From news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!in2p3.fr!oleane!plug.news.pipex.net!pipex!tank.news.pipex.net!pipex!newsfeed.internetmci.com!in1.uu.net!news.new-york.net!news.stormking.com!mpn@ifm.liu.se Mon Mar 25 11:22:51 1996 Article: 4513 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!in2p3.fr!oleane!plug.news.pipex.net!pipex!tank.news.pipex.net!pipex!newsfeed.internetmci.com!in1.uu.net!news.new-york.net!news.stormking.com!mpn@ifm.liu.se From: mpn@ifm.liu.se (Magnus Paulsson) Newsgroups: rec.games.corewar Subject: myConfuser Date: 23 Mar 1996 07:30:37 -0500 Organization: Storm King Ind. Inc. Lines: 96 Sender: server@news.stormking.com Distribution: world Message-ID: <199603230912.KAA22995@ifm.liu.se> Reply-To: mpn@ifm.liu.se NNTP-Posting-Host: valhalla.stormking.com Originator: corewar-l@stormking.com Here is myConfuser! The name comes from a totally different warrior that was designed to confuse pspace switching programs. The idea of that program was to use a qscan that found the opponent 1/3 of the battles, when the qscan found something it would bomb it and then launch Timescape, if it didn't find anything it would launch a version of myVamp. The idea was that the randomness of the strategy would confuse pspace wariors. It wasn't successful. Then I had a look at the hill and there was Clisson and there was Frontwards. The idea to scan for bombs and then put a small stone over the bomb (a safe place in core) had occored to me earlier but it is only succecfull against some stones and bombers and looses to to scanners and paper. But... when you scan for bombs you sometimes find other things ... myConfuser at least survived a few rounds on the hill and made it up fifth place (or was it second?)... The reason it fell from the hill is that it looses realy huge against some programs on the hill (Naced dancer) Since I've got a Mac SE/30 and macPmars 1.99a (which has pspace) is a bit unstable on my computer I haven't made many pspace programs. I think that myConfuser could be used in a switching program with some success. There are some things in the programs that I haven't optimized fully, the step and the direction of the scan, the killer that is placed over bombs (maybee Modron would be good?)... -Magnus ;redcode-94 ;name myConfuser ;author Paulsson ;strategy Confuse pspace warriors by being good against all ;strategy types of wariors (except evolv cap type) ;strategy The first myConfuser used a qscan that only got a hit ;strategy 1/3 of the times and used this as a random factor to ;strategy change between different strategies. ;strategy Now it's a bit of clison type killer + something else. ;strategy Submitted: March 96 ;assert CORESIZE > 1 ;kill my org start+1 st equ 17 ;step size dir equ (-1) ;direction of scan -1 <-> ;backwards decoy dat #dir*150,#dir*150 step spl #dir*st*2,dir*st*2 start add.f step,ptr ;scan 80% sne.i *ptr,@ptr add.f step,ptr ptr sne.i mitten+150*dir+dir*st,mitten+150*dir+dir*2*st nr djn.b start,#(8000-500)/st/4 ;might be changed to ;jmz.f start,cptr mov.i @bptr,>cptr bptr djn.b -2,{bmb1 for 20 dat.f 0,0 rof mitten killer jmz.b clear,nr ;place killer over bomb mov.i >bootptr,>ptr djn.b -1,#4 sub.ab #4,ptr bootptr jmp @ptr,kill kstep equ 1028 kill spl #kstep,-kstep ;cannonade style stone add.f -1,1 mov.i >0,-2 djn.f -2,<-100