From news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!howland.erols.net!feed1.news.erols.com!insync!news.azstarnet.com!news Mon Oct 21 11:26:50 1996 Article: 6306 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!howland.erols.net!feed1.news.erols.com!insync!news.azstarnet.com!news From: Franz Newsgroups: rec.games.corewar Subject: Re: How to improve? Date: Sun, 20 Oct 1996 18:19:55 -0700 Organization: Starnet Lines: 164 Message-ID: <326ACFBB.56BF4564@azstarnet.com> References: <54ef3b$h86@news2.texas.net> NNTP-Posting-Host: usr14ip39.azstarnet.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; I; Linux 1.2.13 i586) > I've thought of upping the increment to 2 to get through core quicker > Thanks in advance nah that will be even worse, it might create a lot of bombers from silks instead of killing them ... and it won't stunn as good .. here's what i thought of doing on the first sight: > ;name Everclear v 0.1 > ;author J.A-Worth > ;strat 2pass clear > ;assert 1 > org start > dat #100, #4 make this spl #0 and make the gate an increment > start spl 0, <-50 > mov @-2, }-2 increment gate and increment a and b field > dat <-53, <-51 > spl #100, #5 kill this dat not needed (actually cmps canners will find this! > dat #0 so after this it would be, also i'll make the gate closer so that imps that start real close can't get through org start dat #100, #4 start spl #0, >-20 mov @-2, }-2 dat >start-20, }start-20 spl #100, #5 next we want to do something about that mov statement make it anti imp too ... lets change the position of the dats and switch a and b fields so that now it will use the b field for the pointer of where to shoot org start ;the imp gate gate dat 0,0 dat 0,0 ;where to bomb where dat #sbomb,#100 ;buffer between gate and warrior for 20 dat 0,0 rof start spl #0, >gate mov *where, >where dat >gate, }gate ;decreased the forward start for the dat round sbomb spl #(start+2)-where,#50 this will do one spl and one dat round ... and then ... (i don't think it will kill itself (too much to go into) but it's not as effective as we want so lets make it spl/spl and then as many dat clears as we need org start ;the imp gate gate dat 0,0 dat 0,0 ;where to bomb where dat #sbomb,#100 ;buffer between gate and warrior for 20 dat 0,0 rof start spl #0, >gate mov *where, >where dat >gate, }gate ;decreased the forward start for the dat round sbomb spl #s2bomb-where,#50 s2bomb spl #dbomb-where,#40 dbomb dat #dbomb-where,#30 ok this looks pretty good but still the speed is .33c (about) ... so we can change the gate type and make it around .5c org start ;the imp gate gate dat 0,0 dat 0,0 ;where to bomb where dat #sbomb,#100 ;buffer between gate and warrior for 20 dat 0,0 rof start spl #0, >gate cc mov *where, >where jmp cc, >gate ;decreased the forward start for the dat round sbomb spl #s2bomb-where,#50 s2bomb spl #dbomb-where,#40 dbomb dat #dbomb-where,#30 ;it's better to include an end so that MARS doesn't get ;confused if there is anything afetr the warrior end this should be quite a good warrior I tried it and it should work. this should be quite good againts silks spl/spl if it catches them fast enough you can make it .66c by doubling the cc line but adding to teh footprint of the warrior ... if your warrior is only a coreclear it would be better to do that ... so here would be the final improved version in .66c ;name Everclear v 0.2 ;author Franz ;strat spl/spl/dat clear ;assert 1 org start ;the imp gate gate dat 0,0 dat 0,0 ;where to bomb where dat #sbomb,#100 ;buffer between gate and warrior for 20 dat 0,0 rof start spl #0, >gate cc mov *where, >where mov *where, >where jmp cc, >gate ;decreased the forward start for the dat round sbomb spl #s2bomb-where,#50 s2bomb spl #dbomb-where,#40 dbomb dat #dbomb-where,#30 ;it's better to include an end so that MARS doesn't get ;confused if there is anything afetr the warrior end have fun Franz