50:14 the point values currently seem maybe a bit high relative to the overall scores 01:12:04 nikheizen: reduce realtime, since that's the one that people have gotten better at 01:17:19 yeah 01:17:53 elliptic: so the rule change would be to make all t3 banners worth additional points be worth +75 points (instead of some being only +25 or +50) 01:18:19 yeah 01:18:52 nchoice is still different from the others because it gets hit by the combo_god_win stepdown (and you can get it multiple times) 01:18:57 I have pushed all the 0.20 changes to the 0.20 branch in the github dcss_tourney repo and have merged that into master 01:19:10 so if you feel inclined to make changes in that repo, please feel free 01:19:24 if I make them, I'll put them in master 01:19:39 since 0.20 is now 'frozen' 01:19:45 sounds good 01:23:02 i like this bad namers list 01:23:14 (what did he try to name his team) 01:29:58 inspect his rc if you care 01:30:07 probably still there 01:31:32 what's the command to view a 0.20 rc? 01:32:05 nm 01:32:18 ah i see 01:36:20 wait, how did the winrate go up this version 01:52:24 Hello world, is anyone around? I'm trying to build from source but running into an error on Mac OS X 01:57:09 Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-51-gea492e9 02:23:00 <3 02:23:13 pretty happy that the serpent of hell is still generating spectacular kills 02:23:15 (: 02:52:04 doy: we made an emote using the dis SoH tile called SnacksOnUV4 02:52:06 Monster database of master branch on crawl.develz.org updated to: 0.21-a0-51-gea492e9 02:52:47 players are really afraid of that one, wich is cool but it's also a bit of a shame that there is disparity between the different versions 02:53:06 I suppose it's also sort of a shame that it tends to be a lot worse than the hell lords themselves, but maybe that's a problem with the hell lords 02:54:18 if it were easier than the hell lords it wouldn't be very interesting, right? you know you're going to run into each of the hell lords for sure, and the SoH is only suspenseful because it makes one of the branches more difficult than you expect 02:55:58 I suppose that's true, but the issue with that is fighting this horrible thing and then fighting the "end boss" that's markedly easier than what you just fought 02:56:37 put the serpent in the same place as the hell lord 02:56:39 it's not very satisfying; maybe the way the SoH mechanic interacts with the level could be changed 02:56:53 amalloy: after you kill the hell lord it becomes the SoH 02:57:08 even more reason not to kill them 02:57:08 and then after you kill the SoH *you* become the SoH 02:57:28 ok after you pick up the rune the SoH appears 02:57:33 and then the hell lord becomes the SoH 02:57:37 and then you become the SoH 02:57:45 joke's on you, i always apport the rune back to the stairs first 02:57:56 foiled again by optimal play 03:20:12 gammafunk: did you see my question about transporters and disconnected levels? 03:20:26 yeah, I responded to you 03:20:38 it's an issue I need to look into 03:20:50 oh, i missed the response one way or another 03:20:55 maybe you missed my response, but it was shortly after your question 03:21:17 ah well, as long as you're aware of it 04:04:25 -!- amalloy is now known as amalloy_ 04:06:35 -!- amalloy_ is now known as amalloy 04:13:01 Weird runed door 13https://crawl.develz.org/mantis/view.php?id=11099 by Yermak 04:13:01 Enslaved monster doesn't stop constricting an enemy when charm is ended. 13https://crawl.develz.org/mantis/view.php?id=11098 by Yermak 04:18:13 MP cost of evoking Olgreb's staff isn't shown anywhere. 13https://crawl.develz.org/mantis/view.php?id=11100 by Yermak 04:23:23 Regen status from spell isn't greyed out with 'no regen' mutation and monsters in sight. 13https://crawl.develz.org/mantis/view.php?id=11101 by Yermak 04:33:45 auto_eat_chunks tries to pick up chunks instead of eating off the ground. 13https://crawl.develz.org/mantis/view.php?id=11102 by Yermak 04:49:19 Switching between gods while training only Invocations gods is clunky. 13https://crawl.develz.org/mantis/view.php?id=11103 by Yermak 04:59:41 Umbra + halo wrongly stops invisibility under TSO. 13https://crawl.develz.org/mantis/view.php?id=11104 by Yermak 05:02:30 -!- amalloy is now known as amalloy_ 05:05:20 Question: why force_more_message doesn't have power of runrest_stop? 05:08:05 Example: I set force_more_message for finding a shaft. Now whenever I find a shaft during autoexplore I'm forced to press spacebar and then exploration continues. It makes no sense. 05:09:18 To solve my problem I have to add runrest_stop for this event too. But the question remains. 05:09:32 Should I file report on this topic? 05:20:26 Autopickup mode shouldn't change on saving and resuming the game. 13https://crawl.develz.org/mantis/view.php?id=11105 by Yermak 05:30:48 Slow status lasts a turn after elephant slug is dead. 13https://crawl.develz.org/mantis/view.php?id=11106 by Yermak 05:41:10 Gaining robust mutation should expand your current hp along with mhp. 13https://crawl.develz.org/mantis/view.php?id=11107 by Yermak 06:23:26 -!- Euph0ria is now known as ScreamingSience 06:23:35 -!- ScreamingSience is now known as ScreamingSilence 06:23:49 !messages 06:23:52 (1/2) Mandevil said (16h 20m 33s ago): When will CUE feed 0.20 data to akrasiac scoreboard? 06:23:56 !messages 06:23:57 (1/1) Medar said (14h 2m 54s ago): FYI rebuilding CUE doesn't show messages in real time, because it's using gzip content-encoding. Doesn't really matter in practice I guess. 06:24:46 !tell Mandevil Not really sure what's wrong. Score files etc. are exposed. 06:24:47 TZer0: OK, I'll let mandevil know. 06:35:22 TZer0: |amethyst has yet to add CUE to CAO scoring, that will be added fairly soon, I think 06:35:29 we were waiting on CDO getting 0.20 as well 06:35:50 so there should be nothing that you have to do, we'll get it added to the scoring pages eventually 07:38:44 How come that when I play console, everything looks fine, but when I spectate someone else I have (sometimes, not always) walls and floor glyphs in blue instead of in darkgrey. It could have to do something with the options bold_brightens_foreground (true), best_effort_brighten_background (false) 07:39:29 But then I wonder why it only happens when I'm spectating. I had this same problem when playing after the options had been introduced, and setting these fixed it. I didn't try spectating back then,. 07:42:14 Kenran: When you're spectating you get their options 07:43:13 Yes, this is a bit unsatisfactory 07:51:44 Pinkbeast: ah thanks, that explains why some people have it correct. They are probably not playing tiles either. 07:53:02 There was a suggestion (by, er, me) that the default should change back with some prominently displayed dark blue and bold black text at the start of a game telling you how to fix it up if you can't see half the text 07:53:46 yeah, I think that's going to be addressed, hopefully in 0.21 07:54:11 johnny0 is working on a nice patch to have a color "calibration" display that might help this happen 07:55:00 untill then spectating and footv will have whatever options the player has 08:03:50 Colour calibration> I was hoping the best (having full options control ingame a la NetHack) was not becoming the enemy of the good (being able to toggle this one option ingame so you can see what you're doing), so that's good news 09:21:41 I'm not really competent with this kind of stuff, but so far everyone I know that plays console actually had to change the default options to make the blue go grey again. Who would be hurt if just the boolean defaults for those options changed? 09:22:31 Side question: gammafunk, you're using OBS on linux right? What method do you use to capture your terminal? 09:28:43 Kenran: new users whose terminal does not brighten colours on bold 09:30:45 Kenran: If your terminal doesn't think bold-black is dark grey but instead black, you can't see those walls etc _at all_ 09:31:54 Kenran: here's an example of a terminal configured that way: https://www.getlucked.com/crawl/img/ccc_animation_no_brighten.gif 09:31:56 my scoring page on CAO hasn't been updated since May 26th, played on both CUE and CXC - any reason for this? 09:32:08 same thing has happened with a friend's scoring page 09:35:15 Hello world, could anyone help me debug an error I'm getting when I try to build from source on Mac OS X? (https://pastebin.com/ETPZ9x6p) 09:36:16 johnny0: I don't understand most of it, but thanks :D The disappearing black-on-darkgrey and vice versa is one of the problems, right? 09:38:04 Kenran: yeah 09:38:27 Kenran: Right, because if your terminal thinks bold black is black not dark grey, then black-on-bold-black is black-on-black not black-on-darkgrey 09:39:15 honestly, sorting this out for me didn't take long, but it was more trial-and-error, since I didn't really know what the new options were about and never worried about how my terminal handled things 09:39:45 to even understand what they do and what values to use, there's some basic understanding of that needed 09:39:54 I think johnny0 was helping me back then himself :D 09:40:15 Kenran: Thing is, in the new world, if it doesn't work, it looks ugly but you can play. In the old world if it doesn't work for you, you can't play. So that's why the default changed. 09:40:35 Pinkbeast: yeah I got that now :) 12:02:27 !tell creyes what build command are you calling exactly? What version of OS X? Also, as a sanity check what happens when you type `which yacc` at the command line? 12:02:30 advil: OK, I'll let creyes know. 12:04:22 is my internet even working 12:05:32 your tubes are clogged 12:21:12 -!- Tickenest_ is now known as Tickenest 12:41:33 -!- Tickenest_ is now known as Tickenest 12:44:49 -!- amalloy_ is now known as amalloy 12:53:25 how do i change which glyph the player uses (not species glyph) 13:03:44 you know, i don't think you can 13:04:16 wow 13:05:31 this seems to be the function that figures out what glyph to use for the player: 13:05:31 !source update_player_symbol 13:05:34 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/player.cc#L570 13:22:23 amalloy: mon_glyph ^= player : x 13:22:47 aha 13:23:00 player confirmed for the real monster 13:23:34 &rc rchandra 13:23:36 http://crawl.jorgrun.rocks/rcfiles/crawl-git/hyperrchandra.rc 13:39:41 !vault minmay_item_behind_shaft 13:39:43 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/variable/mini_features.des#L2820 13:39:50 delet this 13:41:22 fr: delete every file with the string "minmay" in the repo 14:17:53 what do i (a non-dev) do when i have a patch/PR for someone else's mantis bug report? 14:18:09 -!- corax is now known as cut1less 14:26:10 https://underhound.eu:8080/ does only show a loading screen. 14:49:40 cut1less: you could open a PR on github with a link to the mantis issue, and/or comment on the mantis issue with a link to the PR; or if you'd rather not use github you could just comment on the mantis issue with a patch 14:54:57 amalloy: are you planning to upload crawl videos again anytime? I really liked them. 15:11:35 Kenran: i haven't really had much appetite this year for crawl or for recording. i'd like to get back to it sometime, because it was fun, but it's harder to do than just playing games and relaxing 15:12:30 i actually played my first game of crawl in over six months this weekend, so who knows 15:12:35 lol 15:12:56 didn't finish it in time for t, or i'd have a 100% t winrate 15:13:38 got such disgustingly good item drops as well as an early trog altar. a game as lucky as that could have been |amethyst's fourth win 15:14:04 !gamesby neil 15:14:05 neil has played 14328 games, between 2011-08-10 11:04:55 and 2017-06-05 00:53:08, won 3 (0.0%), high score 1719705, total score 38646261, total turns 40091178, play-time/day 0:59:57, total time 88d+13:45:28. 15:14:54 1learn add 3_divby_14328[1] 0 15:16:02 why so many games 15:16:08 at least i'll always have him outpaced 15:24:26 !lg amalloy 15:24:28 417. amalloy the Slayer (L27 VSAs of Trog), escaped with the Orb and 3 runes on 2017-06-12 06:49:07, with 1845893 points after 60056 turns and 6:57:14. 15:26:43 !lg . -2 15:26:44 416/417. amalloy the Martial Artist (L13 OpTm of Okawaru), slain by Sonja (a +2 short sword of distortion) on Lair:3 on 2017-05-13 10:00:33, with 24801 points after 25278 turns and 2:39:44. 15:27:04 oh, i guess i did play some other games 15:55:48 !messages 15:55:49 No messages for TZer0. 16:01:01 TZer0: you noticed underhound.eu:8080 is not working ? 16:18:11 !readall qazlal 16:18:12 The violent god of natural disasters. Gives SH, temporary resists, RMsl, and clouds surrounding you in a small radius (and immunity to all clouds). See also {upheaval}, {disaster area}, and {elemental force}. / While worshipping Qazlal at least one * of piety, the storm surrounding you makes noise every turn. The amount of noise grows as you gain piety, reaching the same volume as fireball at max... 16:21:30 -!- Dracunos_ is now known as Dracunos 17:01:12 ??vaults 17:01:13 vaults[1/2]: Accessed from the main dungeon somewhere between levels 13 and 14. Five levels deep. A large group of guards await on the bottom floor (see {vaults:5}), where you can find several item vaults and a silver rune. There are also ordinary {vault}s. 17:01:22 ??vaults:5 17:01:23 vaults:5[1/4]: You start off surrounded by a ring of 24 {vault guard}s a couple of spaces away. Also scattered throughout the level are many of the game's tougher enemies; be prepared for all sorts of nasty dragons, giants and liches. Have fun! 17:01:53 sorry wrong channel again 18:02:01 Not sure who to report this too, but http://crawl.akrasiac.org/scoring/players/palyth.html is showing duplicate games for 63, 65, 67, 69 (all from CJR). 18:33:20 !lastgames palyth won misk 18:33:24 Last 10 games for palyth: L26 MiSk^Oka (winning), L26 MiSk^Oka (winning) 18:33:45 !log palyth won misk 18:33:45 2. palyth, XL26 MiSk, T:68300: https://crawl.jorgrun.rocks/morgue/palyth/morgue-palyth-20170514-223643.txt 18:33:45 !log palyth won misk -2 18:33:45 1/2. palyth, XL26 MiSk, T:68300: https://crawl.jorgrun.rocks/morgue/palyth/morgue-palyth-20170514-223643.txt 18:34:48 johnstein: oh, will you be able to make a tstats graph? 18:34:48 !lg * recent won cjr trunk 18:34:56 581. Banquo the Conqueror (L27 MiBe of Trog), escaped with the Orb and 3 runes on 2017-05-29 20:11:59, with 1661035 points after 83584 turns and 7:09:50. 18:34:59 amalloy: the duplication is a known issue with cjr 18:35:03 okay 18:35:20 gammafunk: I will start pulling data tonight. 18:35:27 ok, cool 18:36:10 I was hoping to figure out a programmatic way to pull the data but wasn't able to. I'll just overload sequell 18:44:18 I see that we've highlighted the comparison to immediately previous version tourney a lot 18:44:33 but the winter ones seems to systematically get more participation in terms of number of players 18:44:45 !tstats 18:44:51 !tstats t0.19 16 18:44:55 !tstats t0.18 16 18:44:57 !tstats t0.17 16 18:45:05 but in terms of play time I think we're still higher 18:45:10 Stats after 16 days (t): 2997 players, 1062 runers, 543 winners, 1685 wins, 73356 games, winrate 2.30%, total player time 4y+187d+2:02:23. 18:45:48 Stats after t0.19 days (16): 0 players, 0 runers, 0 winners, 0 wins, 0 games, winrate NaN, total player time 0. 18:45:53 Stats after t0.18 days (16): 0 players, 0 runers, 0 winners, 0 wins, 0 games, winrate NaN, total player time 0. 18:45:57 Stats after t0.17 days (16): 0 players, 0 runers, 0 winners, 0 wins, 0 games, winrate NaN, total player time 0. 18:46:04 !lg manman tall s=cv 18:46:06 270 games for manman (tall): 98x 0.20, 75x 0.19, 40x 0.15, 29x 0.18, 24x 0.14, 4x 0.12 18:46:18 oops, messed up those queries 18:50:09 .gfgm map~~its_a_trap s=name,char,xl 18:50:10 2 games for * (kmap= map~~gammafunk splat map~~its_a_trap): Ganima (HOGl (27)), roguesci (VSTm (27)) 18:55:48 !tell ldierk Fixed now. 18:55:49 TZer0: OK, I'll let ldierk know. 18:55:51 gammafunk: did you guys set up any good nick filters that helped sufficiently explain the winrate increase? 18:56:42 nick filters? 18:57:22 we don't really know what caused the winrate increase, so we couldn't really define a nick to explain it 18:57:43 I did create goodplayerspre19 and goodplayerspre20 so we could use something like 18:57:56 !lg !bot !@goodplayerspre20 !boring t / won 18:58:05 !lg !bot !@goodplayerspre19 !boring t0.19 / won 18:58:23 570/59088 games for bot (!@goodplayerspre20 !boring t): N=570/59088 (0.96%) 18:58:29 to show how the winrate increase affected non-goodplayers as much or more than it affected everyone 18:59:04 441/62953 games for bot (!@goodplayerspre19 !boring t0.19): N=441/62953 (0.70%) 18:59:18 this is where I wish I could get a programmatic query for !tstats 18:59:20 but no, we don't know why it's happened, and I will talk about the "Mysterious Winrate Inflation" in my blog post a bit 18:59:53 and another hint seems to be that the increase is more seen for games progressing between xl9 to xl18 compared to start to xl9 and xl18 to win 18:59:55 I'll probably take a stab at an attempt 19:00:28 Well, you really want to avoid just firing off random queries without thinking what each query means 19:00:41 I agree 19:00:43 It's a lot better to start with a hypothesis that can be turned into a query 19:01:07 I was going to pull data for each God separately. maybe for races too 19:01:21 second priority 19:01:26 than to make a bunch of queries and look for bumps, mostly since the increase is significant but not incredibly large 19:01:33 agreed 19:01:42 but yeah, certainly if notice anything specific, feel free to let us know 19:02:18 right, if you subcategorize a lot you'll run into the issue where the numbers involved are really small 19:02:37 yea. there's a nice fancy stats term 19:02:49 that I can't think of 19:03:31 t-test. gaussian. statistically significant. p-value. standard deviation. mean. median. cumulative distribution function. beta distribution. markov... 19:03:56 a small sample of some of the thousands of tags I will add to the blog post to increase its search relevance 19:03:59 it'd be interesting to split the t data into two sets, one to make hypotheses from and the other to test if they are actually true 19:10:02 !kw recentish 19:10:02 Keyword: recentish => cv>=0.17 19:10:02 !kw recentish cv>=0.18 19:10:04 Defined keyword: recentish => cv>=0.18 19:10:05 !kw recent 19:10:06 Keyword: recent => cv>=0.18 19:10:12 !kw recent cv>=0.19 19:10:15 !kw current 19:10:22 Defined keyword: recent => cv>=0.19 19:10:23 Keyword: current => cv>=0.19 19:10:31 !kw current cv>=0.20 19:10:39 Defined keyword: current => cv>=0.20 19:13:51 Sorry i lost internet after posting the dupe games, is this something i should throw in hte issue tracker, or does someone jsut need to re-update the db? 19:17:42 Palyth: I don't know if we have a bug for it yet 19:17:52 if not, a bug report would be good, let me go check right quick 19:17:56 ok 19:18:38 !bug 10951 19:18:39 https://crawl.develz.org/mantis/view.php?id=10951 19:18:59 yeah, looks like we have a bug report 19:19:08 cool, as long as someone knows about it :) 19:20:43 so this affects Sequell as well not just the scoring page. not sure if the fix from Brannock would cover that as well 19:20:58 !lg . won crace=Ba s=char 19:20:59 2 games for Palyth (won crace=Ba): 2x BaGl 19:21:03 i should only have 1 ba win 19:21:12 yes, it affects both, since the games are in two different logfiles 19:21:32 so we need to remove the duplicates from the stable log files (the games duplicated were always trunk, I think) 19:21:42 and then rebuild CAO, and I'm not sure what happens with sequell 19:21:47 but we'll talk to the snark about it 19:21:52 so should i add something to the issue, or would Brannock's proposal fix this as well? 19:22:22 well Brannock isn't quite fully proposing something there 19:22:35 he's just mentioning how CAO will need a rebuild 19:22:51 the fix will be made on cjr as well as CAO/Sequell, but what he says there is accurate 19:22:55 i assume that was his fix, and that it hadnt been done yet cause of hte new scoring page 19:23:27 Well I'm saying that he's not really proposing a fix, just reporting some things that will happen 19:23:39 But yeah, we know what we're going to have to do there 19:23:51 So should i add something about Sequell (really all i am getting too)? 19:24:10 If not, no worries, you guys got this in hand. 19:26:42 Palyth: yeah, feel free to add a note documenting specifically that games are duplicated in sequell 19:26:58 we'll not forget, but not a bad idea to document it for people who aren't aware 19:27:06 ok 19:28:04 Done. 19:28:54 thanks 20:19:05 if i want to reglyph ranged weapons, what would rc line look like? item_glyph += launcher:* ? 20:19:32 item_glyph += ammo:* ? 20:19:47 or do i do each item type individually 20:28:12 i want to say "item_glyph ^= ranged:*" for ranged weapons and "item_glyph ^= missile:*" for ammo/missiles 20:29:26 i want thrown weapons and launcher ammo to have different glyphs, so... hmmm 20:51:31 i suppose you could also try "item_glyph ^= {throwable}:*", too. as long as it comes after the one for missiles, it should probably take priority 20:53:18 these non-canonical glyphs... 20:53:24 what would rchandra say 21:03:44 none of throwable, throwing, tomahawk, tomahawks? work for the toma i found 21:04:38 also how to change the glyph for invis monsters 21:04:55 since im using {} for launcher/ammo now 21:08:07 (ranged and missile do work) 21:08:34 who do I contact about CAO/scoring not recording my online games from CXC/CUE? 21:09:22 I can get my actual stats from the bots but my page hasn't updated since the 26th of May 21:16:08