From news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!howland.erols.net!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!news.idt.net!nntp.farm.idt.net!rotten.com!franz Thu Feb 20 11:54:49 1997 Article: 7110 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!math.ohio-state.edu!howland.erols.net!cam-news-hub1.bbnplanet.com!news.bbnplanet.com!news.idt.net!nntp.farm.idt.net!rotten.com!franz From: George Newsgroups: rec.games.corewar Subject: ADD/ADD scanning Date: Thu, 20 Feb 1997 02:33:13 -0700 Organization: IDT Lines: 65 Message-ID: NNTP-Posting-Host: 199.245.121.178 Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Sender: franz@rotten.com well I'm still up to trying to do the perfect multiprocess scanner but i'm getting too many byproducts ... like now ... I just had an idea for a better (faster) f-scanner ... I don't know if anyone has done this. it's .66c, so it's as fast as cmp scanners ... but it's actually a jmz scanner ... i've done a basic one shot in 6 instructions +1 bomb ... with this engine and it seems to be doing quite well ... it got 82 (i think) on the regular hill and actually got into the bottom of the beginner hill ... it's also 5th on the tiny hill under the name XADD ... So I gave it a name and I thought i'd post it ... has anyone done this before ... i don't think i've seen it anywhere ... it's main advantage is against cmp scanners .. being as fast but smaller, AND it instantly damagaes during scan, since it adds two things in the memory ... so it's got a SLIGHTLY faster reaction time ... it's disadvantage is that it's a one shot and that it's only using a modified dclear ,, which uses the afield which is not as effective against imps ... this code itself could be greatly improved while keeping the oneshot/dclear strategy, but i'm posting this as a new scanning concept ... the only disadvantage is i don't see it extendible to 80%c unless perhaps combined somehow with a cmp scan (i don't see how, i haven't thought about it ... it just came to me) otherwise it's fairly similiar to normal scan in it's effectivness (i think) I have several ideas on how to improve on it generally but i'm not gonna post them ... I don't spend too much time redcoding lately (well i did the past few weeks trying to get a multi-process-scan running) so try this in your scanner see if you see a difference in scores .... :) ... this beats rave about 70% of time ;redcode-94 ;name Archer ;author Franz ;strategy 66% ab scanner ... with some instant damage ;assert CORESIZE==8000 sstep equ 3510 stscn equ (scan-1+sstep) org scan loop add.f step, scan scan add.f stscn,stscn+6 jmz.f loop,@scan step spl #sstep,sstep mov bomb,}scan djn.f step+1,}scan dat 0,0 bomb dat 10,10 end