[01:06] Join: noath joined #corewars [02:54] MSG: Ping timeout: 252 seconds [03:01] Join: noath joined #corewars [03:21] MSG: Ping timeout: 252 seconds [03:27] Join: noath joined #corewars [03:31] MSG: Ping timeout: 252 seconds [03:38] Join: noath joined #corewars [05:44] MSG: Ping timeout: 252 seconds [05:45] Join: Caelian|w joined #corewars [10:03] Join: Core_new_ joined #corewars [12:33] MSG: Ping timeout: 252 seconds [12:36] Join: Metcalf joined #corewars [12:36] Hi :-) [12:45] h M [12:50] MSG: Ping timeout: 252 seconds [12:50] Join: Metcalf joined #corewars [12:53] damnit, i made a working behemot-like bomber, and it sucks [13:11] Hi Mizcu [13:11] What's wrong with it? [13:13] bad scoring, doesnt stunn well enough [13:14] :-( [13:14] Does your code have an airbag? [13:14] yes [13:16] brb [13:36] tell me met, do you prefer clustered (spl/mov) or separated (s ddddd m) incendiaries? [13:47] MSG: Ping timeout: 252 seconds [13:51] Join: Metcalf joined #corewars [13:57] Join: OoS joined #corewars [13:58] MSG: Ping timeout: 252 seconds [14:07] I prefer seperated, but clustered can work well too [14:09] spl/mov ddddd mov would work too [14:09] and about separated: "filling" (spl #m , mov.i s, {s ) or "expanding" (spl #m+1, mov }s ) [14:09] I normally test both [14:09] thats why i dislike incendiaries, too many options, many which work differently against different opponent [14:09] s [14:09] expanding can get out of control and wipe too far [14:12] that behemoty i have currently used sm.....j....d [14:12] with bigger gaps, obviously [14:21] and one more: incendiaries usually need a moment to "cook", which usually a warrior cannot afford.. [14:22] AAAUGH [14:22] stupid me, stupid me [14:25] Join: gurgeh joined #corewars [14:25] greetings, gurgeh [14:25] greetings, Miz [14:26] Hi gurgeh [14:26] Nick Change: gurgeh changed nick to Nenad [14:26] * Nenad streches [14:26] haven't been here for more than a month or so... [14:27] what's up? any tournaments going on or anything like that? [14:28] Hi Nenad :-) [14:28] Nothing at the moment [14:29] no wonder the warrior had some difficulties, i had end warrior instead of end start [14:30] I've done that before [14:38] it still sucks against modern warriors, but its better now that the airbag is actually implemented [14:39] :-) [15:12] now i have a really big small problem [15:12] if, after airbag fails, i have jmp clear, more processes end up in the clear, speeding up the wipe, but it remains an open pointer for imps [15:13] if i use something to extent of jmp clear, {0 , then some processes (own or imp) might end up on incendiary in the space between clear and warrior [15:14] and since i use expanding incendiary, the battle will end up in a tie [15:15] i founded a solution, but it scores worse.. [15:21] http://koti.mbnet.fi/mizcu/behemoty.red - make of it what you will [15:23] 106 wilfiz, 132 wilkies [15:36] (how come the silence?) [15:43] MSG: Ping timeout: 252 seconds [15:44] Join: OoS joined #corewars [15:44] ok.. it scores better with mov.i step-2, 2*step-2 [15:50] MSG: Ping timeout: 252 seconds [16:03] Join: OoS joined #corewars [16:11] Join: Core_new_ joined #corewars [16:12] bah, maybe it just cant get any higher without structural changes when using clustered incendiary [16:12] (no reason to brute-optimize steps, wont make notable difference) [16:15] Have I missed something Mizcu, or are the 3rd and 4th lines of the stone loop the wrong way round? [16:16] it works the both way - but the reason is to get a dat-bomb moving [16:17] because if the switch is done with self-hit, a dat-bomb is lot more friendly trigger than an incendiary [16:17] ironically, with those steps, the switch is done with overwriting the j [16:18] Does the jmp bomb jump to the spl or the mov? [16:18] mov [16:18] which, surprisingly, scored better when i checked the last time [16:19] Wouldn't the spl be better? [16:19] benchmark disagrees by couple points [16:19] why, i have no idea [16:21] most scores lost with the spl-jump is from papers [16:26] yes, actually it does score slightly better with 3&4 swapped [16:27] :-) [16:27] but only by little - what is lost from stones and papers is only barely bettered by the points from scanners [16:27] :-( [16:32] MSG: Ping timeout: 252 seconds [16:32] Join: OoS joined #corewars [16:37] Met, do you remember as i once though about wiping opponents with jmp 4000 | jmp 4000 ? [16:38] Yes, I remember [16:38] well, i made an HSA-stylish, mirrored on-axis scanner with that wiped with spl 4000 | spl 4000 [16:38] Is it any good? [16:38] it _almost_ worked, but the end-game switch turned next to impossible [16:38] :-( [16:39] you cant use HSA-like switch, as you never scan yourself because of the mirror [16:39] and using djn's had some weird voodoo-bug which i dont recall right now [16:39] but it might be worth trying to recode someday [16:42] (Nenad, if you are awake, here is a pop-quiz for you, why did i think spl 4000 | spl 4000 was better in some circumstances than just spl #0 | spl #0 ? can you think of anything?) [16:43] * OoS starts implementing Gnome Sort in redcode! [17:00] MSG: Quit: noath [17:58] OoS: i just had a thought; one problem i had with the onaxis scanner was the end-game: swapping to a dat-bomb didnt really work since wiping didnt start little before the found location [17:58] i switched into lame spl/mov/dat, which worked but not well [17:58] lot better results could be gained it if were possible to mutate the wiper into bishot-clear [18:00] MSG: Ping timeout: 252 seconds [18:00] Join: OoS joined #corewars [18:00] eg. Recon2 -style [18:34] MSG: Ping timeout: 252 seconds [18:37] Join: OoS joined #corewars [19:03] MSG: Quit: Trillian (http://www.ceruleanstudios.com [19:16] MSG: Ping timeout: 252 seconds [19:19] Join: OoS joined #corewars [19:27] MSG: Ping timeout: 252 seconds [20:38] MSG: Read error: Connection reset by peer [21:01] Join: fiveop joined #corewars [21:29] Join: Caelian joined #corewars [21:51] discovery: leap in the dark -style oneshotting doesnt work as well with back-wiping clears [21:52] tried to use my scanning-ssd-clear thatway, and it works worse against some opponents [21:55] but it does beat the shit out of returnofthepapers