[00:04] MSG: Read error: Operation timed out [00:18] Join: michal joined #corewars [02:49] Part: michal left #corewars [05:30] Part: Mike left #corewars [07:43] Join: Mike joined #corewars [08:33] Join: willvarfa joined #corewars [09:05] Hi Will [09:05] It's okay, I'm not a chatbot :) [09:07] hello Mike [09:10] Nothing has happened today [09:11] in corewars, in this channel, or in the world generally? [09:13] in corewars [09:13] on-speed is in 50th place :( [09:17] someone rescue it! :( [09:19] put it out of it's misery, prehaps? ;kill? [09:19] sorry :-) [09:19] I'd be proud of #50 myself [09:19] got to go! [09:19] * willvarfa *waves* [09:19] MSG: Quit: Chatzilla 0.9.68.5 [Firefox 1.0.4/20050512] [09:30] I'd be proud of #50 too [09:59] Join: Mizcu joined #corewars [10:08] Hi Mizcu [11:44] Join: Core29 joined #corewars [11:50] Hi [12:07] MSG: Read error: Connection reset by peer [12:57] Join: Jens joined #corewars [12:57] :) [13:18] hI jENS [13:18] Hi Jens! [13:18] (caps malfunction) [14:16] Join: sascha joined #corewars [14:33] Part: sascha left #corewars [14:33] Join: Fizmo joined #corewars [14:34] hi [14:34] Hi Fizmo! [14:34] * Jens is back in 15 mins :( [14:40] hi Jens [14:41] be back on early evening [14:41] MSG: [15:28] Join: willvarfa joined #corewars [15:33] Join: sf joined #corewars [15:33] Hi [15:33] hello [15:38] Hi sf, willvarfa! [15:38] hello Jens [15:38] Hi Jens, Will [15:39] hello everyone! [15:43] Only 4 successful challenges left and HFB is my oldest warrior on 94nop, that I ever had :) [15:44] Let me see if I can help :) [15:44] Yes, that would be nice. [15:45] By the way the last entry for 94nop only has 62 points [15:45] I think I can beat that :) [15:46] Is it a new version of Quintessence? [15:49] Yes [15:49] :) [15:56] brx: you owe me a favour now! [15:56] sf huh [15:57] * brx_ checks nano [15:57] lol [15:57] it beats type-1 ;= [15:58] * brx_ hugs sf ;) [15:58] that was really sweet [15:58] Not as much as John's warriors do [15:58] * sf blushes [16:06] func l = map (\y -> y+2) (filter (\z -> z `elem` [1..10]) (5:l)) [16:06] becomes [16:07] func = map (+2) . filter (`elem` [1..10]) . (5:) [16:23] MSG: Quit: I don't like you. But Bersirc 2.2 does. Try it out now. [ http://www.bersirc.org/ - Open Source IRC ] [16:26] Wow ... now I have a several score surfaces. When I make an exact surface (1522) and a surface with a reduced number of rounds (100) the maximal score-difference is 66! [16:29] how goes the search for surfaces? [16:30] Every day a suprise! [16:30] I expected the difference between an exact score surface and surface, that was calculated with a reduced number of rounds, to be more or less random, but it isn't [16:32] looked at fcrack? [16:32] Thanks for reminding me [16:32] Jens: pseudo random number generator? [16:33] Join: sascha joined #corewars [16:33] Hi sascha! [16:33] Sascha: I've just calculated the first difference score surface. [16:33] Wait a moment. I upload it to my homepage [16:35] brx_: I know, but I didn't expect to see that much regularities [16:36] depending on how "maximal score-difference" is to be interpretated it might or might not explain the regularities [16:38] The values for the score surface, that was calculated with pmars -P (1522 rounds), are exact. [16:39] I've just calculated the difference to a score surface, that was done with only 100 rounds [16:39] vs muskrat ? [16:39] http://corewars.jgutzeit.de/temp/surface.png [16:39] sascha: yap vs. muskrat [16:40] I'll be back.. [16:40] MSG: Quit: Chatzilla 0.9.68.5 [Firefox 1.0.4/20050512] [16:41] Any comments to that difference-score surface? [16:42] when loading is ready... [16:42] my analog 56k modem is getting slower each day...guess the cable will fail in the next days.. [16:42] :( [16:42] Join: willvarfa joined #corewars [16:42] Jens: well, 100 is an okay sample? :) [16:43] brx_: I don't understand. What do you mean? [16:43] the result of 100 rounds is an okay approximation to -P ? [16:44] the average score difference is 10 and on the normal surface you can see most features [16:44] how does the score diff look like for 200 rounds? [16:44] We are looking which number of rounds are an optimal approx to -P..we calculatin 100,200,300,400 and 500 rounds surfaces [16:45] brx_: I'm just computing the difference for 200 rounds. Just wait a couple of minutes [16:45] I would have probably thought about that before beginning to calc more score surfaces since it is... faster :) [16:45] With 200 rounds the max. difference is 44.7, the average is 7 [16:45] ah [16:46] brx_: Yes, but we need exact score surfaces anyway to compare the other results against. [16:46] Jens: let me rephrase, I meant "more exact score surfaces" [16:47] I am aware of the fact that you need at least one precise one for comparison reasons. also, it might look different for other types of warriors. [16:48] Hmm ... the difference score surface for 200 rounds looks exactly the same when scaled [16:48] but the range is less... [16:49] yes, but apart from that, the green and red areas are at the same position [16:49] (more or less) [16:51] For 400 rounds the average difference is 5.8, the max. difference is 34.5 [16:51] Join: brx joined #corewars [16:51] MSG: Ping timeout: 252 seconds [16:51] 18:46 Jens: let me rephrase, I meant "more exact score surfaces" [16:51] 18:47 I am aware of the fact that you need at least one precise one for comparison reasons. also, it [16:51] might look different for other types of warriors. [16:53] you think that pmars could have a better algorithm for choosing the positions when you have less than all perms to compute? [16:53] Yes, it might be different for other warriors. We'll soon calculate one for tiny stone. [16:53] willvarfa: one could check for that. [16:53] s/one/Jens [16:53] :) [16:54] willvarfa: It is too early to say that. [16:54] well this is historic, but the pmars algorithm is known to be not-especially-random [16:54] look at fcrack etc [16:54] I think there was a lot of stuff about it on r.g.c. a long while back [16:54] yes, I've already read about it [16:55] OK: Forget everything. [16:55] Jens: mmmh? [16:55] I've just calculated a difference score surface for 400 round, but with a difference F-value. [16:56] The pattern looks quite different [16:56] f-value? [16:56] pmars has an F-option to seed the rng [16:57] and? [16:57] what do you think those results tell you? [16:58] The rng of pmars is good enough :( [16:58] when you do -r 100 you get 100 rounds; when you do -r 200 you get 200 rounds, and the first 100 are the same as if -r 100 had been used ..? [16:58] (depending on the rng, the seed value could produce *extremely* repetetive patterns) [16:59] hence fcrack [16:59] (There is only 4 F-values that will produce all different distances at 100 rounds, at coresize 8000) [16:59] willvarfa: what is fcrack? I have a general idea but I would like to have it confirmed :) [17:00] MSG: Ping timeout: 252 seconds [17:00] If you know the F-number of hill, you can precalculate all start-distances [17:00] hills have preset -F random seeds; if you know the seed for the hill, you have an advantage if you optimise for it; fcrack finds that starting number for you [17:01] ah right, yeap, thats it Miz [17:02] Certain Abuse used it to get 95% wins when challenging the hill [17:02] lol [17:02] (But died just next challenge, since it could only store the challenge-numers) [17:14] Join: Core29 joined #corewars [17:14] Part: Core29 left #corewars [17:25] did certain-abuse ever get the write-up in corewarrior as the source promised? [17:32] Brx, you want to play ET tonight? Ill be available in 30mins (unless they take Oasis..) [17:56] * willvarfa *waves* [17:56] MSG: Quit: Chatzilla 0.9.67 [Firefox 1.0.7/20050919] [17:59] Mizcu: sure [18:12] Join: sascha joined #corewars [18:13] Back again [18:15] brx: you choose server [18:15] Mizcu: okay, looking [18:15] preferably one that is not 50-player ultrafraghell, or one with many custom maps [18:15] lol [18:15] i got you [18:38] bye [18:38] Part: Jens left #corewars [18:38] Part: sascha left #corewars [18:44] Join: sayembara joined #corewars [18:44] Part: sayembara left #corewars [18:44] hi Saye, bye Saye [19:50] Join: fiveop joined #corewars [20:21] was home for a week and tiny didn't age :/ [21:55] MSG: [21:56] MSG: Quit: humhum