01:21:07 <Lantell> Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-3260-gdd6b58b38f
01:44:35 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-85-g4d60c92: Document python-is-python3 for Ubuntu 10(13 days ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/4d60c92aebd0
01:45:09 <Cheibriados> 03Elliot Dronebarger02 {Aidan Holm} 07* 0.26-a0-86-gd5eb699: Added SDL_image to dependency requirements in INSTALL.md 10(13 days ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/d5eb6995ac9f
01:56:31 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-86-gd5eb6995ac (34)
01:58:28 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-87-g31e9aea: Trolls and Kobold want unlimited chunks too 10(4 weeks ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/31e9aea6bc67
02:01:31 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-87-g31e9aea6bc (34)
02:08:26 <Cheibriados> 03Elliot Dronebarger02 {Aidan Holm} 07* 0.26-a0-88-g5ee4766: Add habitat information to monster description popup (#11224) 10(4 weeks ago, 7 files, 169+ 10-) 13https://github.com/crawl/crawl/commit/5ee47661b98d
02:27:33 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-88-g5ee47661b9 (34)
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-89-g09c6a6c: Extract view_dungeon() function 10(5 days ago, 1 file, 57+ 43-) 13https://github.com/crawl/crawl/commit/09c6a6c846d7
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-90-g7a3809f: Remove unused slideout animation 10(3 days ago, 2 files, 0+ 30-) 13https://github.com/crawl/crawl/commit/7a3809fd8f55
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-91-g5a24836: Remove unused checkerboard animation 10(3 days ago, 2 files, 0+ 23-) 13https://github.com/crawl/crawl/commit/5a2483679f2b
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-92-g49e1a6f: Extract map tile updating from show_update_at() 10(3 days ago, 5 files, 41+ 16-) 13https://github.com/crawl/crawl/commit/49e1a6fa648f
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-93-g8a085ae: Extract _viewwindow_should_render() helper function 10(3 days ago, 1 file, 14+ 8-) 13https://github.com/crawl/crawl/commit/8a085aef401a
02:33:31 <Cheibriados> 03Aidan Holm02 07[view-dungeon] * 0.26-a0-94-g9804897: Remove global crawl_view buffer 10(2 days ago, 4 files, 5+ 15-) 13https://github.com/crawl/crawl/commit/980489766d16
02:35:04 <gammafunk> I saw some of those weird animations when I was looking into using animations for absolute zero
02:35:39 <gammafunk> in the end I just went with a flash since the animation code appears to be player-based and didn't really have a way to focus animation on an arbitrary point in los
02:40:07 <aidanh> they look pretty cool, but i'd rather not have the cruft
03:22:28 <rchandra> the player banner checklist is awesome, whoever did that
03:30:36 <Kelbell> Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3260-gdd6b58b38f
03:32:24 <travis-ci> The build passed. (master - 5ee4766 #13639 : Elliot Dronebarger): https://travis-ci.org/crawl/crawl/builds/700510149
03:44:35 <Kelbell> Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-2808-gb728f3cfe5
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-89-g09c6a6c: Extract view_dungeon() function 10(5 days ago, 1 file, 57+ 43-) 13https://github.com/crawl/crawl/commit/09c6a6c846d7
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-90-g7a3809f: Remove unused slideout animation 10(3 days ago, 2 files, 0+ 30-) 13https://github.com/crawl/crawl/commit/7a3809fd8f55
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-91-g5a24836: Remove unused checkerboard animation 10(3 days ago, 2 files, 0+ 23-) 13https://github.com/crawl/crawl/commit/5a2483679f2b
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-92-g49e1a6f: Extract map tile updating from show_update_at() 10(3 days ago, 5 files, 41+ 16-) 13https://github.com/crawl/crawl/commit/49e1a6fa648f
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-93-g8a085ae: Extract _viewwindow_should_render() helper function 10(3 days ago, 1 file, 14+ 8-) 13https://github.com/crawl/crawl/commit/8a085aef401a
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-94-g9804897: Remove global crawl_view buffer 10(2 days ago, 4 files, 5+ 15-) 13https://github.com/crawl/crawl/commit/980489766d16
05:02:36 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-95-g294706d: Fix cells out of LOS darkened for mesmerise on local tiles 10(11 minutes ago, 1 file, 3+ 4-) 13https://github.com/crawl/crawl/commit/294706da8a5d
05:46:12 <Cheibriados> New branch created: pr-1404-partial (5 commits) 13https://github.com/crawl/crawl/tree/pr-1404-partial
05:46:12 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07[pr-1404-partial] * 0.26-a0-96-g3dc428b: Remove references to Forest and other removed things 10(5 weeks ago, 11 files, 40+ 42-) 13https://github.com/crawl/crawl/commit/3dc428ba9b08
05:46:12 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07[pr-1404-partial] * 0.26-a0-97-g65d1636: Fix Pan entry placement in TAG_MAJOR_VERSION 35 10(5 weeks ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/65d16368d843
05:46:12 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07[pr-1404-partial] * 0.26-a0-98-g56b65ce: Remove two obsolete mutations from TAG 35 10(5 weeks ago, 1 file, 3+ 1-) 13https://github.com/crawl/crawl/commit/56b65ceb29a8
05:46:12 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07[pr-1404-partial] * 0.26-a0-99-g8d2668a: Delete mottled draconian on TAG_MAJOR_VERSION upgrade 10(5 weeks ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/8d2668ac7942
05:46:12 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07[pr-1404-partial] * 0.26-a0-100-gfd58452: Remove some vestigial blade branch logic 10(5 weeks ago, 1 file, 0+ 18-) 13https://github.com/crawl/crawl/commit/fd584523ee7b
06:09:26 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-95-g294706da8a (34)
06:46:56 <travis-ci> The build passed. (pr-1404-partial - fd58452 #13642 : Alex Jurkiewicz): https://travis-ci.org/crawl/crawl/builds/700535569
06:49:53 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-96-g3dc428b: Remove references to Forest and other removed things 10(5 weeks ago, 11 files, 40+ 42-) 13https://github.com/crawl/crawl/commit/3dc428ba9b08
06:49:53 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-97-g65d1636: Fix Pan entry placement in TAG_MAJOR_VERSION 35 10(5 weeks ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/65d16368d843
06:49:53 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-98-g56b65ce: Remove two obsolete mutations from TAG 35 10(5 weeks ago, 1 file, 3+ 1-) 13https://github.com/crawl/crawl/commit/56b65ceb29a8
06:49:53 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-99-g8d2668a: Delete mottled draconian on TAG_MAJOR_VERSION upgrade 10(5 weeks ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/8d2668ac7942
06:49:53 <Cheibriados> 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.26-a0-100-gfd58452: Remove some vestigial blade branch logic 10(5 weeks ago, 1 file, 0+ 18-) 13https://github.com/crawl/crawl/commit/fd584523ee7b
07:12:53 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-100-gfd584523ee (34)
07:55:45 <Cheibriados> New branch created: view-refactor (1 commit) 13https://github.com/crawl/crawl/tree/view-refactor
07:55:45 <Cheibriados> 03Aidan Holm02 07[view-refactor] * 0.26-a0-101-ga6bfcf2: Make screen cell's flash_colour property tiles-only 10(67 minutes ago, 2 files, 9+ 9-) 13https://github.com/crawl/crawl/commit/a6bfcf29fca6
08:04:44 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-101-gebf84c8: Fix inadvertent copies in loops 10(4 minutes ago, 5 files, 8+ 8-) 13https://github.com/crawl/crawl/commit/ebf84c8062ee
08:04:44 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-102-ga1c194d: Remove tautological comparison 10(40 seconds ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/a1c194d6e7ac
08:16:52 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-102-ga1c194d6e7 (34)
09:36:58 <Euph0ria> CAO current game 0.24, displays an Orc where none is visible even after moving around.  https://imgur.com/EoU9C05.png
09:54:04 <aidanh> Euph0ria: i remember that being a tricky bug we've not really been able to reproduce
11:56:34 <Cheibriados> 03Aidan Holm02 07* 0.26-a0-103-g486a359: Remove unnecessary tiles.update_tabs() call 10(3 minutes ago, 1 file, 0+ 7-) 13https://github.com/crawl/crawl/commit/486a359f8a79
12:01:36 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-103-g486a359f8a (34)
12:18:00 <advil> !kw t
12:18:01 <Sequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
12:18:01 <NotSequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
12:18:10 <advil> does anyone know why cv=0.25|0.25-a?
12:18:55 <advil> there's a bunch of cwz trunk milestones that that is causing to be included
12:22:57 <advil> !lg * t cv=0.25-a src=cpo x=file
12:22:57 <Sequell> No games for * (t cv=0.25-a src=cpo).
12:22:58 <NotSequell> 6. [file=cpo/dcss-logfile-0.25] MrDizzy the Slayer (L26 DDFi of Makhleb), escaped with the Orb and 3 runes on 2020-06-12 23:38:53, with 2041126 points after 47172 turns and 2:59:44.
12:23:13 <advil> alexjurk1 any guesses as to why those are in that logfile?
12:24:47 <ebering> I think in the past the reason for including the alpha has to do w/ hong's deploy schedule on cwz?
12:25:20 <advil> yeah, from 0.23, elliptic found it
12:25:25 <elliptic> it's not specifically cwz
12:25:36 <elliptic> a bunch of servers have had issues with this in the past
12:25:48 <elliptic> !lg * t0.23 cv=0.23-a s=file
12:25:50 <NotSequell> No games for * (t0.23 cv=0.23-a).
12:25:58 <Sequell> 14139 games for * (t0.23 cv=0.23-a): 8767x cwz/soup/trunk/logfile, 3287x cwz/soup/0.23/logfile, 1682x cbro/crawl/meta/0.23/logfile, 403x lld/mirror/meta/0.23/logfile
12:43:44 <advil> !lg * t cv=0.25-a src=cpo x=file,cv,vlong
12:43:45 <Sequell> 6. [file=cpo/dcss-logfile-0.25;cv=0.25-a;vlong=0.25-b1-49-g6eddd28] MrDizzy the Slayer (L26 DDFi of Makhleb), escaped with the Orb and 3 runes on 2020-06-12 23:38:53, with 2041126 points after 47172 turns and 2:59:44.
12:43:46 <NotSequell> 6. [file=cpo/dcss-logfile-0.25;cv=0.25-a;vlong=0.25-b1-49-g6eddd28] MrDizzy the Slayer (L26 DDFi of Makhleb), escaped with the Orb and 3 runes on 2020-06-12 23:38:53, with 2041126 points after 47172 turns and 2:59:44.
12:44:16 <advil> alexjurk1 more specifically, any guesses about the versioning there?
12:45:42 <elliptic> !lg * t cv=0.25 src=cpo x=file,cv,vlong
12:45:43 <Sequell> 1275. [file=cpo/dcss-logfile-0.25;cv=0.25;vlong=0.25.0-1-g7eea58d] thesequoiad the Shield-Bearer (L2 FoFi), succumbed to an adder's poison on D:2 on 2020-06-21 16:12:18, with 9 points after 146 turns and 0:00:35.
12:45:44 <NotSequell> 1275. [file=cpo/dcss-logfile-0.25;cv=0.25;vlong=0.25.0-1-g7eea58d] thesequoiad the Shield-Bearer (L2 FoFi), succumbed to an adder's poison on D:2 on 2020-06-21 16:12:18, with 9 points after 146 turns and 0:00:35.
12:46:24 <elliptic> that is what it would look like normally, for comparison
12:46:28 <elliptic> it might just be a matter of whether cpo was after the 0.25.0 tag in the branch or something like that?
12:48:12 <elliptic> i.e. maybe this is just what happens if a server sets up a release branch before it has been actually released and then doesn't update?
12:49:29 <advil> I think cv is a derived field constructed by sequell
12:50:39 <elliptic> yeah, iirc it just looks at vlong
12:51:53 <elliptic> 0.25-b* or 0.25-a* get turned into cv=0.25-a, while 0.25.x gets turned into cv=0.25
12:53:43 <elliptic> this is what I meant in the other channel by improving sequell derived fields so that we could have the t kw distinguish between 0.25-b and 0.25-a
12:54:23 <advil> ah I see
12:54:33 <advil> heh in fact, var rAlphaQualifier = regexp.MustCompile(`-[a-z]`)
12:54:47 <advil> so any letter will do
12:58:37 <elliptic> oh, maybe I'm wrong
12:59:57 <advil> I think it's checking 'v' instead of 'vlong' but it adds -a as long as 'v' matches that regex
13:00:50 <elliptic> what I meant is actually:
13:01:37 <advil> https://github.com/crawl/go-sequell/blob/master/crawl/xlogtools/crawlxlog.go#L132
13:01:46 <elliptic> !lg * t file="cwz/soup/trunk/logfile" x=vlong
13:01:47 <NotSequell> No games for * (t file=cwz/soup/trunk/logfile).
13:01:47 <Sequell> 3681. [vlong=0.25-b1-12-g2958ca1] MaineCoooon the Skirmisher (L5 GrFi), slain by an ogre (a +0 giant club) on D:4 on 2020-06-21 16:44:28, with 128 points after 2025 turns and 0:01:32.
13:02:03 <elliptic> CWZ's config is just very messed up
13:02:13 <elliptic> and looking at vlong doesn't help
13:02:52 <elliptic> !lg * t file="cwz/soup/0.25/logfile" x=vlong
13:02:52 <Sequell> 1937. [vlong=0.25.0-9-g26727b6] neetthing the Skirmisher (L2 DsFi), quit the game on D:1 on 2020-06-21 16:54:54, with 9 points after 302 turns and 0:00:16.
13:02:52 <NotSequell> 1937. [vlong=0.25.0-9-g26727b6] neetthing the Skirmisher (L2 DsFi), quit the game on D:1 on 2020-06-21 16:54:54, with 9 points after 302 turns and 0:00:16.
13:03:47 <elliptic> it looks like their "trunk" version is actually pre-release 0.25 branch
13:04:15 <elliptic> while their 0.25 version is 0.25.0
13:05:31 <elliptic> so this really isn't sequell's fault :P
13:06:22 <elliptic> I had assumed that CWZ's trunk was just on really old trunk, not really old 0.25 branch
13:10:13 <advil> hm yeah, can confirm that the raw logfile looks messed up
13:10:59 <advil> this is going to be a headache for someone
13:12:05 <elliptic> I think the logfile is probably fine? and the messup is that they just don't have trunk
13:12:33 <elliptic> !tstats 16 t
13:12:41 <NotSequell> Stats after 16 days (t): 2206 players, 779 runers, 431 winners, 1115 wins, 37351 games, winrate 2.99%, total player time 2y+108d+2:44:43.
13:12:44 <Sequell> Stats after 16 days (t): 2404 players, 838 runers, 441 winners, 1127 wins, 40507 games, winrate 2.78%, total player time 2y+141d+22:04:48.
13:12:47 <elliptic> !tstats 16 tfixed
13:12:54 <advil> well, yeah it's fine at a formatting level but it means that their trunk is on the wrong branch
13:12:55 <NotSequell> Stats after 16 days (tfixed): 2206 players, 779 runers, 431 winners, 1115 wins, 37351 games, winrate 2.99%, total player time 2y+108d+2:44:43.
13:12:59 <Sequell> Stats after 16 days (tfixed): 2204 players, 778 runers, 430 winners, 1114 wins, 37343 games, winrate 2.98%, total player time 2y+107d+14:51:44.
13:13:20 <elliptic> !lg * tfixed cdo
13:13:21 <Sequell> No games for * (tfixed cdo).
13:13:21 <NotSequell> 4. scummos the Cudgeler (L4 OgNe of Okawaru), slain by a scorpion on D:4 on 2020-06-15 20:22:08, with 113 points after 3205 turns and 0:05:40.
13:14:44 <elliptic> I think just using tfixed in sequell queries if you care about matching things with the scripts should be fine (aside from the cdo config mess)
13:25:12 <PleasingFungus> hm
13:25:28 <PleasingFungus> in test_hit, we check                  if (randomise_ev)        ev = random2avg(2*ev, 2);
13:26:02 <PleasingFungus> this means we're actually comparing double the defender's EV to the attacker's to-hit, right?
13:27:11 <PleasingFungus> very odd
13:27:29 <PleasingFungus> really wonder if someone misunderstood what random2avg did
13:30:30 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.26-a0-103-g486a359 (34)
13:40:43 <advil> it looks like the 2* comes from a very different context
13:41:22 <advil> i.e. it's not about the random2avg
13:41:24 <advil> %git 9d08f109e0a5
13:41:24 <Cheibriados> 07elliptic02 * 0.10-a0-2269-g9d08f10: Fix some to-hit code. 10(9 years ago, 1 file, 24+ 19-) 13https://github.com/crawl/crawl/commit/9d08f109e0a5
13:41:43 <advil> that moves an earlier *2 (not sure why that exists) into that call
13:42:04 <PleasingFungus> huh
13:42:09 <PleasingFungus> good archaeology
13:42:49 <advil> %git 5e534f531
13:42:49 <Cheibriados> 07Cryp71c02 * 0.10-a0-2222-g5e534f5: Fix EV Scaling and Block / Dodge order 10(9 years ago, 1 file, 55+ 57-) 13https://github.com/crawl/crawl/commit/5e534f531555
13:43:26 <PleasingFungus> "fixes an ev *= 2 that has wild balance impacts"
13:43:30 <advil> indeed
13:43:34 <advil> fixes it by adding it?
13:43:52 <advil> must be some earlier history
13:44:43 <PleasingFungus> the past is a scary place
13:47:20 <advil> looks like Cryp71c did a whole bunch of work on attacks around then, so I wonder if earlier he removed that *= 2 and then readded it in that commit
13:49:04 <elliptic> I mean, "x = random2avg(2*x,2)" is a perfectly reasonable thing to do to randomize x
13:49:44 <elliptic> since it keeps the same mean (up to an additive constant)
13:50:37 <elliptic> but yes that was a terrifying time in crawl history
13:51:13 <elliptic> (I definitely don't remember any details)
14:07:40 <elliptic> hm, I just spent over 300 turns on a level with a gauntlet without getting any messages about it other than the one on entering the level
14:07:55 <PleasingFungus> i hope i didn't break that somehow
14:07:59 <elliptic> I'm pretty sure it's not something weird about my rc since I've heard gauntlet messages before
14:08:02 <PleasingFungus> with the fix for the cloud thing
14:08:07 <elliptic> yeah that's what I was wondering about
14:08:45 <elliptic> I didn't save or enter/exit the level or anything
14:08:45 <elliptic> if that's relevant
14:08:45 <PleasingFungus> i don't think my change should affect it, but i can imagine ways i might have messed it up...
14:09:33 <elliptic> also wow this gauntlet has 3 unrands in it (but I can only get one)
14:09:41 <PleasingFungus> totally unrelated; how sad would we be if umbra/halos applied a fixed delta to EV instead of a random roll? the random roll doesn't seem to change the dynamics very much, but it does make displaying accuracy completely truthfully much harder
14:09:48 <PleasingFungus> i recommend taking morg
14:10:02 <elliptic> morg sounds good on my DsBe
14:11:21 <elliptic> I have vague recollections of halos being a bit weird for some reason and being much stronger than one might expect against high EV monsters
14:11:36 <elliptic> but I don't remember exactly why it worked out this way
14:12:50 <elliptic> like TSO feels like it increases accuracy by a huge amount a lot of the time in practice, but I glanced at the code the other day when you brought up accuracy and I couldn't figure out why this was the case
14:13:05 <elliptic> so I guess I'm saying I don't know :P
14:13:07 <PleasingFungus> it seems like there shouldn't be much difference between 2d(26) and 2d(30 - 1d8) or w/e
14:13:35 <PleasingFungus> especially since it's a contested roll
14:14:15 <PleasingFungus> but yeah my confidence isn't super high either
14:14:15 <PleasingFungus> i'm just going to cheat for now and hope for the best
14:14:51 <elliptic> right, what I mean is I don't understand why 2d30 vs 2d(30 - 1d8) feels like such a large difference so I'm worried that there's some effect of halo that I don't understand
14:15:41 <elliptic> and I would probably do some actual simulating with wizmode, since I'm pretty sure that halo does help accuracy significantly (or did at some point in the past?)
14:23:59 <elliptic> I guess I'm thinking about the formula a little more and maybe it makes sense to me now? since numbers aren't actually that high, right?
14:25:13 <elliptic> a lot of monsters are in the 8-15 EV range, not so much 30
14:25:30 <elliptic> and a lot of players have pretty low accuracy, like 5 or whatever
14:26:05 <elliptic> and at that stage 2d12 vs 2d(12-1d8) is actually pretty huge for low accuracy players
14:27:12 <elliptic> I expect thinking about it that just subtracting the mean instead of 1d8 is going to be a TSO nerf, though maybe not a huge one
14:27:31 <elliptic> and that you might want to subtract a bit more than the mean
14:28:54 <elliptic> basically a low accuracy player would rather face a range from 2d4 to 2d11 rather than always facing 2d(7.5)
14:30:08 <elliptic> PleasingFungus: I think aside from this maybe being a small nerf to halo it probably would be fine though
14:30:44 <elliptic> and could be tweaked (e.g. subtract 5 instead of 1d8)
14:30:52 <elliptic> actually, hm
14:31:54 <elliptic> if I understand things currently then removing the randomness there and making it be a flat 5 would mean that 5 EV monsters and 0 EV monsters would be identical inside a halo
14:32:01 <elliptic> that is probably why the randomness exists, in fact
14:32:43 <elliptic> (and removing the randomness would be a buff when facing 5 EV monsters obviously, so I take back what I said about it being a nerf - it would really depend on monster EV)
14:33:26 <elliptic> the main point I want to make though is that monster EV really matters for thinking about these things, and one should think about the 2 EV case more than the 30 EV case since there are more monsters with 2 EV :P
14:34:26 <elliptic> and possibly a good exercise before doing anything too serious with accuracy/evasion formulas would be to get some data about monster EV distribution at various parts of the game
14:34:50 <advil> floraline possibly Kelbell isn't reporting 0.25 crashes
14:55:28 <Cheibriados> 03advil02 07* 0.26-a0-104-gef3e9dd: Remove an ASSERT for unlinked temporary items 10(3 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/ef3e9dd61fec
14:56:06 <Cheibriados> 03advil02 07[stone_soup-0.25] * 0.25.0-10-g039870e: Remove an ASSERT for unlinked temporary items 10(4 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/039870ee82cf
15:06:20 <advil> can anyone else access cko rebuild? it seems to be unloadable for me
15:06:37 <advil> 504 Gateway Time-out
15:07:24 <advil> oh there it goes
15:07:44 <Kelbell> Stable (0.25) branch on crawl.kelbi.org updated to: 0.25.0-10-g039870ee82
15:07:59 <advil> floraline this is a super low priority request but I suspect that this script shouldn't be behind caching
15:12:52 <PleasingFungus> elliptic: good point about getting more data about monster EV distribution
15:13:00 <PleasingFungus> agree re it being a slight nerf just to use the mean of halo, i was thinking the same
15:14:38 <PleasingFungus> very good point about it being a nerf to haloed monsters with <= 5 EV... i wonder how big an effect that actually is
15:16:37 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-104-gef3e9dd61f (34)
15:18:22 <PleasingFungus> wait wait wait i forgot
15:18:22 <PleasingFungus> halo/umbra doesn't touch ev at all! it only adjusts to_hit
15:32:51 <PleasingFungus> so the question is actually about very low accuracy attacks while in umbra
15:33:22 <PleasingFungus> let's say that the player is wielding a +2 exec axe with 5 dex, 10 fighting, 15 axes (very low all around), and they're trying to hit a servitor
15:34:51 <PleasingFungus> oh, then i need to calculate aevp...
15:36:15 <PleasingFungus> -6 base, +2 slay, +10 from fighting, +15 from axes, let's go wild and give -10 from aevp
15:36:35 <PleasingFungus> *-6 weapon acc
15:36:43 <PleasingFungus> then another 15 base (!) and 2 from dex
15:37:03 <PleasingFungus> so it's like 36 total before umbra?
15:37:30 <PleasingFungus> alternate question: does this make dith/yred umbra too strong? are there monsters that will have extremely low acc?
15:38:07 <PleasingFungus> eh, they have 18 + HD * 1.5 to-hit (or more)
15:39:13 <PleasingFungus> yeah, don't think making halo or umbra constant really hurts anything
15:42:48 <PleasingFungus> i'd also like to move some weird later to-hit modifiers (rmsl, form to-hit UC modifiers, confusing touch to-hit bonus) out of separate rolls and into the big to-hit calc roll
15:43:24 <PleasingFungus> haven't thought that much about the effects though. I think it'd make them slightly weaker if their max value wasn't changed?
16:00:28 <elliptic> oh, halo modifier on the other side? in that case things are a bit different
16:01:49 <elliptic> then there isn't the weirdness with low EV monsters being the same as 0 EV at least
16:04:26 <elliptic> so it is something like 1d(to_hit+4.5) vs 1d(to_hit+1d8)? my guess is that the 1d8 side is still stronger, so you might want to replace 1d8 by 5 or something like that
16:04:43 <PleasingFungus> yeah
16:05:08 <elliptic> I'm going to go back to "I'm a bit surprised halo is so good against high EV monsters" though, so I don't really feel like I have good grasp of the details of what is going on
16:05:54 <PleasingFungus> hm
16:07:50 <PleasingFungus> what do you think about my other proposal of moving the separate rolls for special to-hit effects (confusing touch, form UC bonuses, rmsl) to be constant modifiers before the main roll?
16:10:45 <elliptic> is this again just a matter of 1d(x+1dy) vs 1d(x+(approx y/2)) on the to-hit side of things?
16:12:31 <elliptic> basically all these changes are doing is decreasing the variance on the to-hit roll, which probably amounts to a minor nerf to trying to use low-accuracy attacks against high-EV enemies and a minor buff to the other way around
16:13:32 <elliptic> I'm not sure how large the effects end up being because I don't have a good sense of how large all the different parts of the formula tend to be relative to each other
16:14:54 <elliptic> this story is actually sort of similar to how melee damage works against AC
16:15:03 <elliptic> the distribution for melee damage has a very long tail
16:15:21 <elliptic> where occasionally you do a ton of damage because a lot of random multipliers line up
16:16:10 <elliptic> anyway variance could be added in a more uniform manner if desired probably
16:18:49 <elliptic> I basically don't have any strong feelings about the precise form of the accuracy chance formula, I just think we should make sure we have enough parameters to tweak it
16:20:45 <elliptic> e.g. I think it would be stupid to say "let's restrict ourselves to having the accuracy calculation just be 1d(flat to_hit number depending on factors) vs 2d(flat evasion number) and have the only parameters that we are allowed to change be how various effects alter those two flat numbers"
16:20:53 <PleasingFungus> the separate rolls i was talking about were actually quite different
16:21:08 <PleasingFungus> rather than 1d(x+1dy), they're 1dx + 1dy
16:21:40 <PleasingFungus> i agree we shouldn't limit ourselves to just having 1dX vs 2dY if we have a good goal in mind
16:22:27 <PleasingFungus> but it seems very strange to me that very specifically confusing touch and transmuter combat would have a different probability distribution (and a lower-variance one, i think? well, one that hews closer to the center, anyway)
16:22:36 <elliptic> it's probably fine to say "let's have a to_hit number that depends on stuff about the player, and let's have an evasion number that depends on stuff about the monster, and there's an arbitrary function f(to_hit, evasion) that gives the chance of hitting"
16:22:53 <elliptic> but I wouldn't want to simplify things any further for the sake of formulas
16:23:06 <PleasingFungus> sure, sounds fine to me
16:23:36 <elliptic> PleasingFungus: you mean that when using those things, the comparison is 1d(to_hit) + 1d(something) vs 2d(evasion) or whatever the ev side is?
16:23:45 <PleasingFungus> yes
16:24:19 <elliptic> yeah, that's the opposite direction of lower variance than 1d(to_hit+something)
16:24:51 <PleasingFungus> specifically: if using confusing touch, you roll 1d(to_hit) + 1d(you.dex()) vs 2d(EV)
16:25:12 <PleasingFungus> if using a form, you roll 1d(to_hit) + 1d(UC form bonus) vs 2d(EV)
16:25:22 <elliptic> I think if I wanted to simplify the formulas, I would separate the two sides into (to_hit role that depends only on player) and (evasion side that depends only on monster)
16:25:23 <PleasingFungus> (and if you're using both a form and conf touch then you get both bonuses)
16:25:27 <elliptic> maybe this separation is already basically done
16:25:35 <PleasingFungus> it's mostly done
16:25:44 <PleasingFungus> halos/umbras/rmsl are the exceptions, i think?
16:26:17 <elliptic> well, I mean including status effects affecting each side on the respective side
16:26:27 <PleasingFungus> oh, i lied, there's also invisible monsters and MUT_TRANSLUCENT_SKIN
16:26:49 <PleasingFungus> aside from those things that i mentioned, i think stuff on the attacker only affects accuracy and stuff on the defender only affects EV
16:27:01 <elliptic> I guess you mean that the defender being haloed affects the attacker's to-hit
16:27:04 <PleasingFungus> yes
16:27:05 <elliptic> I think that sort of thing is fine
16:27:24 <elliptic> as long as there are two clear sides and we are thinking about one of them as to-hit and the other as evasion
16:27:27 <PleasingFungus> sure
16:27:44 <PleasingFungus> rmsl re-rolls to-hit, so you roll 1d(1d(to_hit)) vs 2d(EV)
16:27:52 <PleasingFungus> on average halving it
16:28:00 <elliptic> and then the next thing I would do is replace each side by the expected value of that entire roll
16:28:24 <elliptic> so then we just have two numbers with no randomness
16:29:16 <PleasingFungus> hm? what would you do with those?
16:29:49 <elliptic> and then I would try to choose a formula for a chance of hitting that roughly approximates the current behavior in a bunch of standard situations (e.g. without random little modifiers)
16:29:49 <elliptic> the f(to_hit, evasion) function I mentioned above
16:31:05 <elliptic> that's the hard part of course, but you could start with something close to the 1d(2*avg_to_hit_roll) vs 2d(avg_ev_roll) comparison that we currently use when there are no other modifiers
16:31:20 <elliptic> and then tweak it as desired
16:31:44 <elliptic> comparing with current behavior in a variety of simulations
16:32:38 <PleasingFungus> sorry, i'm a little confused. this is the path that you would take to simplify away the weird special-case to-hit modifiers?
16:33:05 <PleasingFungus> i should probably push up a branch so that we can discuss more...
16:33:15 <elliptic> basically what I'm saying is that your approach of moving rolls for special effects to be "constant modifiers before the main roll" is fine, but that at the end there are ways to tweak how the "main roll" works
16:33:27 <elliptic> that should be considered
16:33:43 <PleasingFungus> gotcha
16:33:56 <PleasingFungus> i can think more about that... i haven't been thinking a lot about high-level overhaul of the entire system
16:34:08 <PleasingFungus> just trying to get more visibility into the current system first
16:34:30 <elliptic> I mean, I'm not really talking about high-level overhaul either
16:36:12 <elliptic> I'm just observing that the changes you want to make for code simplicity (having the modifiers all apply at the same time in the calculation) will have balance effects, and that to some extent one can try to minimize those balance effects by replacing the final "1dx vs 2dy" comparison with some slightly different function
16:36:43 <elliptic> (having the modifiers all apply at the same time in the calculation and without extra randomization on some of them)
16:40:13 <PleasingFungus> hm
16:40:47 <PleasingFungus> i think you're right, i'm just not full of bright ideas right now... i'll push an initial version at some point (to a branch, not master) and we can talk more
16:41:22 <elliptic> to some extent I don't think it needs bright ideas, just some tedious work with something like fsim
16:42:10 <PleasingFungus> fun...
16:42:41 <elliptic> replace all the modifiers by their expected value just like you wanted to do, leave the 1dX vs 2dY comparison for now, then compare something like fsim before/after in a bunch of assorted sample characters
16:44:08 <elliptic> yeah, it doesn't sound fun to me either, but I'm not sure I know any better approach - maybe instead of actually doing fsim, you can just have debug output for all the relevant modifiers in some situations and have an auxiliary program simulate it
16:45:04 <elliptic> but I'm not a fan of "just trying to get more visibility into the current system first" if it involves making balance changes of unknown significance first
16:47:41 <elliptic> probably to avoid fsim tediousness it would be best to write down the formulas carefully with all details (including e.g. the formula for how a given value of dex affects base to_hit)
16:47:46 <elliptic> as they are
16:48:42 <elliptic> and then should be fairly easy to write an auxiliary script to simulate things at various skills/weapons/status effects/whatever against different EV enemies
16:54:15 <PleasingFungus> i'll poke at it more
17:21:20 <rchandra> !tell Sequell welcome back
17:21:21 <Sequell> Sorry rchandra, I don't know who sequell is.
17:21:21 <NotSequell> Sorry rchandra, I don't know who sequell is.
17:22:27 <Pinkbeast> rchandra: Bots tend not to be self-aware
17:22:50 <elliptic> hm, I just unwielded distortion and nothing happened
17:23:06 <elliptic> I guess there might be no monsters on this level for the teleport to monsters effect?
17:23:25 <elliptic> no, there's at least one
17:24:45 <elliptic> like I got no message at all
17:41:21 <PleasingFungus> i saw someone report that in discord the other day
17:43:09 <PleasingFungus> (no effect on discord unwield)
18:25:26 <Lantell> Unstable branch on underhound.eu updated to: 0.26-a0-104-gef3e9dd61f (34)
18:33:56 <alexjurk1> advil: sorry I can't see what's wrong with the log file
18:34:05 <alexjurk1> or i'm confused by the order of sequell commands and responses
19:13:59 -!- amalloy_ is now known as amalloy
20:04:48 <gammafunk> ??test
20:04:49 <NotSequell> test[1/44]: $(echo ??test)
20:04:49 <Sequell> test[1/44]: $(echo ??test)
20:04:52 <gammafunk> dang
20:05:12 <gammafunk> !lg
20:05:14 <NotSequell> 2. gammafunk the Conjurer (L5 VpCj of Ru), slain by a centaur on D:5 on 2020-06-16 00:45:42, with 204 points after 7249 turns and 0:15:45.
20:05:14 <Sequell> 784. gammafunk the Conjurer (L5 VpCj of Ru), slain by a centaur on D:5 on 2020-06-16 00:45:42, with 204 points after 7249 turns and 0:15:45.
20:05:42 <gammafunk> .gfgk * day
20:05:43 <Sequell> 7. PaxCecilia the Basher (L11 MiFi of Nemelex Xobeh), slain by an orc warrior (a +0 dire flail) on Orc:1 (gammafunk_ghost_wrathful_warriors) on 2020-06-21 18:53:40, with 12471 points after 12239 turns and 1:10:02.
20:05:44 <NotSequell> 7. PaxCecilia the Basher (L11 MiFi of Nemelex Xobeh), slain by an orc warrior (a +0 dire flail) on Orc:1 (gammafunk_ghost_wrathful_warriors) on 2020-06-21 18:53:40, with 12471 points after 12239 turns and 1:10:02.
20:23:18 <Lantell> Stable (0.25) branch on underhound.eu updated to: 0.25.0-10-g039870ee82
20:29:54 <advil> seems like maybe I should shut of NotSequell now?
20:30:16 <advil> snark said give it a few hours but afaik everything is working
20:30:25 <gammafunk> advil: greensnark says he's working on Sequell today (and I see it's back in channel!).
20:30:32 <gammafunk> oh, I guess I messed where he said that
20:30:59 <gammafunk> I've already switched Cerebot back to Sequell (and beem would be using it automatically, since I never switched it)
20:31:28 <gammafunk> we can watch it for a couple more hours I guess
20:31:28 <advil> yeah, I can give it one more time
20:31:35 <gammafunk> advil: snark also told me that he's not sure he can host Sequell long term but he's working on dockerizing it so that it's easier for someone else to host
20:31:39 <advil> ah
20:31:55 <advil> ok, well I guess I sort of know what is involved now
20:32:02 <gammafunk> depending on its requirements, I could certainly host it
20:32:22 <advil> yeah, I don't have a sense for what it would cost with full load + all the data
20:32:55 <advil> do you know about shallott more generally?
20:33:23 <gammafunk> I don't, know
20:33:47 <gammafunk> johnstein had a login (as did CDO), but not sure they still work
20:34:22 <advil> johnstein has still been offloading ttyrecs there fairly recently I believe
20:35:12 <gammafunk> yes, but his login doesn't currently work and he's had to move ttyrecs to a personal HD
20:35:42 <gammafunk> so I guess it would be wise for us to have a solution for the ttyrec issue that doesn't depend on storing them on shalott.org
20:38:49 <advil> yes
20:39:10 <advil> though personally I'm still in favor of "delete" as a solution
20:40:04 <advil> right, but yeah my question was did snark's comment about hosting sequell extend to shalott in general?
20:50:51 <gammafunk> advil: I guess we could ask, but I sort of assumed that to be the case
20:51:32 <gammafunk> agree that we might have to consider deletion of sufficiently old ttyrecs
21:52:58 <alexjurk1> advil: ping about some log file issue you saw?
21:54:19 <advil> alexjurk1 I think it's benign
21:54:39 <advil> expected behavior if 0.25 is not updated to stable before t start
22:01:14 <advil> what I didn't understand is that -b1 versions get translated to -a versions in the sequell 'cv' field
22:05:31 <gammafunk> !kw t
22:05:32 <Sequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:05:32 <NotSequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:05:45 <gammafunk> guess we might need to update that for cwz?
22:05:51 <gammafunk> !lg * t cwz s=file
22:05:52 <Sequell> 5740 games for * (t cwz): 3771x cwz/soup/trunk/logfile, 1969x cwz/soup/0.25/logfile
22:05:52 <NotSequell> 1969 games for * (t cwz): 1969x cwz/soup/0.25/logfile
22:06:16 <gammafunk> I suppose it's because cwz hasn't updated their trunk in a while
22:07:17 <gammafunk> I think cwz usually does a trunk update so they'll get the next version alpha, but hong added 0.25 fairly early this time
22:07:46 <gammafunk> and probably isn't aware that it causes sequell issues if he doesn't update trunk
22:08:20 <gammafunk> I imagine if we condition t on file that will slow down queries a good deal
22:08:43 <gammafunk> maybe vlong?
22:08:55 <gammafunk> !lg * cwz t file~~trunk s=vlong
22:08:56 <NotSequell> No games for * (cwz t file~~trunk).
22:08:56 <Sequell> 3772 games for * (cwz t file~~trunk): 3772x 0.25-b1-12-g2958ca1
22:09:05 <gammafunk> !lg * cwz t file!~~trunk s=vlong
22:09:06 <Sequell> 1969 games for * (cwz t file!~~trunk): 1300x 0.25-b1-21-g047f711, 669x 0.25.0-9-g26727b6
22:09:06 <NotSequell> 1969 games for * (cwz t file!~~trunk): 1300x 0.25-b1-21-g047f711, 669x 0.25.0-9-g26727b6
22:09:15 <gammafunk> nice
22:09:26 <alexjurk1> advil: oh cool
22:09:41 <alexjurk1> yep, I was asleep when 0.25.0 was tagged so it updated a little late
22:10:26 <gammafunk> !lg * cwz start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr= vlong>=0.25-b1-21-g047f711
22:10:27 <Sequell> Unknown field: time
22:10:27 <NotSequell> Unknown field: time
22:10:54 <gammafunk> !lg * cwz start>='2020-06-12 20:00:00' ((cv=0.25|0.25-a)) explbr= vlong>=0.25-b1-21-g047f711
22:10:55 <NotSequell> 1969. Only3Rune the Cleaver (L14 DDFi of Makhleb), slain by a hell hound (led by Azrael) on D:13 on 2020-06-22 02:04:48, with 51345 points after 15259 turns and 0:16:34.
22:10:56 <Sequell> 1969. Only3Rune the Cleaver (L14 DDFi of Makhleb), slain by a hell hound (led by Azrael) on D:13 on 2020-06-22 02:04:48, with 51345 points after 15259 turns and 0:16:34.
22:11:04 <gammafunk> !lg * cwz start>='2020-06-12 20:00:00' ((cv=0.25|0.25-a)) explbr= vlong>=0.25-b1-21-g047f711 s=file
22:11:05 <NotSequell> 1969 games for * (cwz start>='2020-06-12 20:00:00' cv=0.25|0.25-a explbr= vlong>=0.25-b1-21-g047f711): 1969x cwz/soup/0.25/logfile
22:11:05 <Sequell> 1969 games for * (cwz start>='2020-06-12 20:00:00' cv=0.25|0.25-a explbr= vlong>=0.25-b1-21-g047f711): 1969x cwz/soup/0.25/logfile
22:11:54 <gammafunk> !kw t start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr= vlong>=0.25-b1-21-g047f711
22:11:55 <Sequell> 't' is built-in: t => when=t
22:11:55 <NotSequell> 't' is built-in: t => when=t
22:11:59 <gammafunk> !kw t
22:12:00 <Sequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:12:00 <NotSequell> Built-in: t => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:12:10 <gammafunk> ah, can't even override; I swear you used to be able to
22:12:22 <gammafunk> !kw t0.25
22:12:23 <Sequell> Built-in: t0.25 => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:12:23 <NotSequell> Built-in: t0.25 => start>='2020-06-12 20:00:00' time<'2020-06-28 20:00:00' ((cv=0.25|0.25-a)) explbr=
22:15:49 <floraline> advil: thanks, i think kelbell was actually not ever reporting stable version crashes. should be fixed now maybe
22:17:58 <floraline> advil: regarding rebuild script, was the issue that it appeared not to respond after you actually triggered the rebuild? iirc this is a fcgi issue that won't be resolve for me until i upgrade to buster
22:18:57 <floraline> if i'm thinking of the right issue, the problem is that it is buffering the response meaning you don't get to watch it go
22:23:42 <floraline> i have joins and parts ignored in this channel, does anyone know if i can configure irssi to always show kelbell's activity?
22:25:55 <floraline> is there some way to trigger a crash on purpose to know if it worked?
22:28:55 <amalloy> in wizmode, &^C crashes
22:29:11 <amalloy> there aren't supposed to be ways to force it out of wizmode, because those are exploitable
22:34:12 <elliptic> !kw tfixed
22:34:13 <Sequell> Keyword: tfixed => t0.25fixed
22:34:13 <NotSequell> Keyword: tfixed => t0.25fixed
22:34:18 <elliptic> !kw t0.25fixed
22:34:19 <Sequell> Keyword: t0.25fixed => t0.25 file!="cwz/soup/trunk/logfile"
22:34:19 <NotSequell> Keyword: t0.25fixed => t0.25 file!="cwz/soup/trunk/logfile"
22:34:52 <elliptic> gammafunk: I made ^ earlier, I don't know whethere there are speed issues but it seemed fine for the queries I tried
22:41:57 <gammafunk> elliptic: cool, thanks
22:46:34 <Pinkbeast> floraline: you can do "/ignore except Kelbell JOINS" AFAIK
23:05:17 <Kelbell> floraline (L8 MiFi) ERROR in 'wizard.cc' at line 78: Intentional crash (D:6)
23:06:42 <floraline> amalloy: thank you
23:06:52 <floraline> Pinkbeast: thanks
23:12:50 -!- jfcaron_ is now known as jfcaron
23:23:33 <kate-> hm, the nameless horrors from hell effect miscasts seem to be consistently very weak (like, 10hp and 5 damage), is the scaling for miscast strength off there maybe?
23:24:42 <kate-> i guess they got nerfed because they were too nasty for actual regular miscasts at one point, so maybe they need scaling up a bit on the hell end now
23:25:59 <PleasingFungus> when gammafunk was talking about this in discord a few days ago, he suggested that nameless horrors really shouldn't be hell spawns at all
23:26:03 <PleasingFungus> since that's not really what they're designed for...
23:29:55 <kate-> they seem sort of fine compared to the other miscasts that are just plain damage, but maybe still not ideal yeah
23:46:54 <alexjurk1> i wonder if nameless horrors and abominations should be merged...
23:49:25 <PleasingFungus> someone was talking about giving nameless horrors shadow creatures and/or abjuration
23:51:47 <kate-> they already have abjuration don't they?
23:52:04 <PleasingFungus> oh
23:52:09 <PleasingFungus> @??nameless horror
23:52:09 <Gretell> nameless horror (10X) | Spd: 10 | HD: 15 | HP: 51-73 | AC/EV: 8/2 | Dam: 3012(antimagic) | 11non-living, see invisible, unbreathing | Res: 13magic(immune), 09poison+++, 12drown, 04rot++, 13neg+++, 13torm | XP: 980 | Sp: abjuration [11!AM, 06!sil] | Sz: Giant | Int: brainless.
23:52:12 <PleasingFungus> they do
23:52:21 <PleasingFungus> anyway, it feels like it might be odd for abominations to have abjuration