IRC logs for #openttd on OFTC at 2019-07-14
⏴ go to previous day
00:02:05 *** Smedles_ has joined #openttd
00:10:24 *** andythenorth has left #openttd
00:13:15 <Eddi|zuHause> is that just me or have there been half a dozen factorio clones recently?
00:17:28 *** Wormnest has joined #openttd
01:07:37 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas opened pull request #7654: Change: Update for French town name "Beaujolais" https://git.io/fjXcY
01:13:55 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7654: Change: Update for French town name "Beaujolais" https://git.io/fjXcZ
01:14:16 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7654: Change: Update for French town name "Beaujolais" https://git.io/fjXcY
01:48:46 <Samu> after 30 rounds, the best 2 AIs are neck and neck in score
01:49:36 <Samu> and after, there's 8 AIs which are really close to each other as well
01:50:23 <Samu> then there's another gap
01:51:21 <Samu> 17 more rounds to score...
01:52:30 <Samu> there are now 18 AIs below WrightAI
01:52:58 <Samu> I consider WrightAI the line between good AI and bad AI
01:53:21 <Samu> for no real reason, though
01:53:29 <Samu> maybe because it was made by devs
02:22:12 *** Thedarkb-T60 has joined #openttd
02:29:29 <Hazzard> How far out will sub tropical lumber mills search for wood :s
02:37:10 <Hazzard> 20 tiles in each direction
02:41:47 <Beerbelott> glx: Already knew the doc you pointed out... but I'm somehow struggling at resconstructing the binary chain to send for RCON client (authenticated) commands...
02:41:58 <Beerbelott> Trying to read the code of some lib. Not an easy task.
02:45:40 *** Supercheese has joined #openttd
04:23:48 <Eddi|zuHause> hm... so i have backups of all the partitions that used to be on the broken SSD (made before it broke), but not of the partition table... and i can't seem to make a partition table so it actually boots from this USB device
04:25:00 *** Beerbelott1 has joined #openttd
04:32:39 <Beerbelott1> It seems www.openttd.org is inaccessible
04:33:06 <Beerbelott1> openttd.org is, redirecting to www.openttd.org w/ HTTP 301
04:58:49 <Beerbelott1> Tested from my server too, which sits in a remote location
04:59:03 <Beerbelott1> * Connected to www.openttd.org (2a03:b0c0:2:d0::dc6:7001) port 80 (#0)
04:59:03 <Beerbelott1> > Host: www.openttd.org
04:59:03 <Beerbelott1> > User-Agent: curl/7.52.1
04:59:05 <Beerbelott1> * Recv failure: Connection reset by peer
04:59:30 <Eddi|zuHause> does that do https?
04:59:46 <Eddi|zuHause> maybe it's an IPv6 issue?
05:00:20 <Eddi|zuHause> there maybe was a thing where a thing needed to be updated manually every now and then
05:00:28 <Beerbelott1> lemme test for that
05:01:08 <Beerbelott1> IPv6 pb on your box
05:01:13 <Beerbelott1> Lemme guess... Hetzner?
05:01:46 <Eddi|zuHause> anyway, there is probably nobody around who can fix it right now
05:01:57 <Beerbelott1> OK. Often when that happens there is a correlation w/ that hoster ^^
05:02:04 <Beerbelott1> Well, bad call there, but it's IPv6 indeed
05:02:17 <Beerbelott1> Well, it's not "critical" I suppose
05:02:45 <Beerbelott1> It's a lil bit weird dual-stacks don't test both version when available... DNS answers both addresses alright
05:03:13 <Beerbelott1> OK off to bed now
05:03:19 *** Beerbelott1 has left #openttd
05:28:00 <glx> yup ipv6 fails here too, probably the proxy
05:58:08 <reldred> right, I've finally got what I was wishing for, NRT in jgr's patch pack, I've got my custom made tree set all ready to go for that sweet sweet diferentiating folliage, I've got a few custom tweaked building sets all playing nicely together,
05:58:21 <reldred> and do you think I can maintain attention span long enough to actually PLAY it, no.
06:01:05 <Eddi|zuHause> depends how big of a map you play?
06:01:42 <reldred> been hovering around 2048x1024 or 2048x512
06:02:31 <reldred> nah my add has been legit playing silly buggers lately. I can get really engrosed in some stuff and the others I can't get 5mins in
06:03:33 <Eddi|zuHause> my experience is if i don't get interrupted by work problems or game breaking bugs, i can get about a month out of a game
06:04:32 <reldred> About the best I can get is around 5days
06:04:57 <reldred> And at the present moment it's about 5 mins
06:05:54 <Eddi|zuHause> currently i'm bugged with system instabilities, where about once per day the system freezes up
06:06:48 <reldred> Yeah I was expecting my self-compile of jgr's nrt branch to be buggy as hell but I've not managed to break it yet
06:08:03 <reldred> Can't get that updated rainfall patch to compile, seems to be drawing on a bunch of includes common in linux systems but not windows
06:09:29 <Eddi|zuHause> can you compile the original?
06:09:38 <reldred> Yup, and jgr's patchpack
06:09:50 <Eddi|zuHause> i mean the unupdated patch
06:10:07 <reldred> it was apparently a long way behind master
06:10:27 <Eddi|zuHause> yes, you'd have to check out an old version
06:10:40 <reldred> Yeah I didn't go that far into it
06:11:02 <reldred> the old version was pre NRT so I wasn't terribly interested.
06:11:53 <reldred> Now that does remind me, now that I do have a working build environment there was some other things I was going to start poking around with
07:19:30 <Hazzard> some of this code is wack :D
07:22:36 <Hazzard> Does anyone know if the percentages (of 256) are rounded or truncated?
08:30:44 <reldred> oh custom bridgeheads, how I've missed you
08:32:07 <reldred> I finally compiled his nrt branch
08:32:13 *** sla_ro|master has joined #openttd
08:32:16 <peter1138> I keep meaning to try porting that to master but... map arrays and savegame changes :/
08:32:50 <reldred> jgr's branch also has some magic buttons that handle all that scheduling nonsense that I don't understand
08:32:57 <peter1138> Also I thought my milk was curdling but turns out the fridge is so cold it's just partially frozen o_O
08:33:23 <peter1138> I suppose at least it'll keep fresh for longer :D
08:33:42 <peter1138> It's raw unpasturised milk.
08:34:24 <reldred> and improved breakdowns! *swoon*
08:34:52 <reldred> more failure states, they don't just break and sit there they can limp along, etc.
08:36:39 <peter1138> Definitely not off, it still tastes really good (despite being a week old!)
08:37:39 <reldred> nekminnit brain worms
08:39:15 *** Progman has joined #openttd
08:40:36 <reldred> huh, I never noticed the extended zoom out options
08:44:09 <peter1138> Oh dear, I just looked at today's planned ride. To a Café Nero. I really can't be arsed with that.
08:45:04 <peter1138> And the leader is kinda annoying, slows down or speeds up randomly.
08:45:27 <reldred> I should buy a new push bike
08:45:44 <reldred> since I won't be running any time soon with the arthritis
08:46:00 <reldred> yeah doc reckons I have RA
08:46:01 <peter1138> Yeah, cycling is at least low impact.
08:46:16 <peter1138> Anyway, I should have a day off from cycling, I've ridden all week.
08:46:40 <reldred> The problem is my idiot ass would go and buy a hard tail with small wheels and do dumb things on concrete with it.
08:46:52 <peter1138> Hard tail with small wheels...
08:47:11 <reldred> BMX's are too small for me, I smash my knees something chronic
08:47:30 <reldred> But yeah drop the wheels a few inches, nothing too wild and they do well on concrete
08:47:50 <peter1138> Other than folding bikes, I don't know what other bikes have small wheels.
08:48:37 <reldred> I _could_ do with a decent folder for sticking in my truck though
08:48:38 <peter1138> So, er, there's 26, 27.5 or 29 generally.
08:48:41 <reldred> or get that bmx replaced
08:49:32 <reldred> bought that bmx off a very crazy lady in coastal nsw when I was driving to adelaide from queensland
08:49:39 <reldred> just a lazy 3000km drive
08:49:48 <peter1138> I just want more bikes :/
08:50:14 <peter1138> Like I keep thinking a Brompton would be useful... but kinda wasteful.
08:56:49 <Hazzard> Reading the industry source code, is it only possible for industries with multiple types (farms, oil rigs) to close if both their productions reach zero?
08:57:14 <peter1138> If you've just read the source you probably know more than anyone else ;)
08:58:07 <nielsm> yeah by default rules a multi-output primary industry can only close when all of its outputs are at minimum
08:58:32 <nielsm> also note that that's for smooth economy, not classic economy
08:59:00 <nielsm> and newgrf industries can override the closure rules with the production change callback
08:59:37 <Hazzard> Hmm neat, thanks nielsm
09:01:47 <Hazzard> source code is a bit hard to read, not confident
09:02:19 <reldred> Ahah. The only foul up I've managed to find in jgr's nrt, custombridgeheads plus nrt plus stations underneath are all fine, until a train is in said station and a tram goes over the top and then the poor sprite sorter loses its goddamn mind.
09:02:21 <Hazzard> I guess the clamp means the production never actually shows up as zero
09:02:30 <reldred> So situation normal then.
09:02:31 <nielsm> it's features layered on features with years of special exceptions added
09:03:08 <peter1138> Bridges over stations is a feature I didn't add (after writing it) about 10 years ago, for that very reason.
09:03:52 <reldred> It's fine, and even fine with a train at the station, it's just once a tram goes accross the pylons and everything has a fit.
09:04:10 <reldred> Overall not too bothered, if anything I've been surprised how stable this branch has been.
09:05:12 <reldred> I was thinking the reason he didn't post binaries was because it was unstable but so-far so-good.
09:07:23 <Eddi|zuHause> reldred: sprite sorter is a bit useless if bounding boxes overlap
09:08:01 <reldred> I have fond memories of coaxing behavior out of the sprite sorter when making stations with overpasses, etc.
09:08:06 <Eddi|zuHause> might need a better heuristic to resolve that case, but solving it one way usually creates more problems elsewhere
09:08:13 <reldred> and by fond I mean painful
09:09:19 <reldred> it's that one thing alone that's probably singlehandedly stopped the suburban stations rework dead in its tracks
09:09:31 <reldred> laid out all the sprites then went fuck now I have to actually code it
09:12:23 <peter1138> Urgh, as much as I like my CH yoke, the 8-bit sampling on the analogue axes is kinda frustrating :/
09:12:48 <Eddi|zuHause> maybe the game should throw error messages for bounding box overlaps, if in newgrf-developer-mode?
09:13:41 <reldred> peter1138: it's only 8bit? :O. I was pondering getting a yoke for star citizen since I mainly fly big ships when I do play it.
09:14:04 <peter1138> Yes, the CH hardware is pretty ancient.
09:14:27 <peter1138> Thinking about swapping the circuitry for a Leo Bodnar thing.
09:15:58 <peter1138> Technically it's enough.
09:16:10 <peter1138> But it feels wrong when you can see the steps in movement.
09:20:33 <peter1138> Hmm, seems there is a bit of a deadzone too :/
09:20:42 <peter1138> Wonder when that started. Hmm.
09:21:23 <peter1138> I tried using yoke & rudder in Elite Dangerous but... Elite's yaw pitch is so crap anyway...
09:23:34 <peter1138> Hmm, problem with OrbX TrueEarth UK is... the UK is mostly flat :p
09:23:59 <peter1138> There's a bit up north where mountains are visible but when you're at 4-5000 ft it all looks very tame.
09:47:54 <peter1138> I think the G27 pedals are probably 8 bit too, but that hardly matters either.
09:48:25 <peter1138> I did buy some hall-effect sensors to replace the pots with but never go around to doing it. And then lost them.
10:06:06 *** andythenorth has joined #openttd
10:24:32 <peter1138> Uh, what, to change my callsign I have to... edit the plane? o_O
10:28:15 <andythenorth> in the cheat window, for game date, why does the '<' arrow incremement game date?
10:28:31 <andythenorth> I thought I was mis-clicking, but I'm not
10:28:57 <andythenorth> in fact all the '<' arrows are broken
10:30:06 * andythenorth wonders if James103 will file an issue about it :)
10:33:01 <andythenorth> hmm, not all, maybe just date and max height
10:41:02 *** Flygon_ has joined #openttd
10:45:25 <reldred> we need more tree sets
10:45:41 <reldred> I should do something about that
10:47:08 <andythenorth> could be UI scaling issue
10:47:23 <peter1138> andythenorth, difficult to test, eh?
10:47:44 <andythenorth> I can only repro it in the cheat window
10:47:52 <andythenorth> can't trigger it in settings
10:48:01 <peter1138> Reproduce it with different UI scales.
10:48:21 <peter1138> They're separate windows, so... that's likely.
10:49:21 <andythenorth> yeah it triggers at [1x zoom, 2x font] and at [1x zoom, 2x font] and [2x zoom, 2x font]
10:49:42 <andythenorth> how do I record an animated gif? :P
10:51:10 <peter1138> No need to record, surely?
10:51:17 <peter1138> Or are you unable to describe things clearly these days?
10:53:06 <andythenorth> the video is just way less verbose :P
10:53:58 <peter1138> Less verbose is less useful.
10:58:32 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth opened issue #7655: Cheat window date '<' arrow incremements date https://git.io/fjX84
11:00:29 <andythenorth> I added an extra note, moving the mouse further left does get the expected result
11:05:14 <nielsm> the docs on ClickChangeDateCheat does not match the actual code...
11:05:45 <nielsm> "p1 unused" (appears to contain modified year), "p2 +1 (increase) or -1 (decrease)" (is not used at all)
11:05:54 *** gelignite has joined #openttd
11:11:21 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #7655: Cheat window date '<' arrow incremements date https://git.io/fjX8w
11:13:38 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #7655: Cheat window date '<' arrow incremements date https://git.io/fjX8o
12:01:03 *** HerzogDeXtEr has joined #openttd
13:06:25 *** andythenorth has joined #openttd
13:47:01 <nielsm> reaching dangerous levels of completedness
14:06:48 <DorpsGek_II> [OpenTTD/OpenTTD] michicc approved pull request #7560: Make GSGoal.QuestionClient work correctly at least for clients with ID < 2**16 https://git.io/fjXBD
14:20:24 <andythenorth> I vastly prefer the first option
14:28:40 <LordAro> michi_cc: definitely first
14:28:58 <LordAro> possibly with a little more spacing after the paragraph
14:31:43 <michi_cc> The formatting is just a markdown change, but the spacing would be css messing I think.
14:32:11 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7560: Make GSGoal.QuestionClient work correctly at least for clients with ID < 2**16 https://git.io/fjZjH
14:32:12 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro closed issue #7534: GSGoal.QuestionClient sends message to wrong client https://git.io/fj3Uq
14:32:30 <LordAro> michi_cc: we can worry about that later then
14:33:09 <michi_cc> Please note that the GitHub preview renders differently to our site.
14:37:30 <michi_cc> You may push changes anytime.
14:39:27 <Eddi|zuHause> so, anyone know how one would select the boot partition with gpt?
14:40:10 <Eddi|zuHause> or, where i would get a windows bootloader from?
14:42:15 <LordAro> michi_cc: i feel the "extension" explanation is unnecessary, given we're assuming people are already familiar with source code
14:43:08 <michi_cc> Our use of both .h and .hpp is not common, if that is in, leaving .cpp out looks wiered.
14:44:06 <LordAro> though our differences between hpp & h are usually fairly arbitrary :p
14:44:13 <michi_cc> I.e. both extension are commonly used, but don't have any set meaning.
14:45:45 <michi_cc> And purely from a didactical point of view, IMHO if you split something into parts, simply not saying anything about one part is wrong.
14:48:04 <Eddi|zuHause> nielsm: not entirely sure, but i assume UEFI
14:48:28 <Eddi|zuHause> nielsm: i have a windows partition and a boot partition, but the partition table stuff got lost
14:48:47 <Eddi|zuHause> nielsm: and when i create a partition table and move the partitions in, it won't boot from the disk
14:49:05 <LordAro> probably merge it now so we can see it on staging, then tweak anything else
14:50:44 <nielsm> Eddi|zuHause: if you have the windows boot partition (typically a 500 mb FAT32 thing) then you should just be able to point your firmware at the disk and it should auto-detect the boot partition and run the bootloader
14:51:25 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7654: Change: Update for French town name "Beaujolais" https://git.io/fjXRk
14:51:35 <Eddi|zuHause> nielsm: did i put them in the wrong order?
14:52:21 <nielsm> order shouldn't matter afaik
14:52:44 <Eddi|zuHause> i'm still thinking i'm missing some "this is the boot partition" flag
14:52:59 <Eddi|zuHause> but i've never worked with gpt partitions before
14:53:10 <nielsm> have you tried booting from a windows install usb/dvd and see if the repair boot option can detect the install?
14:53:30 <nielsm> and bbiab, need to take care of laundry
14:53:41 <Eddi|zuHause> i've not, but that should probably be my next try
14:53:53 <michi_cc> LordAro: Should look okay, I even installed jekyll to check. Will still merge though.
14:54:52 <Eddi|zuHause> "Win 10 1703" iso.. that should work, if i just dd it on a stick, right
14:56:31 <DorpsGek_II> [OpenTTD/OpenTTD] Eddi-z commented on pull request #7654: Change: Update for French town name "Beaujolais" https://git.io/fjXRY
15:21:31 <nielsm> what will happen if that number reaches 100% ?
15:33:19 <Eddi|zuHause> nielsm: that will never happen
15:38:32 <andythenorth> well, strictly, it's just counting sprites marked as complete
15:38:46 <andythenorth> it doesn't deal with design issues, codebase issues, bugs
16:33:41 <Samu> scoring the last 6 rounds atm
16:36:18 <Eddi|zuHause> weren't we there like an hour ago? :p
16:36:34 <andythenorth> each 1% is a few trains to draw :P
16:37:03 <Eddi|zuHause> maybe you should switch metrics? :p
16:37:33 <Eddi|zuHause> 90% done, time for the other 90%
16:38:38 <andythenorth> soon it will get worse :P
16:38:59 <andythenorth> 500 trains seems like a lot for one roster eh :P
16:39:22 <Eddi|zuHause> if we add views, your whole project will collapse under its own weight :p
16:39:44 <andythenorth> we probably won't though
16:39:54 <andythenorth> at some point compile time inhibits additions
16:41:01 <andythenorth> "Iron Horse is a minimalist train set"
16:41:55 <Eddi|zuHause> "perfection is not achieved when there is nothing more to add, but when there is nothing more to remove"
16:42:53 <andythenorth> it's a very subtle quote
16:43:18 <Eddi|zuHause> i probably mangled it a bit
16:44:02 <andythenorth> if I could find a way, I'd remove one whole generation
16:44:13 <andythenorth> but I tried it and it doesn't quite make sense
16:44:26 <Eddi|zuHause> i don't think that's your problem
16:45:53 <andythenorth> unusually I don't have many problems with Horse
16:46:07 <andythenorth> I just like deleting things :P
16:46:46 <andythenorth> I could remove some coach / freight wagons
16:47:01 <andythenorth> 'mistakes were made'
16:47:30 <andythenorth> I misunderstood cargo aging, so the wagon design is a little flawed
16:53:02 <andythenorth> strictly speaking, there are 6 luxury passenger coaches that could be deleted
16:53:10 <andythenorth> each with 4 graphical variants
16:53:36 <andythenorth> it's also not clear that every generation needs 4 mail wagons :P
16:53:56 <andythenorth> hmm 6 mail wagons actually
16:54:23 <Eddi|zuHause> how many length paddings do you need?
16:56:06 <Eddi|zuHause> i think in CETS i have mail vans for passenger trains and mail vans for freight trains (what in other countries would be a caboose)
17:01:26 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7353: Feature: Measure vehicle capacity utilisation efficiency https://git.io/fhho4
17:01:45 <nielsm> still need to find a way to measure days (or ticks) spent waiting at a signal...
17:02:08 <nielsm> that works reliably and equally for block and path signals
17:02:45 <LordAro> something horrendously inefficient, like a counter per signal
17:02:52 <Eddi|zuHause> i wanted to add a few flags for which speed limit affects a train, you could incorporate that there (signal would impose a speed limit of 0)
17:03:47 <Eddi|zuHause> LordAro: the signal doesn't know whether a train is waiting on it
17:04:28 <nielsm> I still haven't understood the actual code path taken to pathfind a train past a signal
17:04:31 <LordAro> but what if it did? :p
17:04:44 <nielsm> it seems that for one of the types it needs to run through the code twice for some reason?
17:05:21 <nielsm> and the VRF_TRAIN_STUCK flag is only set for one of the types?
17:05:52 <Eddi|zuHause> i have some vague memory of dealing with that issue years ago with my "stuck trains" hack
17:06:06 <Eddi|zuHause> maybe i should dig that up
17:07:01 <Eddi|zuHause> did i ever post that in the forum?
17:08:55 <nielsm> specifically Train::wait_counter increases at double speed for one signal type
17:09:47 <nielsm> also that commit subtly changes something else that it shouldn't, see if you can spot it
17:17:29 *** sla_ro|master has joined #openttd
17:22:34 *** gelignite has joined #openttd
18:00:28 *** andythenorth has joined #openttd
18:05:14 <nielsm> after waiting at different types of signals for several months there is a slight difference between penalty accrued
18:05:18 <nielsm> but not a major difference
18:05:48 <nielsm> (the dark blue number below the "efficiency this year" line is the raw "potential cargotiles delivered" value)
18:07:15 <andythenorth> very interesting
18:07:25 <andythenorth> feel like adding lifetime earnings? :P
18:07:38 <nielsm> those could be interesting yes :)
18:08:13 <nielsm> even just a display of "earnings this year" in addition to the "profit this year" could be useful
18:09:51 <michi_cc> nielsm: short-circuited if statements with side effects, how nice :p
18:10:17 <nielsm> well after some more analysis the side effect can safely be moved to the line above
18:10:34 <nielsm> nothing depends on the value apart from that code itself
18:11:01 <nielsm> iow. it's okay for the wait_counter to increment even for "wrong" settings of reverse_at_signals
18:14:49 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh updated pull request #7353: Feature: Measure vehicle capacity utilisation efficiency https://git.io/fhho4
18:15:58 <Samu> 1st and 2nd places are not yet decided :p so exciting
18:22:16 <michi_cc> nielsm: Make the vehicle text into two lines and combine the group stats to "Efficiency this (last) year: 76% (73%)" maybe?
18:26:40 <nielsm> how about the vehicle list itself?
18:36:03 <Samu> last round remaining, the round that was generated via random.org
18:36:13 <Samu> should pose some surprises
18:45:12 <andythenorth> 18 trains left to draw
18:53:39 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7353: Feature: Measure vehicle capacity utilisation efficiency https://git.io/fjXue
18:54:41 <Samu> and the tournament winner is...
18:56:03 <Samu> AIAI with 2407 points, NoNoCAB 2nd with 2377 points
18:58:19 <nielsm> Samu: suggestions for next time: decide on a time limit per game ahead of time, so your running time isn't unbounded (maybe 50 or 100 years), decide on some fixed scenarios ahead of time and test each AI pairing in several different scenarios (doesn't have to be "scenario scenarios", just config and map) so the results are properly comparable
18:58:54 <nielsm> and decide on scoring rules before you begin the testing and figure out if you can make your GS collect some more interesting statistics during each game
19:00:10 <nielsm> maybe plan your scenarios so different qualities of AIs are tested, e.g. how well they perform on small maps, large maps, maps with lots of prebuilt roads, maps with lots of water, mountaneous maps, etc.
19:01:17 <Samu> i tried to diversify the worlds as most as I could
19:01:46 <nielsm> yes but if you test combination A vs B on a different scenario than combination A vs C, they the results of the two tests aren't really comparable
19:02:33 <nielsm> maybe A has an advantage in the first and a disadvantage in the second due to scenario differences, and you couldn't tell if A wins first and loses second because it's better or worse than B or C, or because of the different scenarios
19:03:15 <Samu> I know :( but then i'd be repeating the same config
19:03:39 <nielsm> it makes your comparison actually usable
19:03:56 <Samu> there isn't a single config that suits all ais properly
19:04:19 <nielsm> and that's something your testing can uncover
19:04:36 <nielsm> what configs and map types are well suited and poorly suited for each type
19:04:55 <Wormnest> Nice to see that NoNoCab did wel Samu :)
19:05:18 <Samu> Wormnest, WormAI on the other hand...
19:05:58 <Samu> 14th place with 1480 points
19:06:38 <Samu> the scoring rules wasn't something I planned ahead of time
19:07:07 <Samu> i had an idea how I would score them before I started, but then I realise my initial ideas wouldn't be too fair
19:07:09 <Wormnest> Oh well, I know it has its weaknesses and it doesn´t do all vehicle types yet
19:07:17 <Samu> that was in the middle of my 10th or so round
19:13:22 <Samu> my AI finished 3rd place which makes me wonder how biased were my configs...
19:14:31 <Samu> who wants to download 730 MB of savegames?
19:15:45 <Wormnest> Any config will be biased
19:18:17 <Samu> nielsm, I'm quite surprised about EmotionAI and SimpleAI scores
19:18:32 <nielsm> Wormnest yeah, that's why my suggestion is to run every AI pairing on multiple configs and maps
19:18:37 <Samu> regarding the different combinations you mentioned
19:18:51 <Samu> they ended right where they should end
19:19:01 <Samu> they're basically the same AI
19:19:43 <Samu> I was kind of worried they'd be far apart from each other in the end, turns out they didn't
19:21:58 <Samu> as I was scoring this, the battle for first and 2nd places were always between AIAI and NoNoCAB
19:22:25 <Samu> it would be either of the two who would win
19:22:38 <Wormnest> nielsm Yes, but still there´s so many settings that you can´t test all combinations
19:23:54 <Samu> Terron initially was seen as not a real contender, but it managed to reach 3rd place near the end
19:24:13 <Samu> my AI took that place in the last round, lucky
19:24:30 <Samu> my AI was mostly on 4th and 5th places
19:24:56 <Wormnest> In most of my tests Terron did fairly well so I´m not surprised
19:25:14 <Samu> it didn't start that well
19:25:29 <Samu> i mean 7th 8th and such, not really that high
19:25:44 <Samu> for me it was the big surprise
19:26:41 <Samu> trAIns surprised me a big negatively
19:26:47 <Samu> I honestly expected more
19:27:28 <Samu> turns out the great times it was pulling out, were after a bankrupt
19:27:39 <Samu> had I not changed the scoring rules
19:27:45 <Samu> it would probably rank higher
19:34:12 <DorpsGek> Samu: 0.731610942249
19:35:57 <Samu> do I post the savegames, I mean is there any interest?
19:36:10 <Samu> 730 MB to the forum seems massive
19:36:26 <peter1138> I don't think anyone cares, no.
19:37:20 <nielsm> you can put them in a google drive or onedrive shared folder and link that
19:38:56 <peter1138> Ever the pragmatist :D
19:47:01 <Samu> how do i share folders with onedrive
19:48:32 <DorpsGek_II> [OpenTTD/OpenTTD] Ben79487 commented on issue #7655: Cheat window date '<' arrow incremements date https://git.io/fjXzI
19:49:02 <nielsm> I can open it and list the files
19:49:17 <nielsm> and download a random file from it too
19:51:56 <Samu> i should post the OneNote file too
19:52:04 <Samu> has some additional detains
19:52:13 <Samu> like who vs who in each round
19:54:18 <Samu> hmm how do i share one note stuff?
19:55:20 <Samu> can only share via email? that sucks
19:57:50 *** Progman has joined #openttd
20:02:53 <Samu> if neither works, i dunno what to do
20:12:39 <Wormnest> All of those OneNote links work for me
20:32:47 *** Smedles has joined #openttd
21:03:34 *** Thedarkb-T60 has joined #openttd
21:06:42 <Samu> one final thing I wanna do, count all victories/losses, regardless of score
21:11:54 <nielsm> (station window ideas)
21:17:19 <nielsm> that only works when I already have a very good idea of what I want, I put too much work into details then
21:17:35 <nielsm> hand drawn lets me focus more on form
21:19:45 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjXzH
22:09:15 *** Thedarkb-T60 has joined #openttd
22:26:11 <Samu> AIAI v97 2407 66 70 68 46 68 66 66 58 42 64 66 56 18 1 66 64 2 15 52 54 70 36 50 66 2 16 46 70 62 15 66 62 58 62 2 64 62 70 12 66 66 68 60 70 48 68 62
22:31:17 <Samu> how do I create cells on the forum?
22:34:07 <DorpsGek_II> [OpenTTD/OpenTTD] Gymnasiast commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjXgn
22:34:29 <andythenorth> nielsm: picture tells 1000 words? :)
23:38:12 *** Supercheese has joined #openttd
continue to next day ⏵
