2011-06-20T00:08:52 _flag: well yes 2011-06-20T00:08:56 it should be up 2011-06-20T00:09:26 but it seems that it is currently taking 100% of my servers cpu time and not responding 2011-06-20T00:10:03 *** amriedle has quit IRC (Quit: leaving) 2011-06-20T00:40:15 <_flag> Rabidus: I tried connecting awhile ago but it seemed to be perpetually waiting for a game 2011-06-20T00:40:27 so it responded? 2011-06-20T00:40:32 that's something :) 2011-06-20T00:41:02 but yeah, that seems to have some major bugs and i don't have the time to fix em 2011-06-20T00:41:25 so i'll start it again when it is working better 2011-06-20T00:41:35 <_flag> Well at least there's a tcp server on the way :) 2011-06-20T00:41:41 yep :) 2011-06-20T00:43:36 *** computerwiz_222 has quit IRC (Read error: Operation timed out) 2011-06-20T01:04:42 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Read error: Connection reset by peer) 2011-06-20T01:28:58 *** Accoun has quit IRC () 2011-06-20T01:31:13 *** Accoun has joined #aichallenge 2011-06-20T01:31:13 *** Accoun is now known as Account 2011-06-20T01:31:24 *** Account is now known as Accoun 2011-06-20T01:39:05 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-06-20T01:45:23 *** stocha has joined #aichallenge 2011-06-20T02:00:05 *** filter has quit IRC (Ping timeout: 240 seconds) 2011-06-20T02:03:15 *** filter has joined #aichallenge 2011-06-20T02:03:16 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T02:06:35 *** smellyhippy has quit IRC (Ping timeout: 276 seconds) 2011-06-20T02:18:29 *** locutus2 has joined #aichallenge 2011-06-20T02:27:21 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-20T03:28:08 *** filter_ has joined #aichallenge 2011-06-20T03:28:52 *** Palmik has joined #aichallenge 2011-06-20T03:29:11 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-06-20T03:31:21 *** filter has quit IRC (Ping timeout: 264 seconds) 2011-06-20T03:31:24 *** boegel has joined #aichallenge 2011-06-20T03:37:58 well i have a bot up 2011-06-20T03:37:59 it' 2011-06-20T03:38:05 s not very smart when it comes to battles though 2011-06-20T03:41:15 *** sigh has quit IRC (Quit: This computer has gone to sleep) 2011-06-20T03:45:44 *** dr- has joined #aichallenge 2011-06-20T03:48:24 bah, the 10 minute wait sucks when you realise an obvious bug that causes friend collisions :P 2011-06-20T03:50:59 antimatroid: haha, that's so easy to code a check for 2011-06-20T03:51:30 mleise: it's less obvious with what i'm doing, but i put in a check that should work for the moment 2011-06-20T03:51:38 but i had make sure antPlayer != 1 2011-06-20T03:51:41 when it should be 0 :P 2011-06-20T03:52:05 hehe, "off by one" is a common bug 2011-06-20T03:54:28 anyway 'debug painter' got quite a bit faster. I'm still worried about the speed with several layers of debug info though. 2011-06-20T03:55:09 nice 2011-06-20T03:55:23 i have the fixed bot up 2011-06-20T03:55:31 i'll leave it now and see how it goes ranking wise 2011-06-20T03:56:08 mleise: did you see my wtf comment about c++ and priority_queues? 2011-06-20T03:56:17 no, where? 2011-06-20T03:56:27 priority_queues are cool. 2011-06-20T03:56:30 priority_queue, greater > t, t.top() returns the least element in the queue 2011-06-20T03:56:31 priority_queue, less > t, t.top() returns the greatest element in the queue 2011-06-20T03:56:33 I used them in java 2011-06-20T03:57:05 that makes no sense to me :\ 2011-06-20T03:57:06 you mean the sorting is wrong? 2011-06-20T03:57:20 well, wouldn't you expect it to be the other way around? 2011-06-20T03:57:35 D has the same 'issue' :p . Yes, you are right 2011-06-20T03:57:46 is there a reasonable explanation as to why? 2011-06-20T03:58:10 I don't know :( 2011-06-20T03:58:25 let me check Java's queues 2011-06-20T03:58:39 i will cry if they're more reasonable about something 2011-06-20T03:59:30 The head of this queue is the least element with respect to the specified ordering. 2011-06-20T04:00:23 and that would by default be the natural ordering, which is... now im confused 2011-06-20T04:01:10 a < b -> 1,2,3,4,5 2011-06-20T04:01:19 antimatroid: why don't you make a struct vector where you assign a priority number into the struct and then sort that vector by priority number? 2011-06-20T04:01:23 or am i missing something? 2011-06-20T04:01:28 that seems right to me. so Java is doing it the other way round 2011-06-20T04:01:58 Rabidus: isn't that what a priority queue is? i'm just peeved at the counterintuitive way it's done in c++ 2011-06-20T04:02:30 i really have a priority_queue of search information 2011-06-20T04:03:54 I have ie. struct FoodSearch, with operator> defined on FoodSearch, then priority_queue, greater > 2011-06-20T04:06:25 *** locutus2 has joined #aichallenge 2011-06-20T04:08:02 mleise: it makes no sense to me :\ 2011-06-20T04:09:39 *** Naktibalda has joined #aichallenge 2011-06-20T04:11:42 *** zeLark has quit IRC (Ping timeout: 252 seconds) 2011-06-20T04:17:53 ah the Tango library for D has specialized templates MinHeap and MaxHeap to avoid confusion, but it is not ported to D2 yet. 2011-06-20T04:21:58 *** escortkeel has joined #aichallenge 2011-06-20T04:22:11 hi! 2011-06-20T04:23:29 *** escortkeel has quit IRC (Client Quit) 2011-06-20T04:31:53 *** Ice_Harley has joined #aichallenge 2011-06-20T04:36:31 *** stocha has joined #aichallenge 2011-06-20T04:37:23 *** dr- has quit IRC (Ping timeout: 276 seconds) 2011-06-20T04:39:32 looking at the beta server, vs my own framework, i have this strong feeling that you should at least try some* assymetric maps. 2011-06-20T04:40:07 i mean, my measure of strength were as precise as the test server. (though arguably you *might need a litle bit more sampling) 2011-06-20T04:40:22 then the assymetric maps, feels much more diverse. 2011-06-20T04:40:56 it's actually intersting to watch. When the test server is kinda of boring like hell; 2011-06-20T04:41:39 *** dr- has joined #aichallenge 2011-06-20T04:41:43 antimatroid: how is your bot doing ? 2011-06-20T04:42:20 stocha: pretty dismally 2011-06-20T04:42:26 oh god, my poor ref bot droped to 5th. I need to spend time on a better bot. 2011-06-20T04:42:50 This page is generated by Parallels Plesk Panel, the leading hosting automation software. You see this page because there is no Web site at this address. 2011-06-20T04:42:53 ai-contest.com 2011-06-20T04:43:02 looking at the starter bots, i wonder if i shouldn't put in there one weak bot (like pure random with collision avoidance or such) 2011-06-20T04:48:02 Assymetric map forbidenness feels like an arbitrary decision to me. 2011-06-20T04:48:04 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-20T04:48:19 open arena maps 2011-06-20T04:48:20 no water at all 2011-06-20T04:48:38 Rabidus: ? 2011-06-20T04:48:55 ai-contest doesn't open for me 2011-06-20T04:49:21 yeah seems to be down. 2011-06-20T04:49:27 what about the open arena maps ? 2011-06-20T04:50:21 no water at all 2011-06-20T04:50:32 Rabidus: yeah what about them ? 2011-06-20T04:50:59 well... it was just a suggestion :D 2011-06-20T04:51:05 nothing deep 2011-06-20T04:51:44 Rabidus: yeah, i think a lot of people would like to see more diversity. More open (or even in extreme cases full open) maps being one kind of diversity that strike mind. 2011-06-20T04:51:50 rabidus: symmetric_21-30 has a few more open maps 2011-06-20T04:52:02 ah, ok 2011-06-20T04:52:06 antimatroid: the maps on the test server are static ? 2011-06-20T04:52:21 21-40 were added in the last day 2011-06-20T04:52:52 http://aichallengebeta.hypertriangle.com/visualizer.php?game=28462&user=126 2011-06-20T04:52:53 eg 2011-06-20T04:53:40 antimatroid: it's your fooding scheme running ? Your ants have got a depression problem. 2011-06-20T04:54:06 stocha: it's running a few things 2011-06-20T04:54:10 and there are bugs all over the show ;) 2011-06-20T04:55:22 you have rank 6th ... 2011-06-20T04:56:58 seems you added some spreading code :p 2011-06-20T04:58:35 i did :P 2011-06-20T04:58:56 i collect food with a*, then enemies with a* then non-visible squares with bfs 2011-06-20T04:59:19 a* only for food ? 2011-06-20T04:59:27 and for enemies 2011-06-20T04:59:33 it seems you have a rebound mechanisms for food :p 2011-06-20T05:01:16 stilly game (leading bots crashed. While coward bot will finally take them out ) http://aichallengebeta.hypertriangle.com/visualizer.php?game=28482&user=126 2011-06-20T05:01:34 rebound mechanisms? 2011-06-20T05:01:57 yeah i call rebound, when you use a yet to spawn ant to do something. 2011-06-20T05:02:14 ah yeah, i do that 2011-06-20T05:02:19 and where my ant will be in the future 2011-06-20T05:02:19 for example rebound greedy food (you assigne closest ant to food, but can use yet to spawn ants) 2011-06-20T05:02:38 i still have a few tweaks to do to food collection, but the barebones are there 2011-06-20T05:02:48 it's doing a* with multiple sources and multiple targets 2011-06-20T05:03:01 you do a full solver, or only do partial ? 2011-06-20T05:03:13 what do you mean by full? 2011-06-20T05:03:30 i restart the search each iteration, finding the ant that can collect a piece of food the fastest 2011-06-20T05:03:39 i mean, you do it like recursively vs doing it for next 10 turns. 2011-06-20T05:03:56 next 10 turns? 2011-06-20T05:04:01 i just collect every piece of food 2011-06-20T05:04:21 You do that with the optimal gathering, or just an approximation ? 2011-06-20T05:04:36 define optimal gathering? 2011-06-20T05:05:16 it's the shortest time to food, with the least ant used (with non traffic congestion handling) 2011-06-20T05:05:36 no traffic congestion handling 2011-06-20T05:05:45 i have no traffic congestion handling 2011-06-20T05:05:50 then you have the shortest time to food, least ant assigned, with traffic congestion :p 2011-06-20T05:06:07 how much cpu does the fooding need ? 2011-06-20T05:06:13 is it what makes you time out ? 2011-06-20T05:06:15 not sure 2011-06-20T05:06:24 it shouldn't time out at all 2011-06-20T05:06:29 it has the timer checking for that 2011-06-20T05:06:34 i think sir_macelon also use a rebound mechanisms 2011-06-20T05:06:36 like i said, it's quite buggy :P 2011-06-20T05:07:20 I corrected the known obvious bug in my framework. i may work a new generation of bot, if i have time and mood. 2011-06-20T05:07:49 I'm unsure of what to do about combat though. 2011-06-20T05:08:22 for non combat related, i have a few idea, and nothing more. But i guess it would make the bot stronger. 2011-06-20T05:08:46 yeah, i've figured out most things bar combat 2011-06-20T05:08:51 Mainly i have to build a scheduling framework 2011-06-20T05:09:35 maybe i'll first try to do that, and only then see how i can use it to be in the 3 best bot around :p 2011-06-20T05:11:45 i see two possible kind of scheduler. Light (ants have goal, and an expected time of arrival) and heavy. (ants has goal, and a timed precise path) 2011-06-20T05:12:03 How many people in there have scheduler ? who plans on building a heavy one ? 2011-06-20T05:13:17 i guess i'm speaking loud. Given out my ideas :p I guess having good ideas and speaking them louds randomly feels better than winning a such contest. 2011-06-20T05:16:07 For now, i like my map generator better (pure assymetric : random walk with random land mass, and pure random first ant spawn on a empty land). Then for food, i go with a food spawn speed from 0 to 200 (0 being no food, and 0-100 percentage of a food spawning during a turn. And 100-200 100% chance of spawning one food, and partial chances to spawn a second one). Then i have a set percentage of ground map as initial food (from 2011-06-20T05:16:34 *** smiley1983 has joined #aichallenge 2011-06-20T05:16:52 what about trying that in the test server and see how things goes ? 2011-06-20T05:17:08 up to very big maps. 2011-06-20T05:17:57 We need as simple and efficient map generation scheme, at least for 20% of the maps. 2011-06-20T05:18:15 so i would suggest to just use this one with a 20% chance for each game. 2011-06-20T05:19:44 We could use the map generation scheme to find intersting upper bounds. 2011-06-20T05:19:55 this map generation 2011-06-20T05:20:55 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T05:26:27 smiley1983: i have a bot up on the main site :P 2011-06-20T05:26:32 well, beta site 2011-06-20T06:00:23 *** stocha has joined #aichallenge 2011-06-20T06:00:31 here is a nice pathological game : 2011-06-20T06:00:33 http://aichallengebeta.hypertriangle.com/visualizer.php?game=28495&user=107 2011-06-20T06:01:02 stocha, and Sto03RepliCrush are running the same code for the first 600 turns. 2011-06-20T06:01:10 the assymetry is easy to spot. 2011-06-20T06:01:21 Hence symetric maps doesn't bring much to the table really. 2011-06-20T06:03:03 (the map is also a good example of the score system working ... but being defeated by the time out of the leading bot. Which screwed the end ranking by ending the game in favor of the loosing bot. 2011-06-20T06:03:39 don't make a bot win, for only being the last to time out, maybe. 2011-06-20T06:03:47 my bots retarded 2011-06-20T06:04:09 retarded ? 2011-06-20T06:06:06 it's colliding with friends again 2011-06-20T06:08:07 *** Zaphus has joined #aichallenge 2011-06-20T06:08:27 sir_macelon: implemented a digestion algorithm : http://aichallengebeta.hypertriangle.com/visualizer.php?game=28532&user=107 2011-06-20T06:09:11 over all, the sir_macelon is the bot that convinced me the most. 2011-06-20T06:10:40 http://aichallengebeta.hypertriangle.com/visualizer.php?game=28515&user=16 2011-06-20T06:11:05 Zaphus: ? 2011-06-20T06:11:24 just thought it was interesting (apart from the timeout) 2011-06-20T06:11:41 and it's the digestion algorithm (I guess) 2011-06-20T06:13:25 Zaphus: true this is an intresting game, for showing the internals of the sir_macelon bot :p 2011-06-20T06:13:43 I'm sure he will not be happy that his code is exposed so well :p 2011-06-20T06:14:31 it's an interesting behaviour, works pretty well 2011-06-20T06:14:59 to be faire the maps do not have a large variety. 2011-06-20T06:15:13 for example, sir_macelon might well do bad in more open maps. 2011-06-20T06:15:22 this ai-challenge will be intresting 2011-06-20T06:15:26 (although the food collection is pretty good, so it might still do well :p ) 2011-06-20T06:15:39 next few weeks will be interesting 2011-06-20T06:15:50 Rabidus: you planed something for us ? :) 2011-06-20T06:15:55 ;) 2011-06-20T06:16:05 näh 2011-06-20T06:16:16 just watching top5 bots 2011-06-20T06:16:48 time out system, shouldn't end the game. It screw the result. 2011-06-20T06:18:36 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T06:21:34 *** FireFly has joined #aichallenge 2011-06-20T06:27:41 Zaphus: I put a bot up :P 2011-06-20T06:27:48 it keeps colliding with itself :\ 2011-06-20T06:34:09 stocha: the rules state you can't time out, although it would be good if your ants still got any points they earnt 2011-06-20T06:34:38 *** onensora has joined #aichallenge 2011-06-20T06:39:15 once upon a time timing out resulted in being disqualifie 2011-06-20T06:39:15 d 2011-06-20T06:39:39 back in the dark ages :P 2011-06-20T06:44:51 *** Apophis_ has joined #aichallenge 2011-06-20T06:47:54 antimatroid: yes, those were dark days.... any reason timed-out bots couldnt be allowed to accumulate points and possibly win - just not do any more turns ? 2011-06-20T06:48:41 not really sure 2011-06-20T06:48:52 i think we've sort of slippery sloped our way to where we are 2011-06-20T06:48:57 *** nann has quit IRC (Ping timeout: 255 seconds) 2011-06-20T06:49:11 the only reason the ants are left on the map is to try and keep it relatively fair 2011-06-20T06:50:13 technically a timed-out bot could still win, right ? I mean in theory, if they time out near the end *and* the food production isnt too high ? 2011-06-20T06:53:03 yeah 2011-06-20T06:54:25 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has joined #aichallenge 2011-06-20T07:01:31 *** mcstar has joined #aichallenge 2011-06-20T07:02:53 actually that timeout thing must be handled by bot itself. if it sees that he's running out of time, it's his fault 2011-06-20T07:04:34 *** locutus2 has joined #aichallenge 2011-06-20T07:12:35 lol, somehow my bot is up to third 2011-06-20T07:14:07 vn 2011-06-20T07:15:25 have any idea why your bot is timeouting? 2011-06-20T07:17:33 how often is it timing out? 2011-06-20T07:17:39 it's quite buggy 2011-06-20T07:17:46 i just threw together my a* stuff 2011-06-20T07:18:00 i'm not doing anything for battles other than chasing after enemy ants 2011-06-20T07:18:25 i think i fixed what is causing it to collide with friends 2011-06-20T07:26:20 searching for not-visible squares 2011-06-20T07:26:26 anyone got any idea of how to do that not with bfs? 2011-06-20T07:30:21 *** george1914 has joined #aichallenge 2011-06-20T07:30:42 george1914: http://aichallengebeta.hypertriangle.com/language_profile.php?language=C%2B%2B :P 2011-06-20T07:35:17 *** olexs has joined #aichallenge 2011-06-20T07:36:47 antimatroid: hi, looks like you're doing well :) 2011-06-20T07:37:02 george1914: it's still fairly basic 2011-06-20T07:37:08 Antimatroid: I get a blank page with that link (all green where it should have a table) 2011-06-20T07:37:11 i'm doing good food collection etc. though 2011-06-20T07:37:38 thats with IE9, Chrome shows it properly 2011-06-20T07:37:44 it's using a* to collect food, then again to collect enemies, then bfs to "bomb" not visible squares (when i collect a visible square i bomb the visiblity information around it before the next bfs) 2011-06-20T07:37:52 *** smellyhippy has joined #aichallenge 2011-06-20T07:38:13 *** locutus2 has quit IRC (Ping timeout: 260 seconds) 2011-06-20T07:38:57 antimatroid: bomb? you mean you calculate information about it (eg distances etc) and store it for quick use later? 2011-06-20T07:39:38 george1914: when i move an ant towards a not visible square, before searching with the rest of my unmoved ants, i'll update the vision information around the square i was moving the ant towards 2011-06-20T07:39:49 so it doesn't send another ant right near that point 2011-06-20T07:39:56 ./region 2011-06-20T07:40:55 ok, i more or less follow. 2011-06-20T07:41:08 i have to make a copy of the state to do it 2011-06-20T07:43:47 antimatroid: nice expansion rate 2011-06-20T07:44:00 mcstar: :P 2011-06-20T07:44:06 i have a c++ question 2011-06-20T07:44:13 how can i check if something is in a set? 2011-06-20T07:44:15 there is count? 2011-06-20T07:44:33 but surely an elem function would be faster, because it could stop once it found a single instance 2011-06-20T07:45:14 antimatroid: std::set only allows 1. 2011-06-20T07:45:21 It's called "count" for consistency, I think. 2011-06-20T07:45:30 ah fair enough, sillyness 2011-06-20T07:45:31 http://www.cplusplus.com/reference/stl/set/count/ 2011-06-20T07:45:41 c++ is full of wacky stuff for me lately 2011-06-20T07:45:54 i still think the priority queue thing is retarded 2011-06-20T07:46:23 i saw some comments yesterday about an experimental tcp server ... 2011-06-20T07:46:31 not fully functional sorry 2011-06-20T07:47:30 right, i am a bit paranoid - was wanting to use it rather than the beta - not that i have any clever code just yet, but ... 2011-06-20T07:47:41 i guess i should wait ... 2011-06-20T07:47:43 you can have a vactor or whatever you call it in c++ and have it always sorted, so that an elem function could be efficient 2011-06-20T07:48:24 mcstar: He just doesn't like the name "top" for the first element of a priority queue. 2011-06-20T07:48:41 antimatroid: i have implemented a* also - though maybe not so well - sometimes it is slower than bfs, and usually just gives me a factor of 2 or 3 speedup ... 2011-06-20T07:48:43 i see 2011-06-20T07:49:13 george1914: are you doing it with all food as targets? 2011-06-20T07:49:16 i have mainly gone for calculating useful distances etc when i see new squares, and stroing that for lookup ... 2011-06-20T07:49:20 i'm using "future ant location information" too 2011-06-20T07:49:50 antimatroid: no - just calcDistance(loc1,loc2) 2011-06-20T07:50:01 jbroman: top isn't my priority queue problem 2011-06-20T07:50:02 you seem to be doing cleverer stuff ... 2011-06-20T07:50:20 *** computerwiz_222 has joined #aichallenge 2011-06-20T07:50:20 priority_queue, greater > t, t.top() returns the least element in the queue 2011-06-20T07:50:21 priority_queue, less > t, t.top() returns the greatest element in the queue 2011-06-20T07:50:21 that's completely counterintuitive! 2011-06-20T07:51:03 lol, why is that? 2011-06-20T07:51:13 i have no idea :( 2011-06-20T07:51:16 *** locutus2 has joined #aichallenge 2011-06-20T07:51:23 nono, i mean its perfectly ok with me 2011-06-20T07:51:55 it its sorted by >, than the last is the smallest element 2011-06-20T07:52:09 why? 2011-06-20T07:52:34 cause the vector consists of a,b,c; and a>b, b>c 2011-06-20T07:52:39 so c is the smallest one 2011-06-20T07:52:54 top, is the last element 2011-06-20T07:53:25 hmmm okay 2011-06-20T07:53:31 greater or less is the operation that leaves a properly ordered vector invariant 2011-06-20T07:53:33 i can live with that 2011-06-20T07:53:54 mcstar: (Except that it's actually a heap nor a sorted vector :P) 2011-06-20T07:54:10 (But equivalent for the purposes of that discussion.) 2011-06-20T07:54:17 s/nor/not/ 2011-06-20T07:54:20 yep, i should have said IF it were ordered, it would leave it invariant 2011-06-20T07:54:32 george1914: i found it easiest to have like struct FoodSearch, operator> for FoodSearch objects and priority_queue, greater > 2011-06-20T07:55:09 then i also have EnemySearch and VisionSearch 2011-06-20T07:55:32 antimatroid: is <> a type sprcifier syntax? 2011-06-20T07:55:39 r->e 2011-06-20T07:55:53 it's a bit gross, it's how you add templates 2011-06-20T07:56:08 and you can't close with >>, has to be > > > > whatever 2011-06-20T07:56:12 ok, im in the dark there 2011-06-20T07:56:15 *** amstan_ has joined #aichallenge 2011-06-20T07:56:15 *** ChanServ sets mode: +o amstan_ 2011-06-20T07:56:20 apparently tht is fixed with the new c++ thankfully 2011-06-20T07:56:35 mcstar: i basically only know c++ :P 2011-06-20T07:56:52 and not even that well obviously :P 2011-06-20T07:57:00 thats not a problem, just dont use it for scripting purposes :D 2011-06-20T07:57:12 i use it for all purposes :P 2011-06-20T07:58:19 i love how well a* works 2011-06-20T07:58:30 watching the visited squares in searches is neat 2011-06-20T07:58:55 i soo would love to deal with those 2011-06-20T07:59:01 but i cant 2011-06-20T07:59:38 i still have a bug in my engine: for no reason very rarely an ant show up on the side another one 2011-06-20T07:59:45 s 2011-06-20T07:59:56 after a food square is eaten 2011-06-20T08:00:21 a bug in your engine? 2011-06-20T08:00:25 i just use the official engine 2011-06-20T08:00:38 ok, i dont :) 2011-06-20T08:00:59 i just cant put my mind to anything else until i fix that 2011-06-20T08:04:33 *** amstan_ has quit IRC (Ping timeout: 255 seconds) 2011-06-20T08:05:55 ICANN approved the idea of additional gTLDs :/ 2011-06-20T08:06:05 "Internet address names will be able to end with almost any word in any language, offering organizations around the world the opportunity to market their brand, products, community or cause in new and innovative ways." 2011-06-20T08:06:39 So soon we'll get weird domain names like "buy.potatoes" or something. 2011-06-20T08:06:59 (maybe - hopefully not that bad) 2011-06-20T08:07:07 theres gonna be sex.sex 2011-06-20T08:10:48 i wonder if we will be able to access chinese sites 2011-06-20T08:11:05 imagine an address consisting of kanji 2011-06-20T08:11:14 mcstar: \exists a way of encoding this in ASCII 2011-06-20T08:11:50 IDN 2011-06-20T08:12:31 is ipv6 going to replace ipv4? 2011-06-20T08:12:42 im not up to date on this 2011-06-20T08:13:20 Hopefully eventually, but I haven't seen any sea changes in that regard lately. 2011-06-20T08:14:07 its sad, that i dont fully understand ipv4, and when i will do, there isnt gonna be one :D 2011-06-20T08:18:49 im starting to think its not a bug at all 2011-06-20T08:19:12 sorry for the spam, im just a bit frustrated 2011-06-20T08:19:37 and i really cant talk to my dog about these matters 2011-06-20T08:19:57 *** Eruonen has joined #aichallenge 2011-06-20T08:21:58 *** onensora has quit IRC (Ping timeout: 260 seconds) 2011-06-20T08:22:26 *** Eruonen is now known as onensora 2011-06-20T08:24:05 @later tell stocha could you not share your bot code anymore? 2011-06-20T08:24:05 antimatroid: Aye, aye, sir 2011-06-20T08:24:41 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-20T08:25:15 just put up another, hopefully better, bot 2011-06-20T08:25:29 it's still not doing anything about battles, that'll come much later 2011-06-20T08:29:00 jbroman: i didnt know about IDN, i still remember when they tought us "internet" never to use anything other than latin characters 2011-06-20T08:38:32 definitely no bug, hurray, only the visualizer didnt have enough time to show a food square which was spawned in the spwn radius of another ant, so it seemed like it turned into an ant out of the blue 2011-06-20T08:43:21 *** computerwiz_222 has quit IRC (Ping timeout: 264 seconds) 2011-06-20T08:47:48 antimatroid: are your ants persistent? do you associate long term targets to them, or you recalculate targets on every turn? 2011-06-20T08:50:53 *** smiley1983 has quit IRC (Ping timeout: 276 seconds) 2011-06-20T08:52:06 *** smiley1983 has joined #aichallenge 2011-06-20T08:53:10 mcstar: i don't store anything like that between turns 2011-06-20T08:53:17 and never plan t 2011-06-20T08:53:18 o 2011-06-20T08:53:27 *** smellyhippy has quit IRC (Changing host) 2011-06-20T08:53:28 *** smellyhippy has joined #aichallenge 2011-06-20T08:54:00 im surprised, i definitely itend to, but its gonna cost me some time to actually implement it 2011-06-20T08:54:17 i used old information in tron 2011-06-20T08:54:25 but not for something like this 2011-06-20T08:54:36 not to intersect yourself? 2011-06-20T08:54:56 too much stuff is changing 2011-06-20T08:55:02 too messy to really get anything out of it 2011-06-20T08:55:13 at least for my liking 2011-06-20T08:55:27 but otherwise it seems that theres a lot of wasted effort 2011-06-20T08:58:35 hmmm, it may work to do it for food collection 2011-06-20T08:58:45 wouldn't be too much effort either 2011-06-20T08:59:51 i still don't know how to do battles 2011-06-20T08:59:57 i will try game trees, but i worry about it not working 2011-06-20T09:03:06 i realized theres not going to be a bot written in bash 2011-06-20T09:04:00 bash is powerful, cause it can use any utility that accept stdin, but it also means that a new process is spawned every time we use cat/bc/awk/sed 2011-06-20T09:04:13 maybe for even echo, im not sure about that 2011-06-20T09:05:00 oh, yeah, even for $() a new process is started 2011-06-20T09:05:08 hell, it definitely wont work 2011-06-20T09:07:10 *** Zaphus has quit IRC (Quit: Page closed) 2011-06-20T09:18:54 http://ai-contest.com/ down 2011-06-20T09:19:15 @tell beta 2011-06-20T09:19:15 mcstar: (tell ) -- Tells the whatever is. Use nested commands to your benefit here. 2011-06-20T09:19:24 hm 2011-06-20T09:19:30 @show beta 2011-06-20T09:19:31 mcstar: I have no idea what you mean. 2011-06-20T09:19:38 @show boobies 2011-06-20T09:19:39 Rabidus: Run as fast as you can and don't look back. 2011-06-20T09:19:44 :o 2011-06-20T09:20:00 Rabidus: do you know the command to bring up saved strings? 2011-06-20T09:20:16 saved strings? sorry no. :( 2011-06-20T09:20:23 @commands 2011-06-20T09:20:23 Rabidus: action, add, alert, alias, announce, announce add, announce list, announce remove, any, aol, apply, apropos, at, author, azn, ban add, ban list, ban remove, base, binary, bold, cache, calc, call, capabilities, capability add, capability list, capability remove, capability set, capability setdefault, capability unset, capitalize, change, changename, channel, channels, channelstats, chr, (5 more messages) 2011-06-20T09:20:27 amstan always does this 2011-06-20T09:20:28 @more 2011-06-20T09:20:28 Rabidus: cmd, coin, color, colorize, command, commands, concat, config, connect, contributors, convert, countargs, cpu, crypt, ctime, cut, cycle, decode, default, defaultcapability, defaultplugin, dehalfop, deop, devoice, dice, dict, dictionaries, disable, disconnect, dns, do, doctype, driver, echo, eightball, elapsed, enable, encode, errno, export, fetch, field, fight, fit, flush, forget, format, (4 more messages) 2011-06-20T09:20:35 @more 2011-06-20T09:20:35 Rabidus: fortune, game, get, ghost, gnu, google, grab, halfop, headers, hebrew, help, hexip, hexlify, hostmask, hostmask add, hostmask list, hostmask remove, icalc, identify, ignore, ignore add, ignore list, ignore remove, info, insert, invite, ircquote, jeffk, join, kban, key, kick, last, latency, learn, leet, len, levenshtein, limit, list, lithp, ln, load, lobotomy add, lobotomy list, lobotomy (3 more messages) 2011-06-20T09:21:03 @info beta 2011-06-20T09:21:03 mcstar: Error: The command "info" is available in the Factoids and RSS plugins. Please specify the plugin whose command you wish to call by using its name as a command before "info". 2011-06-20T09:21:15 http://ai-contest.com/ down 2011-06-20T09:21:28 yep, but it doesnt matter, theres a beta site 2011-06-20T09:21:28 old news Accoun 2011-06-20T09:21:31 which is saved in the variable beta 2011-06-20T09:21:45 @beta 2011-06-20T09:21:46 Rabidus: beta could be http://aichallengebeta.hypertriangle.com/. 2011-06-20T09:21:50 ;) 2011-06-20T09:21:51 hehe 2011-06-20T09:21:55 too simple 2011-06-20T09:21:55 thats it 2011-06-20T09:22:09 *** stocha has joined #aichallenge 2011-06-20T09:28:18 mcstar: isn't it you span food before your spawn ants ? 2011-06-20T09:29:06 wat's bfs again ? 2011-06-20T09:29:50 bfs = bio fuel system ? 2011-06-20T09:30:12 very good 2011-06-20T09:30:14 bfs vs A* I wish people would stop speaking with abreviations 2011-06-20T09:30:28 stocha: you have wikipedia, right? 2011-06-20T09:30:51 mcstar: yeah. But it doesn't do contextuel abreviation match up 2011-06-20T09:30:54 *** amstan has joined #aichallenge 2011-06-20T09:30:55 *** ChanServ sets mode: +o amstan 2011-06-20T09:31:06 well, belwo pathfinding you can find those 2011-06-20T09:31:25 *** carlop has joined #aichallenge 2011-06-20T09:31:31 and what does belwo stand for? 2011-06-20T09:31:38 blow ... 2011-06-20T09:31:44 hehe 2011-06-20T09:31:47 below 2011-06-20T09:31:51 http://en.wikipedia.org/wiki/Pathfinding no mention of any bfs that i can find 2011-06-20T09:31:59 stocha: i guess you know it is breadth first search ... 2011-06-20T09:32:35 there it is, on the right 2011-06-20T09:32:37 for a*, that is the only name i know for it 2011-06-20T09:32:44 and its written as A* 2011-06-20T09:32:51 i'm non native english speaker, AND i'm a computer more by doing than by learning :) 2011-06-20T09:32:56 A* is a variant of Dijkstra's algorithm 2011-06-20T09:33:16 a much cleverer variant though ... 2011-06-20T09:33:44 those arent my words, i just pasted as a proof for stocha 2011-06-20T09:33:51 who isnt using c anymore 2011-06-20T09:34:30 stocha: if you are computer, then i guess you would be a non-native english speaker :) 2011-06-20T09:35:17 A* i was pointed to a long time ago. Everyone bothering to talk in there seems to be willing to use it. Although i found absolutely no use of it in this game yet. 2011-06-20T09:35:52 i have a*, was a few times faster than bfs, usually 2011-06-20T09:35:56 Maybe as a late optimization of pathfinding. That's a maybe. 2011-06-20T09:36:04 though maybe i should have implemented it more efficiently 2011-06-20T09:36:13 im just think about pathfinding now 2011-06-20T09:36:21 http://ai-contest.com/ down 2011-06-20T09:36:26 unless you bot time out a lot because of it, i see no use for clever pathfinding. 2011-06-20T09:36:41 my algorith works as far as theres a clear view to the target :) 2011-06-20T09:36:44 i doubt pathfinding will be the bottleneck for me. 2011-06-20T09:36:52 Accoun: we just gave you the link to the beta site 2011-06-20T09:37:26 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T09:37:40 good bye stocha, was nice talking to you 2011-06-20T09:37:46 accoun is as annoying as he was half a year ago 2011-06-20T09:38:44 i beleive stoche will win this contest, i told him to use wikipedia :D 2011-06-20T09:39:39 meybe Accoun just wants to look at the old scoreboard 2011-06-20T09:39:55 or watch dead forum 2011-06-20T09:44:33 lol, on this metric, i beleive for pathfinding i only need to consider the overlap of specific rectangles 2011-06-20T09:45:20 the planned path is usually in viewradius anyway 2011-06-20T09:50:07 *** computerwiz_222 has joined #aichallenge 2011-06-20T09:52:18 can i ask about battle resolution rule? 'An ant lives if it has less enemies in range than each of those enemies have in range.' - can i replace the word 'each of' with 'every one of'? 2011-06-20T09:52:49 each and every 2011-06-20T09:52:53 i beleive 2011-06-20T09:53:02 ok, that is clear, thanks 2011-06-20T09:53:15 ask antimatroid he wrote that part 2011-06-20T09:53:38 and attackradius=5 means distance^2 <= 5 ? (it is the = in <= i am unsure about) 2011-06-20T09:53:51 <= 2011-06-20T09:54:03 ok, thanks 2011-06-20T09:54:24 does this rule seem to be fixed now? 2011-06-20T09:54:57 someone else should answer that, but it seems to me its gonna be final 2011-06-20T09:56:06 yes, i guess cannot say 100%, but if it is most likely final then i might go and write my battlecode .. thanks for the info. 2011-06-20T09:58:21 *** carlop has quit IRC (Ping timeout: 240 seconds) 2011-06-20T09:59:46 *** carlop has joined #aichallenge 2011-06-20T10:02:32 http://ai-contest.com/ down 2011-06-20T10:02:56 *** carlop has quit IRC (Read error: Connection reset by peer) 2011-06-20T10:02:58 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has quit IRC (Quit: Leaving.) 2011-06-20T10:05:24 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has joined #aichallenge 2011-06-20T10:09:40 Rabidus: is it a language barrier thing, or he does this on purpose? 2011-06-20T10:10:02 if i could say both, then i'd say both 2011-06-20T10:10:09 mut i prefer maybe more purpose thingie 2011-06-20T10:10:25 he acted very same few months ago in last contest 2011-06-20T10:10:42 and laughed at canadian coders :p 2011-06-20T10:10:48 dunno why 2011-06-20T10:11:13 Accoun: did you know that http://ai-contest.com/ is down? 2011-06-20T10:22:21 he must really like the contests though, as he keeps coming back ... 2011-06-20T10:22:38 yeah well, he was in top10 last year? 2011-06-20T10:22:44 6th or something 2011-06-20T10:24:52 if the site was up, i could check... :) i recall he finished close to me in the rankings... 2011-06-20T10:27:32 *** computerwiz_222 has quit IRC (Read error: Operation timed out) 2011-06-20T10:30:32 he did have a homepage 2011-06-20T10:30:34 let me google a bit 2011-06-20T10:31:17 ty google: http://webcache.googleusercontent.com/search?q=cache:LMM15p2ja0IJ:ai-contest.com/rankings.php+http://ai-contest.com/rankings.php&cd=1&hl=fi&ct=clnk&gl=fi&source=www.google.fi 2011-06-20T10:31:37 *** boegel has quit IRC (Ping timeout: 240 seconds) 2011-06-20T10:33:09 yes 6th - i recall watching the final ranking move after submissions had closed ... we kept swapping places ... 2011-06-20T10:34:16 hehe 2011-06-20T10:34:20 four points difference 2011-06-20T10:38:42 apart from the top 1or2, it was all very close ... 2011-06-20T10:39:19 Rabidus: was it you trying the tcp server yesterday? 2011-06-20T10:39:26 i heard it didnt work out? 2011-06-20T10:39:28 yes 2011-06-20T10:39:44 it worked ~3 games and then used all cpu and didn't work anymore 2011-06-20T10:41:33 ah well, thanks for giving it a go. i'll keep looking out for when one is up ... 2011-06-20T10:43:46 *** sigh has joined #aichallenge 2011-06-20T10:50:26 *** amstan has quit IRC (Read error: Operation timed out) 2011-06-20T10:50:41 *** sigh_ has joined #aichallenge 2011-06-20T10:51:14 *** sigh has quit IRC (Read error: Connection reset by peer) 2011-06-20T10:51:46 *** sigh_ has quit IRC (Client Quit) 2011-06-20T10:55:14 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has quit IRC (Remote host closed the connection) 2011-06-20T11:02:35 *** stocha has joined #aichallenge 2011-06-20T11:02:48 george1914: i wouldn't bet on the fighting system being final now. 2011-06-20T11:06:11 @later tell mcstar it's because i'm deconected that i don't read your replies 2011-06-20T11:06:11 stocha: I think that worked... 2011-06-20T11:07:32 *** dr- has quit IRC (Quit: WeeChat 0.3.0) 2011-06-20T11:08:43 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T11:10:04 *** Ice_Harley has quit IRC (Ping timeout: 252 seconds) 2011-06-20T11:14:27 stocha: ah - am coding now :/ 2011-06-20T11:15:55 stocha: btw - how do you view old chat (ie chat done whilst you were disconnected?) 2011-06-20T11:19:01 george1914: http://contestbot.hypertriangle.com/ 2011-06-20T11:19:36 you can also git clone it and grep for things :) 2011-06-20T11:20:23 /quit/quit 2011-06-20T11:20:25 *** smiley1983 has quit IRC (Quit: leaving) 2011-06-20T11:20:46 thanks smiley :) 2011-06-20T11:24:02 *** computerwiz_222 has joined #aichallenge 2011-06-20T11:25:57 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has joined #aichallenge 2011-06-20T11:27:34 *** _0x47 <_0x47!~slash@p57B3FE84.dip.t-dialin.net> has quit IRC (Client Quit) 2011-06-20T11:29:56 *** locutus2 has joined #aichallenge 2011-06-20T11:30:51 *** filter_ is now known as filter 2011-06-20T11:32:52 *** locutus2 has quit IRC (Read error: Connection reset by peer) 2011-06-20T11:33:01 *** amstan has joined #aichallenge 2011-06-20T11:33:01 *** ChanServ sets mode: +o amstan 2011-06-20T11:33:26 *** locutus2 has joined #aichallenge 2011-06-20T11:33:29 *** ohcrivens has joined #aichallenge 2011-06-20T11:34:22 *** olexs has quit IRC (Quit: Leaving.) 2011-06-20T11:38:05 *** computerwiz_222 has quit IRC (Remote host closed the connection) 2011-06-20T11:38:47 *** locutus2 has quit IRC (Read error: Connection reset by peer) 2011-06-20T11:39:03 *** locutus2 has joined #aichallenge 2011-06-20T11:41:23 *** MuTa-ofd00m has joined #aichallenge 2011-06-20T11:42:13 *** amstan has quit IRC (Quit: Konversation terminated!) 2011-06-20T11:43:38 *** amstan has joined #aichallenge 2011-06-20T11:43:38 *** ChanServ sets mode: +o amstan 2011-06-20T11:49:20 *** sigh has joined #aichallenge 2011-06-20T11:55:32 *** locutus2 has quit IRC (Read error: Connection reset by peer) 2011-06-20T11:56:06 *** locutus2 has joined #aichallenge 2011-06-20T11:58:00 *** nux67 has joined #aichallenge 2011-06-20T12:00:56 *** locutus2 has quit IRC (Read error: Connection reset by peer) 2011-06-20T12:07:11 *** sigh has quit IRC (Quit: This computer has gone to sleep) 2011-06-20T12:10:12 *** locutus2 has joined #aichallenge 2011-06-20T12:12:00 *** sigh has joined #aichallenge 2011-06-20T12:22:35 *** computerwiz_222 has joined #aichallenge 2011-06-20T12:22:54 *** computerwiz_222 has joined #aichallenge 2011-06-20T12:24:43 http://ai-contest.com/ down 2011-06-20T12:27:56 yes, it's known Accoun and will probably remain down for another 10 hours 2011-06-20T12:29:18 janzert: it is a bot, dont talk to it 2011-06-20T12:30:42 but we talk to contestbot 2011-06-20T12:30:49 contestbot: hello 2011-06-20T12:30:50 sigh: I'm sorry Dave, err sigh; I cannot 'hello'. 2011-06-20T12:30:55 :( 2011-06-20T12:31:15 you see? contestbot is intelligent, Accoun bot is not 2011-06-20T12:31:20 Accoun: hello 2011-06-20T12:31:37 yep, i thought so 2011-06-20T12:31:41 :P 2011-06-20T12:31:45 :P 2011-06-20T12:36:05 contestbot: hi 2011-06-20T12:36:06 amstan: User error -- Replace user. 2011-06-20T12:36:43 well. you know what.. i don't like you anymore 2011-06-20T12:37:03 janzert: i just hope jeff got the message 2011-06-20T12:38:40 *** computerwiz_222 has quit IRC (Read error: Connection reset by peer) 2011-06-20T12:39:04 *** computerwiz_222 has joined #aichallenge 2011-06-20T12:45:46 *** Meatkat has joined #aichallenge 2011-06-20T12:47:04 *** amstan has quit IRC (Quit: Konversation terminated!) 2011-06-20T12:54:48 *** computerwiz_222 has quit IRC (Ping timeout: 255 seconds) 2011-06-20T12:54:56 *** computerwiz_222 has joined #aichallenge 2011-06-20T12:55:18 *** ohcrivens has quit IRC (Ping timeout: 252 seconds) 2011-06-20T12:55:49 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-20T12:57:03 aichallenge: McLeopold epsilon * r4cf070d / sql/2_generate_matchup.sql : fix for seed issue - http://bit.ly/msAbOt 2011-06-20T12:57:10 *** amstan has joined #aichallenge 2011-06-20T12:57:11 *** ChanServ sets mode: +o amstan 2011-06-20T12:58:20 *** McLeopold has joined #aichallenge 2011-06-20T12:58:51 McLeopold: that fix seems to mean games in progress are now ignored for choosing a seed? 2011-06-20T12:59:03 seems the opposite of what we want 2011-06-20T12:59:09 no, not quite 2011-06-20T12:59:22 I'm glad you're here, cause we need to discuss it :) 2011-06-20T12:59:26 :) 2011-06-20T12:59:30 worker is positive if a game is in progress 2011-06-20T12:59:39 * amstan is listening if i'm needed 2011-06-20T12:59:44 worker is negative if the game crashed 2011-06-20T12:59:45 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110413222027]) 2011-06-20T13:00:24 still not seeing how the previous led to multiple games with the same seed? 2011-06-20T13:00:26 so, if a game is scheduled to worker 7, but it never reports back and just dies, then the matchup table stays positive, and the seed player never gets seeded again 2011-06-20T13:00:51 *** bhasker has joined #aichallenge 2011-06-20T13:01:09 tasks are never rescheduled after a certain time? 2011-06-20T13:01:09 too many seeds were excluded from being chosen because there are a ton of errored matchups 2011-06-20T13:01:41 tasks need to be rescheduled, that's what I want to discuss, what's the best way? 2011-06-20T13:02:00 we need a max time for matches to be reported 2011-06-20T13:02:10 * antimatroid might have to be terrible and miss the meeting later 2011-06-20T13:02:18 before if a task had been scheduled for more than x time it was resent as a task again 2011-06-20T13:02:54 the manager code can reset matchup worker_id's to null if the game hasn't been reported for awhile 2011-06-20T13:02:58 before it was something like 15 minutes, but of course the time out needs to be longer than the maximum time a game can take 2011-06-20T13:03:25 it only needs to be slightly shorted than the time it takes to rotate through the list of seeds maybe? 2011-06-20T13:04:06 I'd go for just longer than a game can take 2011-06-20T13:04:25 our theoretical max is 5 hours... 2011-06-20T13:04:30 ish 2011-06-20T13:04:55 if we were running serial right now right? 2011-06-20T13:04:56 :) 2011-06-20T13:04:59 but we're not 2011-06-20T13:05:14 also I wouldn't put it in the manager 2011-06-20T13:05:31 the manager should be doing everything needing to be scheduled 2011-06-20T13:05:36 *** computerwiz_222 has quit IRC (Quit: Konversation terminated!) 2011-06-20T13:05:47 *** computerwiz_222 has joined #aichallenge 2011-06-20T13:06:12 I'd just add it as part of the where clause in the task selection 2011-06-20T13:06:26 where worker_id is null or timestamp >= blah ... 2011-06-20T13:06:32 <= 2011-06-20T13:06:43 hmm 2011-06-20T13:06:55 yeah, then it could also update the worker_id 2011-06-20T13:07:04 compile tasks should be rescheduled in the same way 2011-06-20T13:07:06 so, another change might be needed 2011-06-20T13:07:29 I see they're a separate table 2011-06-20T13:07:29 did I hear that there will be a launch meeting? 2011-06-20T13:07:32 I'm sorting by game_id next for picking the seed, but game_id is based on when a game was reported, not started 2011-06-20T13:08:06 I'm thinking I need to change the logic to be a little more fair 2011-06-20T13:08:23 I don't think the difference will be significant 2011-06-20T13:08:26 Otherwise bots in games that take a long time get seeded less often over time 2011-06-20T13:09:11 also are they seeded by last game they played in or last game they were seed? 2011-06-20T13:09:19 If I keep the matchup records around, and store the game_id in them, I can do it 2011-06-20T13:09:24 last game they were seeded in 2011-06-20T13:09:40 ok, next question then is why? ;) 2011-06-20T13:09:55 why order by game_id? 2011-06-20T13:09:57 I vaguely remember discussing it but don't remember the reasoning 2011-06-20T13:10:07 why not order by last game they played in 2011-06-20T13:10:22 why does being the seed make a difference 2011-06-20T13:11:07 I can't think of a reason :) 2011-06-20T13:11:15 I'm pretty sure you made me do it :) 2011-06-20T13:11:22 heh, neither can I at the moment at least 2011-06-20T13:12:08 * janzert goes to find the wiki page on game scheduling 2011-06-20T13:12:26 janzert: i haven't found games take all that long with an agressiveish bot 2011-06-20T13:12:58 antimatroid: that's what I've noticed as well 2011-06-20T13:13:10 https://github.com/aichallenge/aichallenge/wiki/TrueSkill-Matchmaking/80c87a3eb941e2d5f1256c26994b5dd30fe88d42 2011-06-20T13:13:19 McLeopold: yep, looking at it now 2011-06-20T13:13:28 that's a past version, btww 2011-06-20T13:13:53 oh, was looking at most recent 2011-06-20T13:14:04 janzert: have you talked to McLeopold about the idea of "calling the game off early" ? 2011-06-20T13:14:08 most recent shows, ordering by last played 2011-06-20T13:14:26 antimatroid: not really, but I assume he saw the conversation since he filed an issue 2011-06-20T13:14:41 I never liked all the changes bss03 put in there. But I'll give him props for the trueskill formulas. 2011-06-20T13:15:02 * McLeopold waves at antimatroid 2011-06-20T13:15:14 :P 2011-06-20T13:15:29 https://github.com/aichallenge/aichallenge/wiki/TrueSkill-Matchmaking/c6b2a9882b686d84d40dd390f3488a6aa938982c 2011-06-20T13:15:53 pretty sure the future section of that version was mostly my thoughts on how it should go 2011-06-20T13:16:20 McLeopold: do you have many thoughts on how we're going to reasonably reduce max turns? 2011-06-20T13:16:37 antimatroid: yes, and I put them in the new issue 2011-06-20T13:16:48 * antimatroid goes to the issues 2011-06-20T13:16:54 actually maybe all I changed for that was saying seed should be based on last played 2011-06-20T13:17:20 with multiplay and a random number of players, it makes sense 2011-06-20T13:17:24 and that new submissions need to play with already established players 2011-06-20T13:17:37 Plus, with the way things have been going on beta, I think it would be good 2011-06-20T13:17:52 so from what I can tell I was always in favor of going with last time to play not last time as seed 2011-06-20T13:18:17 The trueskill matching *should* take care of that, but I'm not sure if anyone has analyzed how good it's been doing 2011-06-20T13:18:23 besides not being able to come up with a justification for the latter now ;) 2011-06-20T13:18:54 match quality tends to favor established submissions 2011-06-20T13:19:01 ok 2011-06-20T13:19:22 I can create an sql statment to pull the quality of past matches to check our numbers 2011-06-20T13:19:51 seeding is being done by user not submission right? 2011-06-20T13:19:57 * janzert goes to open the sql again 2011-06-20T13:20:31 looks like 2011-06-20T13:21:09 in which case the resubmission block should only be until the current submission finishes compiling 2011-06-20T13:21:19 the 10 minute delay isn't needed anymor 2011-06-20T13:22:28 why not? 2011-06-20T13:23:15 the delay was put in because new submissions used to immediately get a game 2011-06-20T13:23:17 I think it is still a good limitation for users, even if we don't need it for technical reasons. 2011-06-20T13:23:27 so by constantly resubmitting a user could get constant games 2011-06-20T13:23:39 that problem no longer exists 2011-06-20T13:24:03 *** Meatkat has quit IRC (Ping timeout: 255 seconds) 2011-06-20T13:24:16 but people love them some superstitions 2011-06-20T13:24:17 and antimatroid was complaining earlier because he discovered a bug just after submitting but couldn't submit the fixed version because of the delay 2011-06-20T13:24:37 i wasn't so much complaining as i said it was annoying :P 2011-06-20T13:24:40 in that case, his old submission would still get games 2011-06-20T13:24:48 although looking at the the code it seems like the current delay is 1 minute? 2011-06-20T13:25:00 line 69 of submission.php 2011-06-20T13:25:01 yeah, for beta :) 2011-06-20T13:25:14 60 seconds still got to antimatroid? 2011-06-20T13:25:23 it's 10 minutes? 2011-06-20T13:25:24 hmm, well anyone should be able to live with a minute delay :P 2011-06-20T13:25:34 it certainly took longer than a minute 2011-06-20T13:25:49 it's a good anti-spam measure anyway 2011-06-20T13:26:09 yeah, a 1 minute delay seems more like an anti-spam measure 2011-06-20T13:26:17 antimatroid: had it not finished compiling maybe? 2011-06-20T13:26:29 janzert: probably it 2011-06-20T13:26:43 oh, compiling on beta means waiting for the current game to finish 2011-06-20T13:26:45 you can't resubmit until the current compilation finishes 2011-06-20T13:26:53 ahh, yeah 2011-06-20T13:27:28 another possibility is that the worker was down for a while this morning 2011-06-20T13:27:31 perhaps a delete submission feature? 2011-06-20T13:28:08 actually the worker was down just now 2011-06-20T13:28:23 had experience a "no route to host" error 2011-06-20T13:28:39 it would be cool if we gave each user a git repository and made them push changes, then compiling would happen from there 2011-06-20T13:28:48 beta server network connection has been flaky the past few days 2011-06-20T13:29:03 me ssh session seems fine 2011-06-20T13:29:33 McLeopold: it was 3 hours ago or so 2011-06-20T13:29:39 that the error happened 2011-06-20T13:29:55 actually just over 2.5 2011-06-20T13:30:32 also any idea what may have changed on the beta server that is causing all sorts of timeouts if we try and run the worker there? 2011-06-20T13:31:04 not sure, can I try? 2011-06-20T13:31:08 go ahead 2011-06-20T13:31:55 it seems to have started around the time we switched battle method, but I tried switching it back and it was still timing out almost every bot in a game 2011-06-20T13:32:01 looks like api_get_tasks is not returning 2011-06-20T13:32:24 it must be eventually at least 2011-06-20T13:32:30 the ec2 worker is getting tasks 2011-06-20T13:32:35 it took entirely too long 2011-06-20T13:32:47 hehe ^ "it took entirely too long" 2011-06-20T13:32:52 but the timeouts I meant were players in games 2011-06-20T13:33:06 almost every game was ending with all but one of the bots having timed out 2011-06-20T13:33:26 I did notice mysql seems to be using more CPU than I remembered 2011-06-20T13:33:46 I just ran a game with no timeouts... 2011-06-20T13:33:52 * antimatroid doesn't like the ghost thing 2011-06-20T13:33:59 great 2011-06-20T13:34:00 it's gone once you go above it? and just taunts you otherwise 2011-06-20T13:34:07 maybe it resolved itself 2011-06-20T13:34:29 oh wait...2 timeouts 2011-06-20T13:34:42 could be java 2011-06-20T13:36:03 I'm upping loadtime to 10 seconds just to see what happens 2011-06-20T13:36:29 i've found a use for my start up time :) 2011-06-20T13:36:34 I don't think it was always turn 0 although it seems the majority were 2011-06-20T13:37:17 want me to stop the ec2 worker temporarily 2011-06-20T13:37:24 nah 2011-06-20T13:37:33 1 java vm crash, 1 timeout 2011-06-20T13:38:00 28716? 2011-06-20T13:40:35 I think we could only really know something is up depending on the bot 2011-06-20T13:40:47 For instance, starter bots shouldn't timeout 2011-06-20T13:40:50 28716 is 2 timeouts btw 2011-06-20T13:41:32 28717 is ec2 worker 2011-06-20T13:42:06 hunterbot timed out the other day from memory? 2011-06-20T13:42:14 something normal timed out 2011-06-20T13:43:17 amstan: whatever happened to adding some of those images to the site? they were pretty sweet 2011-06-20T13:43:27 computerwiz_222: you're the design guy yeah? 2011-06-20T13:43:30 antimatroid: i need ideas 2011-06-20T13:43:41 antimatroid: yep 2011-06-20T13:43:50 antimatroid: what's up? 2011-06-20T13:44:05 computerwiz_222: I'm not sold on the shading around "Google AI Challenge" and I think i 2011-06-20T13:44:24 d prefer the "sign in|sign up" stuff on the right hand side part at the top of the list 2011-06-20T13:45:07 antimatroid: the idea to have it on the left is to place emphasis on it 2011-06-20T13:45:33 aichallenge: McLeopold epsilon * re827660 / (sql/0_schema.sql sql/2_generate_matchup.sql website/sql.php): added matchup timestamps - http://bit.ly/iMf9cZ 2011-06-20T13:45:38 antimatroid: the shading... i like it - i think it's a personal preference thing. you are free to try it without, just change the css 2011-06-20T13:45:44 29719 is from main server worker and has 2 timeouts one of them being csharp_starter 2011-06-20T13:45:54 err, 28719 that is 2011-06-20T13:46:04 personally i think an emphasized link in any intro parts is good, but slightly annoying if always in your face :P 2011-06-20T13:46:09 not sure what anyone else thinks? 2011-06-20T13:46:29 yeah, i just followed what most large sites do 2011-06-20T13:46:30 antimatroid: I agree 2011-06-20T13:46:58 antimatroid: take facebook: they present every non-logged in user with an entire form to sign up 2011-06-20T13:47:23 antimatroid: maybe we could make it conditional - once you are logged in, it moves into the navbar on the right 2011-06-20T13:47:29 computerwiz_222: did you just use facebook as an example for design? :P 2011-06-20T13:47:39 antimatroid: oh no, but it seems fitting 2011-06-20T13:47:44 gmail does this too 2011-06-20T13:47:55 if you're not signed in, it is easy to locate the register button 2011-06-20T13:48:46 what if "Sign Up" was in bold on the right? 2011-06-20T13:48:52 still more clearly outlined 2011-06-20T13:48:54 but looks cleaner 2011-06-20T13:49:15 we can try it - the decision to put it there was to fill that space 2011-06-20T13:49:28 :) 2011-06-20T13:49:28 janzert: 34,997 errored matchups. That might be why get_task was taking so long... 2011-06-20T13:49:30 computerwiz_222: hi, can i make a suggestion to? its bugging me for some time 2011-06-20T13:49:39 McLeopold: heh, ouch 2011-06-20T13:49:40 my suggestion is to have it say sign up if you aren't signed in, and then once you login your username (and profile link are in the navbar) 2011-06-20T13:50:07 amstan: for images, if we could splash them on the left and right down the specs page, that'd look good 2011-06-20T13:50:11 not really sure otherwise 2011-06-20T13:50:12 McLeopold: how does a matchup get flagged as errored? or at least where should I look to see that 2011-06-20T13:50:25 maybe add one to the background image where the brown is? but not sure that'd look rihgt 2011-06-20T13:50:28 antimatroid: janzert, McLeopold and computerwiz_222 don't seem to like them 2011-06-20T13:50:36 I deleted them, otherwize the matchup table has the info 2011-06-20T13:51:18 mcstar: it's bugging me that I don't know your suggestion yet, spit it out! 2011-06-20T13:51:29 computerwiz_222: i think that green theme is really bad, it doesnt go with the rust colored sides/background, and its hard on my eyes 2011-06-20T13:51:44 antimatroid: sorry, i was waiting on the confirmation of computerwiz_222 2011-06-20T13:52:02 i dont want to tire him with silly requests 2011-06-20T13:52:09 mcstar: objection overruled :p 2011-06-20T13:52:43 hmm, another stderr spewing bot... 2011-06-20T13:52:44 mcstar: we had about 4 designers who claimed they would make a new design, the only one that went through with it was computerwiz_222 2011-06-20T13:52:49 McLeopold: I mean what code is actually making the change to flag a matchup as errored or why? 2011-06-20T13:52:51 mcstar: it's either this design, or the old one 2011-06-20T13:53:04 yeah.. everyone likes to jump on the designers lol 2011-06-20T13:53:16 like.. i want it YELLOW! 2011-06-20T13:53:21 computerwiz_222: <3 sorry if it seems like hate, we're just good at pointing out what bugs us 2011-06-20T13:53:22 janzert: api_game_result, if the engine detects it crashed without finishing the game 2011-06-20T13:53:27 overall I do think it looks good 2011-06-20T13:53:28 antimatroid: yeah yeah i understand 2011-06-20T13:53:30 ahh, ok 2011-06-20T13:53:32 antimatroid: i'm being a baby :P 2011-06-20T13:53:44 webdesigns by cometee usually end bad: http://theoatmeal.com/comics/design_hell 2011-06-20T13:53:48 I committed a mistake and let the server run overnight :( 2011-06-20T13:53:49 one interesting thing: chromium breaks ranking to two lines for me 2011-06-20T13:53:55 firefox shows it on one line 2011-06-20T13:54:18 amstan: yes, that is exactly what happns 2011-06-20T13:54:21 ah, so hopefully it's resolved 2011-06-20T13:54:28 but what is really interesing: if i zoom out in chromium, the break remains, but if i zoom in, it will stay in one line 2011-06-20T13:54:46 computerwiz_222: can you add a live feed if things my cat tweets on the internet? 2011-06-20T13:55:02 McLeopold: huh?? 2011-06-20T13:55:05 of* 2011-06-20T13:55:14 McLeopold: lol 2011-06-20T13:55:21 mcstar: and by zoom you mean ctrl+ ? 2011-06-20T13:55:21 lol 2011-06-20T13:55:27 mcstar: works fine for me 2011-06-20T13:55:29 ctrl scroll 2011-06-20T13:55:43 all I can say is that I'm sorry you don't like it -- if you can get a few other people and a formal vote is called we can change it no problem 2011-06-20T13:55:49 mcstar: i can't reproduce it 2011-06-20T13:55:53 it is very hard to fulfill everyone's desires 2011-06-20T13:55:56 mcstar: rankings.php right? 2011-06-20T13:56:03 yes 2011-06-20T13:56:12 if i change it to blue then everyone will complain that they don't like green lol 2011-06-20T13:56:16 *** Apophis has joined #aichallenge 2011-06-20T13:56:19 s/green/blue/g 2011-06-20T13:56:51 amstan: the last character, the arrow or - will be on another line, right under skill 2011-06-20T13:57:11 * antimatroid wonders what it would look like if the top and bottom green were the same as the rhs green :P but I have no css skills 2011-06-20T13:57:25 mcstar: what version of chrome? 2011-06-20T13:57:35 latest 2011-06-20T13:57:46 12.0.742.91 (0) 2011-06-20T13:57:48 *** Apophis_ has quit IRC (Ping timeout: 255 seconds) 2011-06-20T13:58:48 one explanation can be that my font sizes might not be the default, but nevertheless, i would expect it to not break the lines 2011-06-20T13:59:04 mcstar: i have s/91/100/g 2011-06-20T13:59:08 mcstar: works fine for me 2011-06-20T13:59:37 that small change must not matter 2011-06-20T13:59:59 i have 14pt default fonts 2011-06-20T14:00:33 mcstar: does the ubuntu font load properly? 2011-06-20T14:00:39 yes 2011-06-20T14:00:44 nvm, that wouldn't matter 2011-06-20T14:00:44 mcstar: I can put a nowrap css style in 2011-06-20T14:00:47 it's only for the headers 2011-06-20T14:01:07 no, the whole page is in ubuntu isnt it? 2011-06-20T14:01:10 lok closely 2011-06-20T14:01:13 look 2011-06-20T14:01:20 McLeopold: that would be good i think 2011-06-20T14:01:32 McLeopold: every game from worker 3 seems to be having multiple timeouts 2011-06-20T14:01:40 http://aichallengebeta.hypertriangle.com/visualizer.php?game=28723 2011-06-20T14:01:44 mcstar: what os? 2011-06-20T14:01:51 Arch 2011-06-20T14:01:55 I don't think FlagCapper has been timing out at all 2011-06-20T14:02:11 amstan: windows 99 :D 2011-06-20T14:02:17 my default browser fonst are the dejavu sans lgc family 2011-06-20T14:02:34 but for me the whole page is in ubuntu 2011-06-20T14:02:37 mcstar: i'm installing google chrome right now 2011-06-20T14:02:44 janzert: okay, not sure what the issues is, other than to say it's the server somehow 2011-06-20T14:03:23 mcstar: oh.. right.. the whole page is ubuntu 2011-06-20T14:03:25 ok, I wonder if it's related to the network issues it's been having too 2011-06-20T14:03:29 :D 2011-06-20T14:03:48 now you must pay licence 2011-06-20T14:03:54 mcstar: it's free 2011-06-20T14:04:00 janzert: if it is a vm, maybe the vm itself is getting starved? 2011-06-20T14:04:11 * mcstar always jokes when he is not serious 2011-06-20T14:04:14 mcstar: that font is free 2011-06-20T14:04:17 certainly a possiblity 2011-06-20T14:04:30 it's like Myriad Pro but for the open source community :P 2011-06-20T14:04:51 computerwiz_222: do you have to wear a turtle neck while designing with myriad pro? 2011-06-20T14:05:06 computerwiz_222: do you know about freetype2-infinality? 2011-06-20T14:05:21 amstan: always, i wear two! 2011-06-20T14:05:32 mcstar: no, never heard of it 2011-06-20T14:05:54 well, for the moment for linux its like the best font renderer 2011-06-20T14:06:17 i just installed it a couple of days ago, and my gtk fonts look beautiful 2011-06-20T14:06:26 as nice as qt fonts 2011-06-20T14:06:38 i just installed google chrome to find out that the bug you suggest is not reproduceable :P 2011-06-20T14:06:39 (of course i mean fonts in appőlications) 2011-06-20T14:08:42 *** Dlayne has joined #aichallenge 2011-06-20T14:10:43 computerwiz_222: if you did, now you can use chromium for all tasks, im glad you installed it 2011-06-20T14:10:54 but tell me the dedault font sizes 2011-06-20T14:10:57 f 2011-06-20T14:11:13 mcstar: he's anti-chrome for some reason 2011-06-20T14:11:13 mcstar: i see what you did there 2011-06-20T14:11:18 even though it has less buttons than firefox 2011-06-20T14:11:22 :) 2011-06-20T14:11:42 i'm not anti-chrome, i just need more reasons to use it (but i'm not asking, please refrain) 2011-06-20T14:11:53 also.. 2011-06-20T14:12:00 we just started a browser war 2011-06-20T14:12:02 windows is best 2011-06-20T14:12:04 no linux! 2011-06-20T14:12:08 pfff 2011-06-20T14:12:11 lol 2011-06-20T14:12:14 i dont even wqnt to talk to you 2011-06-20T14:12:22 i'm a ubuntu user 2011-06-20T14:12:23 no worries 2011-06-20T14:12:31 now tell me the font sizes from chromium pls 2011-06-20T14:12:42 i cant remmeber if it 16 or 15 pt 2011-06-20T14:12:51 mcstar: where do i find this? 2011-06-20T14:12:59 amstan: if he knew, he would go there himself 2011-06-20T14:13:08 hehe 2011-06-20T14:13:11 preferences 2011-06-20T14:13:14 mcstar: it says medium.. 2011-06-20T14:13:18 mac os x > windows > linux 2011-06-20T14:13:19 under the hood 2011-06-20T14:13:25 webcontent 2011-06-20T14:13:32 customize fonts 2011-06-20T14:13:40 oh boy, here we go 2011-06-20T14:13:52 ;) 2011-06-20T14:14:00 just trollin' 2011-06-20T14:14:02 lol 2011-06-20T14:14:03 yeah, linux i ahead of all 2011-06-20T14:14:11 top of the queue 2011-06-20T14:14:13 yep, comes out on the right 2011-06-20T14:14:30 mcstar: looks like 16 2011-06-20T14:14:35 thx 2011-06-20T14:14:56 and then if it's FILO, we can pop OS X and then windows shortly after and then the program crashes because windows was there and linux rules them all 2011-06-20T14:15:55 also trollin' 2011-06-20T14:16:04 :) 2011-06-20T14:16:16 i realized long ago when i left windows, that there are different needs, no point in bashing any of them 2011-06-20T14:16:23 mcstar: agreed 2011-06-20T14:16:52 mcstar: i probably wouldn't handle a business payrole on linux because there is no one around to update the tax rates and so on 2011-06-20T14:17:03 hey don't agree that. we want to argue! 2011-06-20T14:17:09 Rabidus: how do max and bsd compare? 2011-06-20T14:17:23 McLeopold: thanks :D 2011-06-20T14:17:40 wtf is max 2011-06-20T14:17:41 :D 2011-06-20T14:18:03 mac, a type 2011-06-20T14:18:05 0 2011-06-20T14:18:16 http://en.wikipedia.org/wiki/MAX_%28Linux_distribution%29 2011-06-20T14:18:18 * McLeopold bangs head on keyboard 2011-06-20T14:18:20 lol 2011-06-20T14:18:24 :D 2011-06-20T14:18:52 ok, back to the ants 2011-06-20T14:19:11 vi vs emacs? anyone? 2011-06-20T14:19:14 m 2011-06-20T14:19:19 nano! 2011-06-20T14:19:20 vi 2011-06-20T14:19:21 yay 2011-06-20T14:19:31 well, i use them both :D 2011-06-20T14:19:35 notepad vs wordpad? anyone? 2011-06-20T14:19:37 :D 2011-06-20T14:19:40 but i use nano for most config files for some reason 2011-06-20T14:19:45 Rabidus: that is just silly 2011-06-20T14:19:46 force of habit 2011-06-20T14:19:49 yes. :D 2011-06-20T14:20:04 bash vs tcsh? 2011-06-20T14:20:10 bash 2011-06-20T14:20:12 lol 2011-06-20T14:20:20 i will form an opinion on most things :P 2011-06-20T14:20:36 i said it earlier, i realized today that theres not going to be any bot written in bash :( 2011-06-20T14:20:51 mcstar: why not? 2011-06-20T14:20:55 mcstar: we can do bash just fine 2011-06-20T14:21:00 Rabidus: notepad++ 2011-06-20T14:21:07 bash spawns processes by the tousands 2011-06-20T14:21:15 hehe 2011-06-20T14:21:23 never used again after ftp plugin trashed somehow one of my files 2011-06-20T14:21:44 amstan: consider that even a simple $() spawns a new process 2011-06-20T14:22:11 and then comes any utility u use, sed/awk/bc/echo/cat 2011-06-20T14:22:13 mcstar: is there really no other way? 2011-06-20T14:22:41 can't you just use simple string concats? 2011-06-20T14:22:59 well, if we very very restrict ourselfes, it might work 2011-06-20T14:23:12 but i guess it would make the code awkward 2011-06-20T14:23:23 and bash's main idea is to interface commands 2011-06-20T14:23:52 yes, but that's not allowed 2011-06-20T14:23:54 muhahah 2011-06-20T14:24:13 oh, another flame: gnome/kde/awesome 2011-06-20T14:24:28 you guessed right, im using the last one 2011-06-20T14:24:59 what is awesome? 2011-06-20T14:25:02 lol 2011-06-20T14:25:28 one of the best tiling window managers 2011-06-20T14:25:39 oh, it's tiling? 2011-06-20T14:25:40 very ncie 2011-06-20T14:25:43 mcstar: how does it compare to xmonad? 2011-06-20T14:25:49 ok, since everyone seems to be trying to stir things up here, i am going to contribute by posing the question ... 2011-06-20T14:25:51 what about ants? 2011-06-20T14:26:00 Rabidus: what about ants? 2011-06-20T14:26:13 what is the (estimated) timeline for the contest? 2011-06-20T14:26:15 ants you say... 2011-06-20T14:26:15 umm.. some talking about it :) 2011-06-20T14:26:27 george1914: idk, we're having a meeting tonight to discuss that 2011-06-20T14:26:32 amstan: what time? 2011-06-20T14:26:34 i was about to use xmonad, but it was quite a change to switch to tiling wms, and learning at the same time lisp and haskell seemed hard 2011-06-20T14:26:38 computerwiz_222: see topic 2011-06-20T14:26:44 george1914: when I fix stuff, which is taking awhile because I'm busy with life 2011-06-20T14:26:49 so i went with awesome + lua 2011-06-20T14:27:07 ew, lua 2011-06-20T14:27:38 yeah, i know, i dont want to know it, but for writing a config file its pretty straightforward 2011-06-20T14:27:40 I thought lua was cool? 2011-06-20T14:27:55 lol 2011-06-20T14:27:56 you only have to mod. already existing ones 2011-06-20T14:28:32 McLeopold: the only advantage it has over python is the sandboxing 2011-06-20T14:29:06 whoa, Pontifical Catholic University... 2011-06-20T14:29:54 McLeopold: what? 2011-06-20T14:30:06 I'm reading the lua website 2011-06-20T14:30:23 that sentence was hard to read, i read whoa in an all excited way but then calmed way down about the University... lol 2011-06-20T14:31:00 i didnt like lua, cause it comes from portugese, and it a usually warm country, i might eat their dishes, but i dont think i want to use their software 2011-06-20T14:31:10 mcstar: lol 2011-06-20T14:31:33 It comes from Brazil, not portugal 2011-06-20T14:31:58 well the name is prtugese 2011-06-20T14:32:09 means the moon 2011-06-20T14:32:12 ok, thanks for that - i'll keep looking in. very good of you folks to work on this .... - if/when you have any guesses it'd be great to know - i saw dates somewhere like september, and also winter? - so i have no idea at all when things might begin officially - the anticipation/excitement might have killed me by winter you see ... .) 2011-06-20T14:32:23 but the same argument holds 2011-06-20T14:32:46 george1914: yes, i was hoping we could have released earlier, but that didn't happen 2011-06-20T14:32:55 oh, im not racist, but i have stereotypes 2011-06-20T14:33:29 If your stereotype is about a people group, then that means your rasist. 2011-06-20T14:33:58 what if i hate stupid people? am i racist? 2011-06-20T14:34:18 I would say no, because any people group can contain stupid people. 2011-06-20T14:34:19 McLeopold: how can you not have a stereotype about a group of people? 2011-06-20T14:34:24 stupid people aren't a stereotype, that's a trait 2011-06-20T14:34:29 amstan: very hard not to 2011-06-20T14:34:31 being stupid isn't a stereotype* 2011-06-20T14:34:38 god that sentence was rough 2011-06-20T14:34:45 :) 2011-06-20T14:34:49 also.. why are we talking about grammar? 2011-06-20T14:35:01 because i said so. don't poke the bear 2011-06-20T14:35:15 hell, grammar, have any of you heard of the Voynich manuscript? 2011-06-20T14:35:20 computerwiz_222: say @roulette 2011-06-20T14:35:27 @roulette 2011-06-20T14:35:27 computerwiz_222: *click* 2011-06-20T14:35:35 computerwiz_222: do it a few times 2011-06-20T14:35:36 ohhh 2011-06-20T14:35:39 @roulette 2011-06-20T14:35:39 computerwiz_222: *click* 2011-06-20T14:35:41 @roulette 2011-06-20T14:35:41 *** computerwiz_222 was kicked by contestbot (BANG!) 2011-06-20T14:35:42 * contestbot reloads and spins the chambers. 2011-06-20T14:35:55 @roulette 2011-06-20T14:35:55 *** antimatroid was kicked by contestbot (BANG!) 2011-06-20T14:35:56 * contestbot reloads and spins the chambers. 2011-06-20T14:35:58 *** computerwiz_222 has joined #aichallenge 2011-06-20T14:36:00 hehe 2011-06-20T14:36:02 well that was fun 2011-06-20T14:36:11 welcome back to the living 2011-06-20T14:36:11 everyone turned into lemmings now 2011-06-20T14:36:20 *** antimatroid has joined #aichallenge 2011-06-20T14:36:25 is that random? 2011-06-20T14:36:36 or is it 3 times and then i'm gone 2011-06-20T14:36:38 @roulette 2011-06-20T14:36:38 antimatroid: *click* 2011-06-20T14:36:41 @roulette 2011-06-20T14:36:41 antimatroid: *click* 2011-06-20T14:36:43 hehe, decide for yourself after a couple of hundred runs 2011-06-20T14:36:44 :P 2011-06-20T14:36:50 lol 2011-06-20T14:37:20 so, no Voynich? 2011-06-20T14:37:32 its really interesting though... 2011-06-20T14:37:37 nope 2011-06-20T14:37:53 *** Palmik has quit IRC (Remote host closed the connection) 2011-06-20T14:37:57 no dice 2011-06-20T14:37:58 its a manuscript written in an unknown languge made probably in the 14th century 2011-06-20T14:38:17 if the language is unknown, how do you know it's a language? 2011-06-20T14:38:25 it could just be an interesting series of paintings :P 2011-06-20T14:38:28 good question 2011-06-20T14:38:41 from soe statistical methods, it resembles a language 2011-06-20T14:38:43 some 2011-06-20T14:38:48 neat 2011-06-20T14:39:12 and there are some peculiar drawings in the book too, very interesting 2011-06-20T14:39:30 You all need to do something constructive. Free T-Shirt to the person who solves this. https://github.com/aichallenge/aichallenge/issues/191 2011-06-20T14:39:45 McLeopold: what kind of tshirt? 2011-06-20T14:39:58 amstan: you are supposed to be on my side! 2011-06-20T14:40:08 just put it in a mathematical theorem prover 2011-06-20T14:40:19 McLeopold: i am, but i have to know how much enthusiasm to dedicate 2011-06-20T14:40:43 amstan: you need to get the others to have enthusiasm, not yourself 2011-06-20T14:41:00 ok everyone! free 3 wolf moon tshirt! 2011-06-20T14:41:08 :D 2011-06-20T14:41:08 It will have a hole for you head, your torso, and 2 holes for your arms 2011-06-20T14:41:38 what about my nipples? 2011-06-20T14:41:47 no holes for nipples 2011-06-20T14:41:52 whoever makes that 3 wolf moon tshirt must love that meme 2011-06-20T14:41:53 McLeopold: what about the holes in between atoms? 2011-06-20T14:41:56 then no enthusiasm 2011-06-20T14:41:57 or whatever you would call it 2011-06-20T14:41:58 enough space for my manboobs? 2011-06-20T14:42:25 hehe 2011-06-20T14:42:30 Rabidus: you will get a salad instead of a t-shirt 2011-06-20T14:42:35 heheh 2011-06-20T14:42:40 what kind of salad? 2011-06-20T14:42:44 lettus 2011-06-20T14:42:47 meat-salad 2011-06-20T14:43:02 McLeopold: where are you from? 2011-06-20T14:43:02 hmmh 2011-06-20T14:43:09 its lettuce 2011-06-20T14:43:19 you're a lettuce 2011-06-20T14:43:26 http://www.cafepress.com/ZeroViz.128448216 2011-06-20T14:43:33 lettuce start finding a solution 2011-06-20T14:43:59 i want nipple holes! 2011-06-20T14:44:13 i mean.. in shirt... holes for my nipples 2011-06-20T14:44:16 JSkills means java right? 2011-06-20T14:44:19 Rabidus: you may cut them yourself! 2011-06-20T14:44:33 Rabidus: there's this tool, made out of metal, it has 2 moving parts, and sharp blades, you can use it to cut things 2011-06-20T14:44:33 then they are not original 2011-06-20T14:44:40 We can discuss the mutilation of merchandise after the trueskill has been fixed. I will be abliging 2011-06-20T14:45:11 McLeopold: what was wrong with the python implementation again? 2011-06-20T14:45:20 was it not agreeing with the microsoft online one? 2011-06-20T14:46:20 yes, that was the issue 2011-06-20T14:46:26 jskill was much more accurate 2011-06-20T14:46:29 did we bugreport it? 2011-06-20T14:46:34 no 2011-06-20T14:46:36 *** Redgis has joined #aichallenge 2011-06-20T14:46:41 shouldn't we? 2011-06-20T14:47:03 If you want the t-shirt, then yes 2011-06-20T14:47:20 *** choas has joined #aichallenge 2011-06-20T14:47:23 McLeopold: if you can give me the test data, i'd gladly do it 2011-06-20T14:47:36 you have access to the server? 2011-06-20T14:47:38 100kB?? thats long 2011-06-20T14:47:44 yes, but i'm not sure what to look for 2011-06-20T14:47:45 and its compressed 2011-06-20T14:47:57 it's in the issue 2011-06-20T14:48:55 McLeopold: which one? the one you linked has 2 lines of text 2011-06-20T14:49:08 amstan: is your avatar a chevron from an SGU stargate? 2011-06-20T14:49:14 mcstar: yes 2011-06-20T14:54:17 McLeopold: i'll talk to you when i get home 2011-06-20T14:54:50 *** amstan has quit IRC (Quit: Konversation terminated!) 2011-06-20T14:58:09 computerwiz_222: how about this: http://imgur.com/BQiL7 2011-06-20T14:58:27 *** computerwiz_222 has quit IRC (Ping timeout: 250 seconds) 2011-06-20T14:58:27 eww 2011-06-20T14:58:34 that's disgusting 2011-06-20T14:58:45 are you color blind? 2011-06-20T14:58:49 pretty strange color combos idd 2011-06-20T14:58:50 no 2011-06-20T14:59:00 i find it appealing 2011-06-20T14:59:02 :) 2011-06-20T14:59:18 i just changed the hue and contrast of the original 2011-06-20T15:01:35 you know, this game is full of colors, all the different players have diff. colors, i thought the site could use them as well 2011-06-20T15:01:40 and i like pastel colors 2011-06-20T15:01:58 they are nice in the terminal too 2011-06-20T15:02:19 I'm being honest, and sorry if I offend, but it really does look really bad. 2011-06-20T15:02:41 i can uderstand that 2011-06-20T15:05:09 McLeopold: those are my terminal colors what do you think about them? http://imgur.com/jQHgv 2011-06-20T15:11:38 I think a terminal color scheme is a lot different than a website colorscheme 2011-06-20T15:12:14 im just asking your personal opinion about my colors 2011-06-20T15:12:43 I don't think I would have choosen the purple or pink 2011-06-20T15:13:17 hmm... 2011-06-20T15:13:18 yeah, i love them 2011-06-20T15:13:22 actually, I would have 2011-06-20T15:14:05 I like this: http://ethanschoonover.com/solarized, which is similar to yours, but not quite 2011-06-20T15:16:03 not bad, but i like a black background, and a bit stronger colors for the terminal 2011-06-20T15:16:26 *** sigh has quit IRC (Quit: This computer has gone to sleep) 2011-06-20T15:16:53 *** Accoun has quit IRC () 2011-06-20T15:20:23 McLeopold: it seems it is still taking 15-30 seconds to get the next task from the server 2011-06-20T15:20:33 yeah, I'm looking at it 2011-06-20T15:20:36 ok 2011-06-20T15:29:17 aichallenge: McLeopold epsilon * r0305b6b / (4 files in 3 dirs): update to get_task to assign matchup to alternate worker after 20 minutes with no result - http://bit.ly/lii4MX 2011-06-20T15:33:30 *** Accout has joined #aichallenge 2011-06-20T15:33:42 McLeopold: I assume you have no problem changing the name of the current battle resolution method from power to occupied? (Since I think you might have been the one to come up with it) 2011-06-20T15:34:11 Yes, I think overall it may be more descriptive. Maybe "Less occupied" is better. 2011-06-20T15:34:30 Perhaps we need a really good description first, then we can come up with a name 2011-06-20T15:34:49 yeah, except we need one word and 'occupied' seems better than 'less' ;) 2011-06-20T15:35:10 idle? 2011-06-20T15:35:25 hmm, I think I like occupied better still 2011-06-20T15:35:30 me too 2011-06-20T15:35:36 power could be anything 2011-06-20T15:36:06 I'm going to add the name occupied, basically as an alias of power 2011-06-20T15:36:12 to ants.py 2011-06-20T15:36:17 ok 2011-06-20T15:36:18 and update playgame.py 2011-06-20T15:36:27 outnumbered? would be good except that == kills you also 2011-06-20T15:39:53 aichallenge: janzert epsilon * r58b0b3b / (ants/ants.py ants/playgame.py): Make occupied an alias for power attack method and set it as the default - http://bit.ly/kg4I4v 2011-06-20T15:46:38 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-20T15:46:46 if ants from opposing teams are both close enough to spawn a food what happens? I think the food is simply destroyed 2011-06-20T15:47:07 yes 2011-06-20T15:49:42 <_flag> Is it possible to download bot input for games where I crash/timeout? 2011-06-20T15:50:23 _flag: we are not saving that info 2011-06-20T15:50:25 _flag: just the game replay, which should be convertible back to the bots input 2011-06-20T15:50:39 the plan is to get a game result converter that will create that input for you 2011-06-20T15:50:42 unfortunately no one has written a conversion utility that I know of yet 2011-06-20T15:51:04 <_flag> Okay, thanks 2011-06-20T15:52:10 <_flag> Has the schedualing bug been fixed? 2011-06-20T15:52:14 *** astevenson has joined #aichallenge 2011-06-20T15:53:38 antimatroid: where is your mathy explaination of the occupied method? 2011-06-20T15:54:39 janzert: can you turn off the worker for awhile? 2011-06-20T15:54:43 *** amstan has joined #aichallenge 2011-06-20T15:54:43 *** ChanServ sets mode: +o amstan 2011-06-20T15:54:49 yo 2011-06-20T15:54:53 sure 2011-06-20T15:55:14 just a short time or an hour or more? 2011-06-20T15:55:23 it's stopped now 2011-06-20T15:55:25 about 1 hour 2011-06-20T15:55:50 *** sigh has joined #aichallenge 2011-06-20T15:55:53 ok, I'll go ahead and shut it down altogether and bring up a fresh one later then 2011-06-20T16:02:20 janzert: here's my attempt at an explaination https://github.com/aichallenge/aichallenge/wiki/Ants-occupied-battle-resolution-method 2011-06-20T16:03:25 * janzert is working on a narrative explanation now 2011-06-20T16:03:45 I'm down to 2 lone ants of opposing colors ;) 2011-06-20T16:04:09 my link was a narrative 2011-06-20T16:04:37 not in quite the same way mine is, I'll post what I've got so far in just a bit 2011-06-20T16:04:53 okay, maybe together we can make it easy to understand 2011-06-20T16:05:16 I'm okay if it takes a little for users to grok battle resolution. That's where jeff disagrees. 2011-06-20T16:05:37 especially when the simple cases are really straightforward 2011-06-20T16:05:45 1 against 1 kills both 2011-06-20T16:05:47 I think seeing the game as a whole, people will think "Oh, you need to gather food and attack the enemy, cool" 2011-06-20T16:05:49 2 against 1 kills 1 2011-06-20T16:06:29 Also, when they see a good bot not lose anything, then they will be inspired to learn the "battle tricks" which is always fun. 2011-06-20T16:07:59 the github description confused me today. 'An ant lives if it has less enemies in range than each of those enemies have in range.' - maybe replace the word 'each of' with 'every one of'? 2011-06-20T16:08:30 george1914: try reading the link I just posted and tell me what you think 2011-06-20T16:08:36 ok 2011-06-20T16:08:39 reading 2011-06-20T16:10:24 https://github.com/aichallenge/aichallenge/wiki/Occupied-battle-resolution-narrative 2011-06-20T16:10:37 there is the start, it doesn't quite finish yet 2011-06-20T16:11:01 and of course is a very rough first draft 2011-06-20T16:12:34 McLeopold: i understood it best by reading the code. the description above was correct, but i would find it a little confusing if i didnt already know it. maybe that is just me. 2011-06-20T16:12:43 ok 2011-06-20T16:12:50 we are working to make it better 2011-06-20T16:13:23 i would be thinking 'hang on, doesnt ant A die anyway? (it does, but like you say, you dont choose to discuss that as well). 2011-06-20T16:13:51 so i guess it is ok description, but not perhaps as unambiguous as it could be .... 2011-06-20T16:13:58 george1914: any thoughts on the 'story' based approach? 2011-06-20T16:14:54 which isn't meant to be the completely definitive explanation even when completed of course, but a way to give people an intuitive understanding 2011-06-20T16:15:41 description seems fine and clear. though does not cover all cases (yet) 2011-06-20T16:18:54 janzert, first time reading all this... this line sounds scary "This process is of course quite complex", it is not immediatly obvious that it is complexity that the developer doen't need to worry about (that it is just a cute part of the story) 2011-06-20T16:20:06 sight: "This action is of course quite complex, taking much fine and detailed work by the ant" 2011-06-20T16:20:08 any better? 2011-06-20T16:20:44 not sure, you'll need to ask someone with no background knowledge, but be aware of it 2011-06-20T16:21:20 it's just that when I was reading it I wondered whether you had made the ant spawning more complex until I put things in context :) 2011-06-20T16:21:21 the whole story is meant to be a rather light hearted treatment as if the ants were actual physical creatures 2011-06-20T16:21:44 '(1) for every ant -j- that is fighting, count how many enemies it fights (ie within attackradius), call it e[j]. (2) now, every ant -j- kills every other ant -k- within its battleradius IF e[j] <= e[k]' 2011-06-20T16:22:08 *** Kingpin13 has joined #aichallenge 2011-06-20T16:22:37 george1914: and I think every person without a background in either math or programming got lost between word 2 and 4 ;) 2011-06-20T16:22:47 maybe :/ 2011-06-20T16:27:15 an ant dies, if it has the highest number of enemy ants of all of its enemies in its battle radius, if its a tie, all of those ants die 2011-06-20T16:28:07 janzert, the best suggestion I can make it is to use a different word than complex... I think delicate is a good choice since it is easily disrupted by enemy ants 2011-06-20T16:29:13 *** computerwiz_222 has joined #aichallenge 2011-06-20T16:30:05 is my explanation is correct at least? 2011-06-20T16:30:10 I'm also hoping the ease with which reader ends up resolving that 'complex' will at least subconsciously relax them when they encounter the second 'complex' at the end of the paragraph. :) 2011-06-20T16:30:16 delete second is 2011-06-20T16:30:57 janzert, yeah the second complex is actually something the programmer has to worry about :P 2011-06-20T16:31:03 mcstar: I believe so 2011-06-20T16:31:08 thx 2011-06-20T16:31:30 its the simplest so far for me 2011-06-20T16:32:01 *** Dlayne has quit IRC (Remote host closed the connection) 2011-06-20T16:53:34 *** carlop has joined #aichallenge 2011-06-20T17:01:08 *** bhasker has quit IRC (Quit: bhasker) 2011-06-20T17:01:29 janzert: look at the "wall punch" senario https://github.com/aichallenge/aichallenge/wiki/Ants-occupied-battle-resolution-method 2011-06-20T17:04:21 *** bhasker has joined #aichallenge 2011-06-20T17:05:09 yes? 2011-06-20T17:07:36 amstan: I could see that ant art you had going with this https://github.com/aichallenge/aichallenge/wiki/Occupied-battle-resolution-narrative 2011-06-20T17:11:10 McLeopold: https://github.com/aichallenge/aichallenge/wiki/Ants-occupied-battle-resolution-method looks good to me - nice examples :). 2011-06-20T17:11:28 maybe change radius=5 to radius^2 = 5? 2011-06-20T17:11:28 thx 2011-06-20T17:11:34 oh, yeah 2011-06-20T17:11:49 and there are some "?" in the maps that i suspect are typos? 2011-06-20T17:12:02 they are not 2011-06-20T17:12:41 When is the contest officially starting? 2011-06-20T17:12:53 astevenson: when it's ready ;) 2011-06-20T17:13:06 no date yet 2011-06-20T17:13:07 :) Ok so maybe Summer, maybe Winter? 2011-06-20T17:13:25 maybe when Duke Nuke Forever comes out? :) 2011-06-20T17:13:33 it'll all be over by christmas 2011-06-20T17:13:34 I really hope we're measuring time to start in less then months at this point :) 2011-06-20T17:13:46 ah, sounds good :) 2011-06-20T17:13:55 Ok sounds good, just curious.. 2011-06-20T17:14:29 I personally think the measure is still in weeks, but others might debate that we're closer 2011-06-20T17:15:24 but that I think is about the best idea anyone has of start date still 2011-06-20T17:15:31 the narrative looks ok also - needs to link to the 'rigorous' description also though i guess. btw maybe it is spelt 'peek' in no peeking? 2011-06-20T17:15:55 oops, yes 2011-06-20T17:16:13 I'm not talking about summiting a mountain anyway 2011-06-20T17:16:30 that is good to know - i can see why you dont want to have a fixed date - is nice to know very roughly when it is though, so thanks for that info :) 2011-06-20T17:24:06 I thought the start date was supposed to be back in April :P 2011-06-20T17:25:51 *** bhasker has quit IRC (Quit: bhasker) 2011-06-20T17:26:49 *** bhasker has joined #aichallenge 2011-06-20T17:36:03 *** MuTa-ofd00m has quit IRC (Quit: Bye) 2011-06-20T17:37:34 for help understand the battle resolution method, i think the best way should start with "Calculate for each ant the number of enemy ant in his attack range (this is how much an ant is "occupied").", and then telling when an ant survive or nto 2011-06-20T17:37:40 seem clearer to me 2011-06-20T17:39:38 so maybe the line "the enemy is surronded by less (or the same) of its own enemies than the ant" can become "the enemy is less or equal occupied than the ant" 2011-06-20T17:41:52 a part from this, I think that McLeopold description work 2011-06-20T17:42:30 *** acieroid` has joined #aichallenge 2011-06-20T17:44:17 *** amstan_ has joined #aichallenge 2011-06-20T17:44:17 *** ChanServ sets mode: +o amstan_ 2011-06-20T17:45:24 *** acieroid has quit IRC (Disconnected by services) 2011-06-20T17:45:28 *** acieroid` is now known as acieroid 2011-06-20T17:47:02 carlop: I does sound nicer, but then you shouldn't define "occupied" with the word "occupied" 2011-06-20T17:47:11 aichallenge: McLeopold epsilon * r1522a2a / (sql/0_schema.sql sql/2_generate_matchup.sql): make choosing seed player faster - http://bit.ly/jZAUCZ 2011-06-20T17:47:11 aichallenge: McLeopold epsilon * rff6acca / (ants/ants.py ants/playgame.py): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/iUFR9U 2011-06-20T17:47:20 sigh: the original date was february :) 2011-06-20T17:50:22 *** computerwiz_222 has quit IRC (*.net *.split) 2011-06-20T17:50:22 *** amstan has quit IRC (*.net *.split) 2011-06-20T17:50:22 *** Redgis has quit IRC (*.net *.split) 2011-06-20T17:51:23 Zannick, there was actually a date back then? I don't remember it being specified even to the month then, just the season 2011-06-20T17:51:38 it was more like "let's aim for february" 2011-06-20T17:51:51 ie. rather than march 2011-06-20T17:52:00 ah fair enough 2011-06-20T17:52:10 well, the way things are that seems like it might be a valid plan :P 2011-06-20T17:52:18 bahaha 2011-06-20T17:52:43 *** Zaphus has joined #aichallenge 2011-06-20T17:54:45 *** ltriant has joined #aichallenge 2011-06-20T17:58:00 McLeopold: hmm, only the information for 2 of the players get entered in this game http://aichallengebeta.hypertriangle.com/visualizer.php?game=28796 2011-06-20T17:58:18 was that from something you're in the process of working with? 2011-06-20T17:58:24 no 2011-06-20T17:58:38 :( 2011-06-20T17:58:56 mcLeopold... do you see the "versions" in a crazy order on this page ? http://aichallengebeta.hypertriangle.com/profile.php?user=107 I see 17, then 11, then 33 - that can't be right ?! 2011-06-20T17:59:29 Zaphus: looks like I need an order by... 2011-06-20T17:59:57 except the version 17 and 11 games happened after the version 33 games 2011-06-20T17:59:59 Zaphus: actually, it looks like some old submissions got activated 2011-06-20T18:00:05 ouch 2011-06-20T18:00:28 maybe something with an old matchup got put back in the queue with the recent changes? 2011-06-20T18:00:35 oh, that's it 2011-06-20T18:00:51 so, that part is working :) 2011-06-20T18:01:25 Ah, I see you've made some changes 2011-06-20T18:05:25 *** bhasker has quit IRC (Quit: bhasker) 2011-06-20T18:07:08 *** bhasker has joined #aichallenge 2011-06-20T18:07:14 Is anyone responsible for the www.ai-contest.com site ? It's been gone for 24 hours now 2011-06-20T18:07:35 *** Redgis has joined #aichallenge 2011-06-20T18:10:00 Basically we're leaving it for Jeff to fix, he seems to have broken it through the hosting control panel :( 2011-06-20T18:10:25 2330 ants?!?! 2011-06-20T18:10:31 and he's the only one that has access 2011-06-20T18:11:07 McLeopold: can you show me how to do an SQL query that gets the player ID's and scores for each match ? I assume the data is in that SQL dump but I struggle to find it 2011-06-20T18:12:08 select * from game_player where game_id=; will give you everything stored about each player in the game for the game 2011-06-20T18:13:33 just the game score and player id would be select player_id, game_score from game_player where game_id=1234; 2011-06-20T18:14:03 you probably really want the user_id though 2011-06-20T18:14:56 I assume player_id is the player number in the game? 2011-06-20T18:15:07 not positive on that 2011-06-20T18:15:32 ok, cool.. so there's only 192 games in the file I have. Just looking at how something like this: http://pastebin.com/f63gK0Cb could be done (thats from my local testing framework, not the sql) 2011-06-20T18:15:56 janzert: matchup fairness http://pastebin.com/CN4XLZT0 2011-06-20T18:16:39 game_delay is time since last game? 2011-06-20T18:16:55 number of games played since last game 2011-06-20T18:17:00 ahh 2011-06-20T18:17:05 this only looks at the last 1000 games 2011-06-20T18:17:43 janzert: you can fire up the worker again 2011-06-20T18:17:55 ok 2011-06-20T18:17:58 hopefully things will even out in a few games 2011-06-20T18:19:06 *** mleise has quit IRC (Ping timeout: 252 seconds) 2011-06-20T18:19:16 you just use that for seeding, and the rest of the players get selected via another method right ? 2011-06-20T18:19:28 *** Alex__ has joined #aichallenge 2011-06-20T18:19:57 *** carlop has quit IRC (Ping timeout: 250 seconds) 2011-06-20T18:20:22 Zaphus: essentially, yes 2011-06-20T18:20:24 Zaphus: seeding is now basically just player with the least recent game 2011-06-20T18:20:57 perfect. as long as the other players are chosen by a different method - so that you dont get the same bunches of players every time round the loop 2011-06-20T18:20:57 Hi all! When will the forum work? 2011-06-20T18:21:22 Alex__: hopefully by tomorrow morning 2011-06-20T18:21:57 thanks 2011-06-20T18:22:12 *** carlop has joined #aichallenge 2011-06-20T18:23:17 I noticed that the in last game was used my old bot (version1) instead of the latest version 4 - http://aichallengebeta.hypertriangle.com/profile.php?user=139 2011-06-20T18:23:22 *** RPleonasm has joined #aichallenge 2011-06-20T18:23:58 a bunch of old games that hadn't been played when they were supposed to be got played finally 2011-06-20T18:24:18 once they get finished everything should return to normal 2011-06-20T18:24:29 ok 2011-06-20T18:25:23 aichallenge: McLeopold epsilon * rbbba691 / sql/check_matchup_fairness.sql : added matchup fairness sql - http://bit.ly/jN2yWx 2011-06-20T18:27:16 McLeopold: you could now add a "next game in %d games or earlier" to the website for each player :-) 2011-06-20T18:31:01 why do I get the message "compiled but failed test cases" after uploading? 2011-06-20T18:31:27 my bot runs fine against test_bot 2011-06-20T18:34:40 *** stocha has joined #aichallenge 2011-06-20T18:35:02 Zaphus: http://pastebin.com/TKMd4dPm 2011-06-20T18:35:11 i like the work occupied (though maybe i use the french similar word). I picture it as "busy" 2011-06-20T18:35:22 So a ant is as busy as there is ennemy in sight. 2011-06-20T18:35:35 stocha: yeah, that's the idea 2011-06-20T18:35:44 Zaphus: can you open an issue on the next game idea 2011-06-20T18:35:53 reverse would be attentive. so busy can be distracted also. 2011-06-20T18:36:07 RPleonasm: what's the bot name? 2011-06-20T18:36:07 So the most attentive ant, will kill any more distracted ants. 2011-06-20T18:36:17 stocha: exactly 2011-06-20T18:36:19 (it focus on distracted ant and tore it to pieces) 2011-06-20T18:36:30 oh it was all already settled ? 2011-06-20T18:37:29 RPleonasm: you got caught on the bad server, upload again and it will be fine 2011-06-20T18:37:43 oh ok, thanks :) 2011-06-20T18:38:15 maybe only one term should be used. So let's go for distracted maybe. 2011-06-20T18:38:50 Every ennemy ant an ant can see distract it. The less distracted ant will kill any more distracted ennemy in sight. 2011-06-20T18:38:59 maybe focus? 2011-06-20T18:39:16 If an ant is more focused than an enemy, it will kill it 2011-06-20T18:39:31 i'd rather have it distracted. 2011-06-20T18:39:42 then it's direct proportionnale to the ennemy count. 2011-06-20T18:40:02 that's why I stated it both ways in the wiki 2011-06-20T18:40:05 my goal was to have only one term. 2011-06-20T18:40:32 stocha: too late for that we already got two, you can only add more now :} 2011-06-20T18:40:49 mm, i'll use distracted in my own head then :p 2011-06-20T18:41:03 at least someone will profit from my hard reflections :p 2011-06-20T18:41:19 doesn't really change the math does it. 2011-06-20T18:42:08 that's probably the real problem with this method is that the math came first and we're trying to find good natural language ways of stating 2011-06-20T18:42:27 *** bhasker has quit IRC (Quit: bhasker) 2011-06-20T18:42:28 the math came second 2011-06-20T18:43:23 *** bhasker has joined #aichallenge 2011-06-20T18:43:53 i don't really like the term occupied. For me it reffers to "occupied territory" :p 2011-06-20T18:44:09 i'd rather have it "busy" or something more standard for business than occupied. 2011-06-20T18:44:25 stocha, you may be right 2011-06-20T18:44:39 distracted/busy/focus are all better terms 2011-06-20T18:45:07 let's update that word then. 2011-06-20T18:45:14 to throw another one out there, I found myself using 'attention' a lot when writing the narrative 2011-06-20T18:45:37 I like distracted as well 2011-06-20T18:45:39 occupied as well as attention are from french origins :p 2011-06-20T18:45:56 you wouldn't want to use french words would you. 2011-06-20T18:45:58 janzert: just thought about devided attention :) 2011-06-20T18:46:24 e->i 2011-06-20T18:46:34 no 2011-06-20T18:46:47 the thing is, the more the word is close to core english (the most used english words) the better. 2011-06-20T18:46:58 yes 2011-06-20T18:47:04 usually word from french origine are farther appart from the core. 2011-06-20T18:47:15 they are more refined or something :p 2011-06-20T18:47:40 although my knowledge of english being limited, i can't say for sure. 2011-06-20T18:47:43 i have to go. 2011-06-20T18:47:49 *** stocha has quit IRC (Quit: Page closed) 2011-06-20T18:48:08 McLeopold: that search is very close to what I want, only difference is that I would change it to class a 'win' against another player as finishing higher than them on a game (e.g. 2 beats 3) 2011-06-20T18:48:30 http://ngrams.googlelabs.com/graph?content=occupied%2Cattention%2Cdistracted&year_start=1950&year_end=2011&corpus=0&smoothing=3 2011-06-20T18:48:40 by usage metric then attention wins ;) 2011-06-20T18:49:34 Zaphus: good point 2011-06-20T18:50:09 Zaphus: that requires a rewrite 2011-06-20T18:54:52 Zaphus: http://pastebin.com/TKMd4dPm 2011-06-20T18:55:38 of course, by total percent is inverted :/ 2011-06-20T19:04:11 *** Kingpin13 has quit IRC (Ping timeout: 240 seconds) 2011-06-20T19:04:30 *** nux67 has quit IRC (Quit: Page closed) 2011-06-20T19:10:33 *** amstan has joined #aichallenge 2011-06-20T19:10:33 *** ChanServ sets mode: +o amstan 2011-06-20T19:10:58 *** Kingpin13 has joined #aichallenge 2011-06-20T19:13:17 McLeopold: I've been thinking the sigma increase should be changed to be based on percentage of the population that has changed such that 100% of population change would set sigma to the initial value 2011-06-20T19:14:16 *** choas has quit IRC (Quit: leaving) 2011-06-20T19:14:20 *** amstan_ has quit IRC (Ping timeout: 260 seconds) 2011-06-20T19:14:54 janzert: so if we want to take the leader down from the top of the board, a bunch of us (all of us) should resubmit and their sigma will plummet :) 2011-06-20T19:15:26 except everyone elses will have too :P 2011-06-20T19:15:44 and even worse the leader will still have his mu 2011-06-20T19:16:07 so much for a way to overthrow the leader :} 2011-06-20T19:18:02 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-06-20T19:19:04 McLeopold: just noticed the check_matchup_fairness scripts don't filter out users who don't have a currently active submission? 2011-06-20T19:20:08 janzert: it's just a quick check, so no 2011-06-20T19:20:30 janzert: I agree with the sigma change. 2011-06-20T19:20:41 I think I asked you that question last week? 2011-06-20T19:20:48 ok, I was wondering why three users weren't getting games still 2011-06-20T19:21:57 hmm, actually why is amstan still not getting games http://aichallengebeta.hypertriangle.com/profile.php?user=2 2011-06-20T19:22:22 I disabled him awhile ago 2011-06-20T19:22:36 janzert: i also have 101 pages of games 2011-06-20T19:23:01 McLeopold: why did you disable me? 2011-06-20T19:23:11 you had a starter bot 2011-06-20T19:23:25 ahh, ok I see the submission status is still ready to play but latest=0 2011-06-20T19:24:31 dinner and a movie time for me, be back for the meeting 2011-06-20T19:24:40 @topic 2011-06-20T19:24:40 McLeopold: Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || Code Repo: http://github.com/aichallenge/aichallenge || Beta testers needed: http://aichallengebeta.hypertriangle.com/ (amstan) || Launch Preparation Meeting http://bit.ly/kYYbD4 (amstan) 2011-06-20T19:25:04 *** McLeopold has left #aichallenge 2011-06-20T19:27:00 *** Redgis has quit IRC (Quit: ... mains libres) 2011-06-20T19:28:22 *** Alex__ has quit IRC (Ping timeout: 252 seconds) 2011-06-20T19:31:42 *** RPleonasm has quit IRC (Quit: Page closed) 2011-06-20T19:38:49 *** Apophis has quit IRC (Read error: Connection reset by peer) 2011-06-20T19:40:01 *** Apophis has joined #aichallenge 2011-06-20T19:40:18 ahh, turns out i should be fine for the meeting 2011-06-20T20:00:19 *** FireFly has quit IRC (Quit: swatted to death) 2011-06-20T20:06:16 *** Apophis has quit IRC (Read error: Connection reset by peer) 2011-06-20T20:08:01 oh man, flag is owning, look at his ants chase the pink ones: http://aichallengebeta.hypertriangle.com/visualizer.php?game=28870&user=7 2011-06-20T20:08:21 <_flag> amstan: :) 2011-06-20T20:08:51 _flag: though, i noticed that if your ants are not together, but attack an enemy ant from different sides, they hesitate 2011-06-20T20:08:54 *** Apophis has joined #aichallenge 2011-06-20T20:09:28 is that a game mechanic thing(are you not more powerful?) or is it just your code? 2011-06-20T20:09:48 *** sigh has quit IRC (Quit: This computer has gone to sleep) 2011-06-20T20:10:13 <_flag> amstan: This is because I'm assuming that the enemy will go in one of the various directions they can in order to secure a 1v1 trade, so I try to move such that I will always kill more ants (or do nothing at all) than he will 2011-06-20T20:11:02 <_flag> The fact that there are only four directions is also a bit of a problem because I can't go towards an enemy and stay alongside it if it's running in a certain direction at the same time 2011-06-20T20:11:08 _flag: it takes you a really long time to finish games off :P 2011-06-20T20:11:46 _flag: are you using a decistion tree for your ants? 2011-06-20T20:11:51 <_flag> I'm thinking of adding a condition that will be more agressive if I'm winning, but I don't think it's that big of a deal 2011-06-20T20:11:55 and multiple ants at once or just a single ant at a time? 2011-06-20T20:12:19 <_flag> antimatroid: contest secret :) 2011-06-20T20:12:22 :P 2011-06-20T20:12:43 i haven't worked out how i'm going to do battles yet 2011-06-20T20:14:29 *** sigh has joined #aichallenge 2011-06-20T20:19:16 *** Apophis has quit IRC (Ping timeout: 258 seconds) 2011-06-20T20:20:25 *** Apophis has joined #aichallenge 2011-06-20T20:42:50 *** Evious has joined #aichallenge 2011-06-20T21:00:08 contestbot: seen j3camero 2011-06-20T21:00:08 amstan: j3camero was last seen in #aichallenge 2 weeks, 5 days, 0 hours, 29 minutes, and 58 seconds ago: Okay I am out for the night. Catch y'all later. 2011-06-20T21:00:19 Yo 2011-06-20T21:00:24 Somebody say my name? 2011-06-20T21:00:30 j3camero: omg! 2011-06-20T21:00:38 j3camero: dude.. i need your phone number 2011-06-20T21:00:49 You can have all seven of them if you want. 2011-06-20T21:00:51 j3camero: the one you're listing on facebook/resume belongs to one "emma chi" 2011-06-20T21:01:12 Is she nice? 2011-06-20T21:01:34 well, she only says her name on the answering machine 2011-06-20T21:01:40 Wanna light up some EC2 instances? 2011-06-20T21:01:45 she seems nice from that.. but that's the most i got 2011-06-20T21:01:48 janzert: ping 2011-06-20T21:02:31 j3camero: sure 2011-06-20T21:03:24 that new seeding method looks like it is working fantastically, every bot I have looked at is getting a similar rate of games (I've looked at about 20) 2011-06-20T21:03:55 Okay let's do this thing. I will light up some instances then pass the API details to the peoples. Then we can light up a real contest instance on those. 2011-06-20T21:04:33 j3camero: ok, decent 2011-06-20T21:04:37 What sorts of servers do we want for the workers? 2011-06-20T21:04:48 I think janzert had some picked out that he wanted. 2011-06-20T21:04:49 *** onensora has quit IRC () 2011-06-20T21:04:57 all i know is ubuntu natty 64 bit, janzert knows more details 2011-06-20T21:05:51 janzert's probably still movieing 2011-06-20T21:06:01 might as well ping mcleopold too 2011-06-20T21:06:30 Okay well no worries just now. I will fire an email to the list asking him which servers we should use, and light them up when he replies. 2011-06-20T21:07:05 antimatroid: ping too 2011-06-20T21:07:45 j3camero: ok, cool 2011-06-20T21:07:48 j3camero: though i expect them to be here soon 2011-06-20T21:08:26 j3camero: so what's going on with the old server? it's like the apache files got overwritten 2011-06-20T21:09:12 Yeah I don't really know. 2011-06-20T21:09:18 I haven't looked at it too much. 2011-06-20T21:09:27 It was getting only a few dozen pageviews per day anyways. 2011-06-20T21:09:32 I think. 2011-06-20T21:09:54 people were using the forums for some deving 2011-06-20T21:10:04 Oh haha make that 250 ish per day. 2011-06-20T21:10:39 What's it mean in the rankings when someone's name appears in a lighter font? 2011-06-20T21:11:02 they were there before, but they submitted another bot and they're doing worse now 2011-06-20T21:11:17 Oh I see. 2011-06-20T21:11:26 its good to have, but very annoying when there are lots, should be optional 2011-06-20T21:11:39 we can probably tweak it 2011-06-20T21:11:51 My thought is to disable it by default, as cool as it is. Or at least explain it. It's really confusing. 2011-06-20T21:12:10 or... if you are logged in then show your ghost by default but not the others ? 2011-06-20T21:12:22 That would be good yeah. 2011-06-20T21:12:36 Or even not. It would be so much clearer to display a user's peak rank on their profile. 2011-06-20T21:12:43 Along with the time when that rank was last achieved. 2011-06-20T21:12:46 Something like that. 2011-06-20T21:13:09 j3camero: Wouldn't that be misleading: people who submit early would naturally have a very good peak rank. 2011-06-20T21:13:17 That's okay. 2011-06-20T21:13:20 go the whole hog, add a rating or rank graph over time ? 2011-06-20T21:13:29 Yeah why not. 2011-06-20T21:13:59 Having ghosts in the rankings is pretty confusing. We could explain it, but most people would not read the explanation and be confused anyways. 2011-06-20T21:14:19 If your new bot sucks more (or other people have improved their bots around you) then your ranking should go down. 2011-06-20T21:14:28 *** antimatroid-uni has joined #aichallenge 2011-06-20T21:14:46 bah, sorry i'm late :P 2011-06-20T21:15:04 if it only persisted until you'd played games (eg 5) then that might help, just waiting until your new bot gets a ranking 2011-06-20T21:15:39 *** McLeopold has joined #aichallenge 2011-06-20T21:16:44 ok, noted: https://github.com/aichallenge/aichallenge/issues/195 2011-06-20T21:16:51 Is j3camero around? 2011-06-20T21:16:57 McLeopold: yes 2011-06-20T21:17:14 McLeopold: now we need janzert to tell us what ec2 instances to get 2011-06-20T21:18:04 doesn't look like a missed much :) 2011-06-20T21:18:13 *** bhasker has quit IRC (Quit: bhasker) 2011-06-20T21:18:32 well, till then we could move to the rules finalization 2011-06-20T21:18:41 /game parameters 2011-06-20T21:19:50 I guess we could do that now. 2011-06-20T21:20:09 Though I feel many IRC regulars are probably better qualified for that by now. 2011-06-20T21:20:22 i like spawn2 = 1, attack2 = 5 and view2 = 55 2011-06-20T21:20:54 j3camero: we've gone back to the more confusing battle option (quite recently) 2011-06-20T21:21:05 Okay. 2011-06-20T21:21:05 we discovered that the damage option allowed blocking far too easily 2011-06-20T21:21:23 and in general seemed to favour the "defender" much more in such situations 2011-06-20T21:21:39 What do we do if someone discovers a degerate or semi-degenerate tactic with the latest battle logic? 2011-06-20T21:21:58 we can make maps more open depending on how much that helps 2011-06-20T21:22:02 As long as we have a backup ready to go in the engine then I'm cool. 2011-06-20T21:22:07 otherwise we're always going to have that doubt to some degree 2011-06-20T21:22:17 there are other options in the engine worst comes to worst 2011-06-20T21:22:27 Sounds good to me. 2011-06-20T21:22:47 * janzert is here now; checking archives 2011-06-20T21:22:48 does anyone have any complaints with scoring? 2011-06-20T21:23:12 Hey janzert. I was gonna light up some EC2 instances now. Which ones should we get? 2011-06-20T21:23:46 @amstan: FYI this issue: https://github.com/aichallenge/aichallenge/issues/174 is also related to the ghosts/high water 2011-06-20T21:23:47 Zaphus: User error -- Replace user. 2011-06-20T21:23:52 amstan: FYI this issue: https://github.com/aichallenge/aichallenge/issues/174 is also related to the ghosts/high water 2011-06-20T21:26:20 janzert: what are your thoughts on max turns? 2011-06-20T21:26:21 per McLeopold the ghosts are only meant to hang around for a bit while the new submissions rank stabilizes, they hang around forever now to help him during beta 2011-06-20T21:26:28 i think we should reduce it quite a bit for the moment 2011-06-20T21:26:45 oh I guess McLeopold showed up before I did anyway :) 2011-06-20T21:27:23 janzert what sort of EC2 instances do we want? 2011-06-20T21:27:37 j3camero: for what, workers? 2011-06-20T21:27:57 don't just fire a bunch up though 2011-06-20T21:27:59 Yep 2011-06-20T21:28:05 Okay 2011-06-20T21:28:27 m2.xlarge for workers is what I've been using 2011-06-20T21:28:31 I was thinking of starting 2 workers tonight. 2011-06-20T21:28:49 but you need the right ami and user data before requesting them 2011-06-20T21:30:22 *** sigh has quit IRC (Quit: This computer has gone to sleep) 2011-06-20T21:30:42 antimatroid-uni: I've been working on downloading game data today so I can actually build that histogram to see what length games have been taking 2011-06-20T21:33:07 i think we can afford to cut it down further than that might imply 2011-06-20T21:33:29 look at any of my bots games, it isn't smart with battles at all but reasonably aggressive at chasing after enemies 2011-06-20T21:33:40 when it wins it doesn't usually take even 500 turns 2011-06-20T21:33:40 but at least it will give us an idea what is happening now 2011-06-20T21:33:49 yep 2011-06-20T21:33:51 yep 2011-06-20T21:34:07 i think we all agree it's okay to make people work for the food bonus 2011-06-20T21:34:14 it seems many reasonably good playing bots finish in under 500 2011-06-20T21:34:19 although it might open the door up for farming again? 2011-06-20T21:35:01 if there's a high chance of missing out on the bonus, let them stay just above the level where the game would be called off, i'm thinking about ending the game early, but yeah that^^ 2011-06-20T21:38:14 I'm also still enamored with making a per-map dynamic turn limit ;) 2011-06-20T21:38:38 it affects the food bonus amount 2011-06-20T21:38:42 but so does the spawn rate etc. 2011-06-20T21:38:52 * janzert nods 2011-06-20T21:41:06 *** McLeopold has quit IRC (Ping timeout: 252 seconds) 2011-06-20T21:42:21 janzert: ok, well, should we get a server up then? 2011-06-20T21:42:38 I messaged j3camero with details 2011-06-20T21:42:44 but I think he went afk 2011-06-20T21:42:52 damn 2011-06-20T21:42:54 Im here. 2011-06-20T21:43:00 Listening. Lurking. |Waiting. 2011-06-20T21:43:10 did you get my pm's? 2011-06-20T21:43:28 Yep just saw them now that you mentioned them. 2011-06-20T21:43:33 Will start those up now. 2011-06-20T21:43:34 :) 2011-06-20T21:43:49 decent, decent 2011-06-20T21:45:12 there should be a "previous next" page link section below the leaderboard 2011-06-20T21:47:58 *** McLeopold has joined #aichallenge 2011-06-20T21:53:46 What are we doing about a web server? 2011-06-20T21:54:36 janzert: are the workers more stable this time? Do they have to be rebooted often? 2011-06-20T21:54:45 they seem to be much more stable 2011-06-20T21:54:50 It's fine if they do, I am just wondering what percentage of uptime we get from each one. 2011-06-20T21:54:52 For game length, I'd rather not turn it down just yet. 2011-06-20T21:55:09 I'm going to put in some logic to cut games off and we can see how it does on beta. 2011-06-20T21:55:18 I've seen our own script crash out a few times for bugs that I've then fixed :/ 2011-06-20T21:55:32 but not seen the OS actually lock up so we should be pretty good 2011-06-20T21:55:39 That means we could turn game length up very high and it only goes long when it needs to. 2011-06-20T21:55:45 McLeopold: i would like to get away from the beta server, we have like 2 days left till we're charged for 2 months 2011-06-20T21:56:02 Sounds good McLeopold. Game length is touchy, since it's something that everyone sees. Games cut short can confused people and make them perceive the contest as not fair or broken. Just things to keep in mind. 2011-06-20T21:56:28 I went trawling through WHT forum for good dedicated server providers and came up with one the other night, but of course now I can't remember the name :P 2011-06-20T21:56:32 The official discussion is here. https://github.com/aichallenge/aichallenge/issues/193 2011-06-20T21:56:34 Sounds good janzert. OS locking up is an issue that has plagued both Tron and Galcon contests. 2011-06-20T21:56:35 let me see if I can dig it back up again 2011-06-20T21:57:07 The goal is to make sure everyone agrees that the games that were cut off are the ones that should have been. I think we will find a good metric. 2011-06-20T21:57:07 ahh, yes singlehop.com 2011-06-20T21:57:17 anyone have first hand experience with them? 2011-06-20T21:57:40 I like the "overpowering" idea - if someone dominates the game for a significant length of time, end it. Seems fair on everyone, and unlikely to ever change the final result 2011-06-20T21:58:06 Nope no experience. 2011-06-20T21:58:11 What do you like singlehop? 2011-06-20T21:58:22 There are 2 types of games that need to be cut off, 1) all stupid bots that can't fight, 2) 1 dominate bot that won't end the game 2011-06-20T21:58:25 they do have ppl smiling 2011-06-20T21:58:40 My experience with our last host was pretty good. Nothing to complain about. Us having a shitty server was my fault. 2011-06-20T21:58:41 people seemed to say good things about them on WHT forum and the look like a good price 2011-06-20T21:58:53 Nice nice. 2011-06-20T21:59:00 WHT is webhostingtalk.com btw 2011-06-20T21:59:15 Maybe I will snag one of those and direct the ai-contest.com domain at it. 2011-06-20T21:59:21 amstan: can we still have some type of beta server going? 2011-06-20T21:59:29 lot of providers seem to hang out there 2011-06-20T21:59:39 For now let's stick with amstan's beta server with a weird URL to avoid a false start. 2011-06-20T21:59:39 McLeopold: yes, we'll setup a new beta on the new servers 2011-06-20T21:59:49 oh? 2011-06-20T21:59:53 McLeopold: the official server can simply be in beta for a while 2011-06-20T22:00:16 Which means I'm allowed to touch it and break things? 2011-06-20T22:00:25 well, the beta server that we got now is from demerzel 2011-06-20T22:00:53 McLeopold: is there any objections to making you an admin? you did say you don't want to compeete 2011-06-20T22:01:03 hmm, that's right once we move to whatever will be the official main server we'll need to lock out everone that doesn't want to give up participation privileges 2011-06-20T22:01:09 On the ghost issue. I plan to make it on by default, and allow users to turn it off if they want. 2011-06-20T22:01:48 McLeopold: can they disappear once the new submissions sigma is below a certain value? 2011-06-20T22:01:57 that really makes the most sense to me 2011-06-20T22:02:27 I think it's a good idea to have two web servers running (even if they're on the same physical box). One "staging area" at staging.ai-contest.com and then the main one at ai-contest.com. That way we can deploy to the staging area first without worrying about breaking stuff. 2011-06-20T22:03:04 j3camero: i'm thinking about an archive server, we could put the staging on the same one 2011-06-20T22:03:06 yes, I like that idea 2011-06-20T22:03:28 *** slkjcna_ has joined #aichallenge 2011-06-20T22:03:32 j3camero: if it's the same machine I think we still need to lock out any non-organizers 2011-06-20T22:03:54 *** _flag_ <_flag_!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-20T22:04:02 Yeah these things can all go on one physical box. The staging area plus the archive would get a few hundred pageviews per day or something easy like that. 2011-06-20T22:04:12 janzert: why? 2011-06-20T22:04:27 j3camero: the testing people should have root 2011-06-20T22:04:41 so.. it would make it hard to lock them out if they're on the same comp 2011-06-20T22:04:44 because they could gain access to the submissions of the contest 2011-06-20T22:05:04 I think we are not talking about the same thing. 2011-06-20T22:05:06 and in any case it would be viewed as suspicious I believe 2011-06-20T22:05:58 I am talking about another instance of apache and another instance of mysql that is just a mirror of the main server. Except that we deploy there first before deploying to the main instance. It's just so we can catch breakages before deploying for real. 2011-06-20T22:06:10 okay, gtg, do a @later if you need me to do something 2011-06-20T22:06:38 j3camero: we won't be able to catch much with just a mirror website 2011-06-20T22:06:44 yeah i'm off to the shop 2011-06-20T22:06:50 i'll be back in a bit if you need anything from me 2011-06-20T22:06:58 maybe not, I mean that only people that give up participation should have access to the machine that hosts the website/database or workers for the main contest 2011-06-20T22:07:08 j3camero: so it would not be running matches itself ? just using a copy of the SQL DB ? 2011-06-20T22:07:27 yes janzert agreed. Only admins can have shell access to the machines. 2011-06-20T22:07:59 Zaphus: I am thinking something like the current beta server. 2011-06-20T22:08:23 ok, which means a staging area on the same machine won't be accessible to McLeopold, mleise and such 2011-06-20T22:08:40 j3camero: but people with access to the current Beta have access to the bot submissions, right ? - and if you want people to upload bots to test the staging they need to be protected 2011-06-20T22:08:43 j3camero: what i'm thinking about is a totally separated host that could be used for most developers in testing 2011-06-20T22:08:54 *** Apophis has quit IRC (*.net *.split) 2011-06-20T22:08:54 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (*.net *.split) 2011-06-20T22:08:54 *** smly- has quit IRC (*.net *.split) 2011-06-20T22:08:55 *** slkjcna has quit IRC (*.net *.split) 2011-06-20T22:09:03 Zaphus: yes, they have access to them 2011-06-20T22:11:23 *** bhasker has joined #aichallenge 2011-06-20T22:11:38 Okay we have got some EC2 instances up. 2011-06-20T22:11:43 janzert can access them. 2011-06-20T22:12:05 I think the dual servers is causing the same match to be played twice: http://aichallengebeta.hypertriangle.com/profile.php?user=16 2011-06-20T22:13:11 j3camero: can i have access? 2011-06-20T22:13:42 you actually already do ;) 2011-06-20T22:14:01 you just need to know the secret to get to 'em :P 2011-06-20T22:14:45 yes.. wel 2011-06-20T22:15:37 Security by obscurity FTW. 2011-06-20T22:18:22 j3camero: isn't this cryptography business all obscurity? 2011-06-20T22:18:41 It's true. 2011-06-20T22:19:03 I know these two huge prime numbers. They're so obscure you've probably never heard of them before. 2011-06-20T22:19:08 The ultimate hipster. 2011-06-20T22:19:39 lol 2011-06-20T22:21:07 * janzert just realized we could easily just do away with worker login access completely and just kill and start new ones as needed :) 2011-06-20T22:22:39 <_flag_> So is there a projected contest start date yet? 2011-06-20T22:22:49 j3camero: so we have this cool bashrc script that tells us info when we "su contest": http://imgur.com/vPpQq 2011-06-20T22:22:55 _flag_: yes, good question 2011-06-20T22:23:09 i would say 2 weeks 2011-06-20T22:23:36 * amstan has said that many times, but it never happened, i hope it will this time 2011-06-20T22:23:36 here's a game where 2000 turns wasnt quite enough for one player... he would have won if there were a few more turns (I would have won if I wasnt so stupid). 300:50 ants on a small map and I run away :-) http://aichallengebeta.hypertriangle.com/visualizer.php?game=28997&user=78 2011-06-20T22:24:00 Cool. 2011-06-20T22:24:12 fyi, from july 1-9 I'll be out of town again and not around much or any 2011-06-20T22:24:34 ok.. 10 july? 2011-06-20T22:24:50 <_flag_> What still needs to be done that will take 2 weeks? 2011-06-20T22:25:25 sounds like a good spot to shoot for at least 2011-06-20T22:25:33 _flag_: https://github.com/aichallenge/aichallenge/issues?milestone=3&state=open 2011-06-20T22:25:41 _flag_: email, completely stable trueskill, etc. 2011-06-20T22:25:44 I think it's a good goal. Let's not go advertising that just yet though. 2011-06-20T22:25:59 yeah, we're not ready to advertise a date for sure 2011-06-20T22:26:17 not even on github? 2011-06-20T22:26:34 you can reset the launch milestone date if you wish ;) 2011-06-20T22:27:49 j3camero: so at an average price of $0.175 per hour per worker how many workers can we afford? 2011-06-20T22:28:10 I am gonna go with two for now. 2011-06-20T22:28:13 *** sigh has joined #aichallenge 2011-06-20T22:28:15 so we can figure out what what game parameters are actually feasible 2011-06-20T22:31:37 *** Kingpin13 has quit IRC (Quit: this used to be the life, but I don't need another one) 2011-06-20T22:34:08 Zaphus: 28991 and 28990 both came from the same worker somehow 2011-06-20T22:34:53 janzert: odd... I've also spotted two games with only one bot difference, and one game apart 2011-06-20T22:36:22 29014 and 29030 - same map, same 7 bots (different result) 2011-06-20T22:37:11 can you file an issue for it? 2011-06-20T22:37:22 yeah, trying to find better examples 2011-06-20T22:37:28 I'm guessing it's a result of the changes McLeopold made earleir 2011-06-20T22:38:06 not sure exactly how of course, but the parts of code that could case this were touched in those 2011-06-20T22:41:37 *** McLeopold has left #aichallenge 2011-06-20T22:43:00 time for lunch :-) 2011-06-20T22:45:20 *** carlop has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110413222027]) 2011-06-20T22:49:03 contestbot: later tell McLeopold looks like maybe something in the new scheduling code is allowing a matchup to run twice https://github.com/aichallenge/aichallenge/issues/196 2011-06-20T22:49:03 janzert: Job's done. 2011-06-20T22:51:40 j3camero: well, i'm going to go now, but i would really want a better way to contact you in the future 2011-06-20T23:00:03 Text me at 416 464 3096. 2011-06-20T23:00:55 j3camero: ok, cool, thanks