[02:39] Join: myndzi\ joined #corewars [02:43] MSG: Ping timeout: 245 seconds [03:00] Join: myndzi joined #corewars [03:03] MSG: Ping timeout: 245 seconds [08:43] Join: yoR joined #corewars [08:44] So Miz, what was your idea (in the logs 31-8?) [09:14] Join: fiveop joined #corewars [10:48] Join: sh0ne joined #corewars [10:49] Hello [11:12] hi sh0ne [11:13] Roy: a type of paper [11:38] ugh, i cant figure out what to say without spilling the beans [12:18] Heh its ok :) [12:22] i figured out how to get "free" work out of a paper without adding instructions, but it requires the rest of the paper to be carefully sculpted, so the point is kinda blunt [12:23] no, not related to the expanding attack-carpet [13:16] Sounds intresting enough, why don't you try it instead of telling JM every time ;-) [13:56] MSG: Quit: Leaving [15:23] greetings humansx. [15:23] er.. humans. [15:23] :) [15:24] how goes fiver? [15:26] I'm well, thanks :) [15:26] anything new in canada? [15:38] not much. [15:38] first day the students are back on campus. [15:39] what a fucking horde. [15:39] :) [16:00] So, anyone hacking anything of interest? [16:18] Yeah? [16:18] what? [16:18] I'm trying to get a bit higher score in RecMath... [16:18] But not much 'luck' yet [16:19] (not really luck I guess) [16:19] * yoR just sucks :) [16:21] hm [16:21] You guys should try to, I always liked primes and the recmath competitions, so I really enjoy this one [16:22] Good oppertunity for bvowk to get revenge, I've teased him enough now [16:26] sounds like a number crunshing exercise [16:34] heh [16:34] hey roy. [16:34] we're thinking about it. [16:34] yay :) [16:34] * bvowk goes to check out roys score. [16:34] fiveop: It does, but there are lots of ways to crunsh numbers, and I'm not planning on calculating ALL the combinations (which is kind of impossible) [16:35] *spoiler* I'm currently using a very basic GA algorithm to calculate random scores and mutate/crossover.. but the scores aren't anything near 'good' [16:36] wow.. [16:36] 75. [16:36] ouch roy ;) [16:36] But during a discussion with a friend we might have found another better method to get combinations, and if it works I'll put the GA algorithm on that (smarter) method [16:36] * bvowk ducks. [16:36] Words don't hurt, better scores do, good luck ;-) [16:40] But the contest is harder then I thought, I really thought I had some good scores, but it turns out to be bad :) [16:40] Although I am pleased to have found perfect solutions for 4 and 5... the others are much much harder [16:45] that's how they usually work :) [16:46] I noticed, but in the crosswords puzzle it went easier then I thought until I got in the top 5/6... [16:46] Much closer to the top.. [16:47] The ceiling of my current approach is just shockinly low, so there must be a whole different better take on this challenge I have to find [17:06] yoR: random cross over sounds stupid [17:07] you should take subsets that together already produce a majority of the primes and put them together with others [17:07] Join: CoreOld joined #corewars [17:08] fiveop.. shush! ;) [17:08] that's basically just reminding of the obvious :P [17:08] :) [17:09] Yeah well, thats what a GA does fast enough for you, combinations with good scores cross with other combinations with good scores [17:09] well [17:09] obviously the obvious isn't that obvious to certain people [17:09] :) [17:10] heh [17:10] I just don't want you helping roy. [17:10] ;) [17:10] But a subset that creates primes (for example a solution for N=4) doesn't really create a good base for N>4 (I've seen in tests) [17:11] it's starts with the obvious.. and then ends with solid proofs for toy problem sets ;) [17:12] bvowk: I don't mind the secressy, but I think I even told you exacly how my algorithm worked during the crosswords challenge, and still my laptop with a silly Java program scored (magrinally) better [17:13] someone's competitive [17:14] heh [17:14] roy: yeah, the crossword one didn't go so well.. [17:15] although we did have a really fucking slick dawg hack [17:15] but the placement algorithm sucked somehow really hard.. [17:15] 1. I am, 2. Personal competition(s) drives me to score better, and create more fun/tension [17:16] wil made a seriously impressive C dawg that worked like lightning. [17:16] it was in cache for the scoring, so it was fast fast :) [17:16] My program just created a list of 'good' words which had a lot of overlap and it tried to place them in a way that had as much overlap as possible (pretty straight forward actually) [17:17] pretty much the same for us. [17:18] With this competition the scoring isn't so hard, you can easily put all the possible primes in a hit/no-hit list, fastest prime checking possible, and I've also added some checks that where even faster (eliminating multiples of 2,3,5,7) [17:20] But no doubt it could be done ever quicker in some cases, but I'd rather put my time in figuring a good algorithm, just as it showed in the crosswords, a kick-ass scoring system doesn't score by its self... better algorithms, even slow ones (in Java) can kick yer ass [17:21] But I'm going home now, been here waaaaay too long (here equ work) [17:21] MSG: Read error: Connection reset by peer [18:43] Join: Metcalf joined #corewars [18:44] * Metcalf enters rec.math :-) [18:47] hi Met [18:47] :) [18:47] Hi Mizcu, Fiveop [18:47] I've got quite a way to catch up with Roy [18:48] Has anyone else entered? [18:50] haven't and won't [18:52] This one isn't too bad [18:52] I didn't like all the geometry contests [18:53] hey guys [18:53] what's this contest you're talking about? [18:53] recmath.org [18:53] mhm [18:54] ah [19:18] greets! [19:19] Hi Bvowk [19:21] how's it going met? [19:24] Not very well yet [19:25] I'm just writing a program for recmath [19:39] MSG: Quit: humhum [20:16] MSG: Quit: Trillian (http://www.ceruleanstudios.com [20:31] Join: CoreOld joined #corewars [20:40] aw. raised the benchmark s cores quite a bit, but it didn't help me at all [20:40] not enough scanners perhaps [20:54] MSG: Ping timeout: 245 seconds [20:55] MSG: Quit: Trillian (http://www.ceruleanstudios.com [21:16] hrmmmm [21:17] looks like optimax + optimax + corewin = crash [21:17] lol. [21:20] optimax always crashes my computer after a few hours [21:27] it ran by itself just fine for weeks [21:27] i'd forgotten it was running [21:27] but i tried to run a benchmark in corewin while optimax was going and woops [21:27] it happened the other day too [21:27] not every time, just sometimse [21:27] times* [21:27] :-( [21:27] hard to believe it's overheat, too [21:28] since my proc varies like 4c max [21:28] and now that truecrypt is installed i can't remote desktop home and run everything again lol [21:31] i'm gonna have to get vmware or something [21:31] there's another app i have to run in it too [21:32] stuff should run pretty well, i have a recent proc [21:32] was it virtualbox that was on slashdot the other day? [21:36] I think I've only been to slashdot once [21:52] oh well, i think it is [21:52] it compared favorably in many aspects to vmware, but it's free so .. :>