[01:56] MSG: Quit: humhum [08:14] bah. [08:14] missed fizzer [08:22] MSG: Read error: Operation timed out [08:27] MSG: Read error: Operation timed out [09:31] MSG: Ping timeout: 245 seconds [09:31] Join: crem joined #corewars [10:09] Join: Mizcu joined #corewars [10:19] Join: bvowk joined #corewars [10:20] yay, Linksys's drivers didnt allow adding wlan with hidden ssid [10:21] didnt take more than 2 mins to get the card to work in windows vanilla [10:24] and QoS doesnt work in the router, but otherwise things work [11:11] hrm. [11:13] hiya [11:14] whats up mizzer? [11:14] we have a distance-working week at school [11:14] so obviously its a veiled holiday [11:15] heh [11:16] now i just have to figure out how i get that server puteron DMZ (or routed past the box) [11:17] sure, theres DMZ option in the box, but it requires me to specify an external ip.. [12:04] is there a selection anywhere in windows that i can twiddle to make the swapping a little less aggressive? [12:05] it's windows d00d. [12:05] was afraid of that [12:06] i have gig of ram so that my system would starts swapping when running out of memory - but now it swaps stuff out of memory to HD because it hasnt been used in the last five minutes [12:06] ok, so I don't think diversity hills will work [12:07] at least not unmodified.. [12:09] how so? [12:09] experimentation failed? [12:11] Join: fiveop joined #corewars [12:14] yeah.. it's too flappy [12:17] hi [12:23] heya fiver [12:23] how goes? [12:23] i dont think it the problem of diversity hill, but its settings [12:23] I'll leave for france tomorrow :) [12:23] how are you guys? [12:24] well.. maybe it works if I use the scores to pick the best and not the discriminant [12:24] I kick things off the hill using the diversity rules, but everything else uses raw scores.. [12:24] that seems to work pretty well [12:24] it's still pretty flappy [12:24] use the relative diversity score as a multiplier to total score? [12:24] diversity? [12:25] does it echo in here? [12:25] http://www.corewar.co.uk/diversit.htm [12:26] hmm, well, diversity is kinda hard to turn into a singular multiplier, since there is atleast there are atleast three variables [12:26] (WLT, and then you get a chain of WWTTLLTW which you compare to WWLLTTLW) [12:30] (which turns into 0-1-2, but how to compare 001122 and 221100 ?) [12:31] I think perhaps a scaling factor for the diversity score might do it.. [12:35] hrm.. no.. that doesn't work either.. [12:36] perhaps the difference between the min distance of a warrior vs the global minimum? [12:40] i think the diversity is diamond-shaped [12:40] well.. it looks like the diversity space includes all the space where warriors suck. [12:40] you start from low tip where everything scores 0. With improvement you get scoring of 100, 010 and 001 which are the middle-sides [12:40] and top would be 111 for all warriors [12:41] so, the diversity scores for really shitty warriors puts them right up there :) [12:42] however, the diamond comparison sucks when you notice that you get only stones and papers, no scanners [12:49] this is what i have in mind right now: bench all the warriors: when a warrior y scores better aginst z than x, y gets a point. repeat for all warriors, see who has the most scores [12:49] "scores better" means recieving more ties than he gets losses and wins than he gets ties [12:50] in short, turning the diversity in to competition [12:50] "i beat the guy who beat you, i am better than you!" [12:51] im playing Civ on the side, so this is though with only one half of brain [13:10] (what is this silence i hear?) [13:11] so why are you going to france fiver? [13:12] to ski [13:12] well thats a pretty sweet deal :) [14:54] Join: Metcalf joined #corewars [14:54] Hi [14:56] hi [14:58] Hi Fiveop [14:58] Anything new? :-) [15:00] hrm. [15:00] greetings metster! [15:01] nothing new [15:01] Hi Bvowk [15:01] * Metcalf has just been writing a bookmarklet [15:01] Will upload in a minute [15:02] bookmarklet [15:02] hum [15:02] :) [15:04] hrm.. [15:04] so if I use the average distance instead of the minimum distance it works better for my purposes.. [15:05] the minimum distance would flap pretty hard, but the average forces warrior groups that are the same down, and their weaker members get pushed off and then the stronger ones rise again [15:05] I hope you're going to write up your results bvowk. [15:06] hrm [15:06] that sounds like work tho met ;) [15:07] but if I use the average, the optimization makes a steady walk towards better [15:07] instead of a random walk that kinda sorta leans towards better [15:08] and the diversity seems to be better.. as in the warrior distance doesn't seem to collapse after a while [15:09] we'll leave it running a while and see what happens [15:09] so what are you hacking on met? [15:11] at the moment, koth.org html. Just uploaded the front page with bookmarklet added [15:52] Join: OoS joined #corewars [15:54] MSG: Ping timeout: 245 seconds [15:55] Nick Change: OoS changed nick to Metcalf [16:32] Hmmm... [16:33] I've discovered a vunerability in a popular website which lets me generate free credits ;-) [16:40] credits? [16:42] MSG: Ping timeout: 245 seconds [16:42] Join: Metcalf joined #corewars [16:43] The website works on a credit system [16:43] You can buy them, or earn them [16:48] and apparently, steal [16:52] Written a JavaScript to automatically steal them now ;-) [17:11] tzz :) [17:25] MSG: Ping timeout: 245 seconds [17:26] Join: Metcalf joined #corewars [17:58] MSG: Ping timeout: 245 seconds [17:59] Join: Metcalf joined #corewars [18:05] MSG: Ping timeout: 245 seconds [18:29] MSG: Quit: humhum [18:31] Join: crem___ joined #corewars [18:31] MSG: Read error: Connection reset by peer [20:41] Join: jon____ joined #corewars [20:41] MSG: Read error: Operation timed out [23:23] Join: Metcalf joined #corewars [23:27] MSG: Ping timeout: 245 seconds [23:27] Join: Metcalf joined #corewars [23:33] MSG: Ping timeout: 245 seconds