[00:30] Join: flyduck joined #corewars [00:52] MSG: Quit: This computer has gone to sleep [01:01] Time to go, I hope everyone has a fantastic Christmas [01:01] Join: baudchan joined #corewars [01:40] hi everyone [01:40] around andrew? : ) [01:40] AndrewBC (beep!) [01:46] Heya flyduck [01:46] I juuust got home from work :) [01:46] cool :P [01:46] How's it going? [01:47] ive been working on a warrior for the beginners hill.. [01:47] oh? [01:47] do tell :D [01:47] best score so far is 57.8, i need like double that to broach the bottom of the hill :o [01:47] hmmm [01:47] what's the premise of the warrior? [01:48] im still working with that replicator concept [01:48] but have added in a bomb to it [01:48] ooh :) [01:48] so it replicates itself and SPL's to the new copy, then the old copy starts bombing [01:48] nice [01:50] i think the concept is ok but now im doing a lot of offset math :P [01:50] yeah [01:51] are you working on anything? [01:53] well, not really. I reformed my DoubleDwarf last night, and it did pretty well as you saw (45.5% wins), but I'm still struggling to understand some of the concepts, mostly because of my very very dusty assembly programming skills [01:53] 45.5 on the nano hill? thats pretty good ! [01:53] (I've only done a little bit of assembly in an academic setting. So it might as well have been none) [01:53] yeah :) [01:54] yeah i didnt have much assembly skill before starting either :P [01:54] im mainly a perl programmer [01:54] is it on the nano hill now? [01:54] python here [01:54] no, it was 51 :( [01:55] it was the one called Twenty Five or Six to Four [01:56] I bet that if I had it replicate itself in addition to it's current tactic, it would do reaaally well [01:57] but I'd have to figure out a way to do what I'm doing now with less commands, as it's already at 5 [01:57] I'm sure there's a way to optimize [02:02] im trying to figure out a way to write a perl program that would let me see how effective certain offsets would be [02:03] right now im just doing inline perl in the command line though because its a fairly straightforward thing, but its not that effective for really giving me the big picture [02:03] hmm [02:03] like [02:03] you might try running tons of simulations per change in offsets [02:03] and do a bunch of data crunching [02:03] perl -e 'print ((($_*410)*8000)."\n") for ( 0 .. 20 )' [02:04] well see [02:04] one problem with running a simulation is that the result from it would be relative to the warrior it was fighting [02:05] true [02:05] but what im looking for really is a combo of a replicator step and bombing step that will fill up the core in the best way [02:05] ahh [02:05] because that would do the best against the most warriors [02:05] well, run the simulations as if there were no other warriors [02:05] thats true, but how do you get results? :o [02:06] all the results you're looking for is the actual core status, as in, what spaces are filled with what [02:06] the only output from pmars is the result of a match, unless you get into debugger output, which would be a whole other ball of wax [02:06] hmmm [02:06] hmm... [02:06] in theory [02:06] you could design a warrior to just take up certain spaces, or something, and see how it fared against that? [02:06] well, I'm not sure how, but CoreWin does it [02:06] im not sure what that warrior would look like though [02:07] it runs the simulation at least, though, no results [02:07] Oh, I have that warrior [02:07] --- [02:07] wait JMP wait [02:07] end wait [02:07] --- [02:08] it's called, rather brilliantly, Wait [02:08] ohhhh... [02:08] hm ! [02:08] :) [02:08] i think i read about wait before, and i didnt know what i was reading [02:08] but that works rather well ! [02:08] just load a bunch up throughout the core [02:08] yup [02:09] run the simulation for a certain # of steps [02:09] get your results, which will be how many of the waits your warrior killed [02:09] that would work ^)^ [02:09] yeah, it wouldn't be as definitive as seeing exactly what went on during the battle, but yeah [02:10] it seems like thats a large enough task that figuring out the math would be on-par, too [02:11] heh [02:11] yeah [02:16] MSG: Quit: This computer has gone to sleep [02:18] MSG: Ping timeout: 240 seconds [02:33] I wonder if I could utilize python to make a layer on top of redcode, to simplify things a bit [02:33] you mean like [02:33] I mean, I know it can be done, but it would be better if I understand redcode completely [02:33] programming a compiler? :o [02:33] thatd be quite the task ! [02:33] well, not really a compiler [02:34] just maybe start with some well-known redcode idioms [02:34] i think theres some stuff like that out there, pmars at the least has a for loop construct kinda deal [02:34] that expands into multiple lines of redcode [02:34] give them form in python, with pythonic syntax, and then have it be outputtable as redcode code [02:35] this is probably just my yearning for a safe place, in my knowledge [02:35] will have to learn redcode complete [02:35] ly* [02:35] and not worry about having to learn, because learning is fun! [02:36] i think the hardest part about redcode is learning how it self-modifies [02:36] some stuff gets quite "O_o" [02:36] yeah, exactly [02:36] I'm still not sure -exactly- how my DoubleDwarf is doing what it's doing [02:36] like a warrior that bombs a core all the way through and then bombs itself at a certain location to then make it do something different [02:36] I just happened to get lucky, and it worked when I was trying to do something [02:37] woah, heh [02:37] hm, well, if there's a compare on the place where it bomb [02:37] like, comparing using that place, then the bomb might have a certain b value [02:37] so it could change the result of the compare easily [02:38] i just submitted a new offset version of my warrior [02:38] oh? [02:38] lets see if it can beat my old best of 57.8 : ) [02:38] :D [02:38] its competing right now on the beginner hill [02:38] on beginner's? [02:38] k [02:40] maybe I should stop hammering the server [02:40] it won't update any faster with me doing so [02:40] :P [02:41] 59! [02:43] haha [02:43] moderately better i guess :P [02:43] hello. [02:43] progress is progress [02:43] nice to see all the activity :) [02:43] heya bvowk :) [02:43] hey bvowk :P [02:43] What can I say? I'm the life of the party. [02:44] why the long tongue flyduck? [02:49] How are you, bvowk? [02:50] I'm not bad. [02:50] I really hate xmas tho. [02:50] Hah, so do I [02:50] bah humbug. [02:52] Indeed. [02:57] ok [02:57] another version submitted with new offsets [02:57] let us see how it does.. [02:58] mkay :D [02:58] i think after this i should abandon this or make a different attack method tho :P [02:58] tweak it a bit maybe [02:58] i might have kind of hit a wall, messing with offsets isnt going to double my points when i already have a fairly good set i think [02:58] yeah, offsets aren't going to change much, methinks [03:02] new one got 63.6 :P [03:02] keeps going up ! [03:02] nice! [03:03] think I'm going to take some time off to get warm and watch a movie. Be back later! [03:04] cya ~ [03:04] highlight me if you get a nice score [03:04] ;p [03:05] k ;p [03:59] Join: baudchan joined #corewars [04:00] MSG: Client Quit [04:00] Join: baudchan joined #corewars [04:00] AndrewBC, 71.5 :P [04:02] i went with a different attack strategy [04:03] smaller, but less effective, so my W% went down but i got 38.7% ties [04:08] interesting [04:08] what was the new tactic? [04:11] MSG: Quit: This computer has gone to sleep [04:12] mov 0, 1 :D [04:12] i just changed the bomb to an imp, and i also came up with a good offset [04:12] hmmm [04:13] that explains the ties [04:13] im trying something new now [04:13] which hopefully might up the wins.. [04:13] now it throws a SPL 0 along with the code [04:15] interesting [04:17] ok, trying something new and kind of crazy now, a huge SPL 0 pad that it attacks with [04:17] which then gets overwritten by the imp after the replication finishes and SPLs to the new spot [04:17] if that makes sense [04:18] hm, 69.8 [04:18] thats with one SPL [04:18] now lets see what it gets with 5 :) [04:18] :P [04:24] 69.9 [04:24] heh [04:27] niice ;p [04:27] a whole .1 points! [04:30] hrm, what the deuce [04:31] I'm doing mov, but instead of moving a mov, it moves a dat [04:31] oh [04:31] nope [04:31] * AndrewBC scratches head [04:33] oh! [04:36] aha! [04:40] ok, newest tactic [04:40] i copy+pasted an imp spiral in for the attack [04:44] probably not the best coz im not even 100% sure how it works, heh [04:44] but well see how it does :P [04:44] c JMP imp-i_step [04:44] launch SPL c [04:44] SPL imp+i_step+1 [04:44] imp MOV 0, i_step+1 [04:44] weird, huh ? [04:47] 73.5 ! [04:53] Wow! [04:53] nice :) [04:53] I just made a double-sided imp gate [04:54] in 3 lines [04:54] ---- [04:54] splode MOV.I 3, >jump [04:54] change MUL.AB #-1, 1 [04:54] jump JMP -2, 2 [04:54] --- [04:54] bad formatting! bad! [04:54] stupid IRC clients and their tabs [04:56] the problem with it being double-sided though, means that it isn't guaranteed to catch every imp that comes at it, I don't think [04:57] yeah, against a single imp, it only wins 13.6% of the time [04:57] loses 0% [04:57] and ties the rest [05:01] it wasn't what I was wanting to accomplish [05:01] I was wanting to progressively bomb outwards with fewer lines, but the only thing holding me back from doing that with this is that when I convert the number to a negative one, incrementing it adds one when I wish it would subtract one! [05:01] curses! [05:02] 74.4 : ) [05:02] :P [05:03] andrew i dont understand that code you posted because [05:03] MOV.I 3, >jump [05:03] since jump has a 2 in the b-offset, ... [05:03] oh wait, nvm :o [05:03] :D [05:04] i always forget that with indirect addressing [05:04] it uses the address of the instruction that gets indirected to [05:04] rather than the original instruction that specifies the indirection [05:04] I know, it's a bit hard to wrap my head around that myself 99% of the time [06:08] Join: nescience joined #corewars [06:26] hi nescience [06:47] MSG: Ping timeout: 240 seconds [10:09] Join: waknuk joined #corewars [10:16] * waknuk wishes the idlers (and future viewers) a merry .Xmas [10:16] Part: waknuk left #corewars [16:18] Join: flyduck joined #corewars [17:28] ugh [17:30] yesterday cousin's computer underwent a spontaneus fuckup, and i just spent 6 hours fighting with it [17:30] still havent got a clue which would be the most likely point of failure, other than the buildings powersupply, or the systems allergy towards me [17:35] What's happening with it? [17:36] yesterday, i needed to download some drivers to cousin's newer computer, so i switched older computer on; worked perfectly [17:37] next time, i get problems with the Wifi because i forgot the password -> to router to change the password [17:37] the older computer starts, and dies in 10 seconds [17:38] after start, the power supply runs for 10 seconds, then shuts down [17:38] i took the system apart, literally, and went for one-by-one check for what might be te [17:38] the cause [17:38] i switch power-supply and motherboard from other computer, and plug in the chassis [17:38] works [17:38] i plug in the rest of the parts, it works - for one boot [17:39] i start mixing and matchin parts between the two, and i get a system to the point that the power supply works, but neither display adapter gives out anything [17:40] one of the parts inbetween those two has gone reaper, think ill need to dig my dad's own set and replace stuff with it [17:40] crazy [17:41] since that was technically a left-over machine for second cousin, its not in a hurry and nobody cares if i push some leftover-parts in it, as long as MSN works [17:42] yeah [17:42] sometimes you've got to wonder if you picked the right hobby, so long ago [17:43] being the "fix my computer! it's broken!" guy can be hell [17:43] i just need a shot of Salmari, can of energy-drink and im ok again [17:44] That's the spirit! [17:44] :) [18:10] hi guys [18:32] im watching this thing on TV about toys from the 70s and 80s [18:32] so many weird things ive never heard of o_O [18:33] about the only one of these i had was a gameboy [20:01] MSG: Read error: Connection reset by peer [20:01] Join: nescience joined #corewars [20:52] MSG: Read error: Operation timed out [21:54] Join: baudchan joined #corewars [22:59] Join: flyduck joined #corewars [23:13] MSG: Read error: Operation timed out [23:49] Join: flyduck joined #corewars