Path: ibmpcug!gwen.pcug.co.uk!dish.news.pipex.net!pipex!tank.news.pipex.net!pipex!swrinde!emory!nntp.msstate.edu!nntp.memphis.edu!nntp.rhodes.edu!hal.mathcs.rhodes.edu!graham Newsgroups: rec.games.corewar Subject: Porch Swing Message-ID: <1995Jul20.174441.2807@rhodes> From: graham@hal.mathcs.rhodes.edu (Randy Graham) Date: 20 Jul 95 17:44:41 -0500 Nntp-Posting-Host: hal.mathcs.rhodes.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 105 Well, I've played with Swing a little and played with Withershins Thrice a while, and decided that, unfortunately for me, Withers has a better design, even though I move through core faster. The problem is, my djn-stream slows down Withers and lot, and a decoy at the end slows Rude Wind II, but I still can't beat up on all others as well as Withers. So, I toyed with Swing some to figure out why she couldn't stay in first. I decided that the higher speed through core wasn't enough because - a) I only lay one bomb per loop, b) I use an even step size, c) I am one line longer than Withers in the bomb/scan loop. Well, what to do? The obvious solution was try to figure out a better bomb/scan routine than Withers. So, I changed Swing to bomb/bomb/scan like Withers, but used a different bomb/scan pattern. Now I lay a bomb high and low, and scan in between : -----+------+------+------+---- | | | | bomb scan scan bomb Don't really know why this gives a better result then Paul's bomb high and higher, scan low and lower, but the results are quite good. So, I had this idea, used it, and it did as well as, or better than, Swing on everything I tried it against. But, when I sent it to the hill, it was in the same position as Swing. Examining the scores, I found I did miserably vs. TimeScape and Marcia Trionfale. Odd, since against paper Swing did great. I can't tell why, but when I added the for EXTRA loop you see below and set EXTRA = 4, my wins vs paper shot up from 13 and 19 up to 39 and 56. Set EXTRA to 0 and get Porch as first sent to the hill with. Set it to 4 and get the current Porch. In addition, this extra dat space gave a couple more wins vs Agony II (by the way, has source for this ever been published?) and even more against other scanners. I also scan at an odd interval. At least this way a reflection can only fool Porch or Withers. This seemed to help a little against stones (the odd, smaller interval), but I doubt it will make a difference against cmp-scanners (since they are already >= the original step of 12). So, having wasted this much room explaining the way she is now, here are Porch Swing's results and source (again without EQU lines). Versus W L T ----------------- --- --- --- Withershins Thrice 62 38 0 Agony II 23 74 3 (not great, but better then Swing) Porch Swing (old) 53 47 0 HeremScimitar 51 48 1 Illusion-94 67 31 2 Juliet Storm 49 35 16 Copyright Notice 49 51 0 Rude Wind II 65 20 15 Torch t17 53 39 8 Blue Funk 3 56 31 13 TimeScape (1.0) 39 32 29 (and this distresses me) Thermite 1.0 59 38 3 Quiz 52 36 12 Iron Gate 1.5 56 41 3 (Swing was about 40/50/10) Taking Names 67 29 4 Marcia Trionfale 56 27 17 Sling 66 30 4 Miss Carry 64 35 1 Miss Treatment 70 28 2 Miss Treatment 70 28 2 (yes, there are 2 on the hill) overall score: 56/37/6 - 175 Randy ----------------------------------------------------------------- ;redcode-94 ;name Porch Swing ;kill Porch Swing ;Author Randy Graham ;assert 1 ;strategy Swing with a little wider range. ;strategy Now 80% bomb/scan with djn-stream once-thru dat.f #gate-10, step-gate+5 gate dat.f #gate-10, sneer-STEP+1 dat2 dat.f #gate-20, step-gate+5 dat1 dat.f #gate-25, step-gate+5 site2 spl.a #gate-30, step-gate+5 site spl.a #gate-40, step-gate+5 for EXTRA dat.f 0, 0 rof adder sub.f sweeper, sneer hithigh mov.i step, *sneer hitlow mov.i step, @sneer sneer sne.i @gate+STEP*6-1-EXTRA, *gate+STEP*3-EXTRA ;so we bomb step looper djn.b adder, gate swinger djn.b mover, {site step dat.f NNTP-Posting-Host: sw24-128.iol.it X-Newsreader: Forte Free Agent v0.55 Apparently-To: cbx.rec.games.corewar Having got the O.K. from Randy, I hope Paul has nothing against it, I'm publishing how I managed to score 56/42/2 wins against Porch Swing and 61/36/3 against Whitershins Thrice; all this with a warrior, Tornado 1.2, scoring well 97.2 points on the 94 hill. Whitershins, and then Porch, shocked the hill when they entered. They got really high scores and, most important, outscored nearly every other warrior with the only exception of Agony. Having to make a multi-component p-warrior, and hoping to score very high , I began thinking at how to beat these two deadly guys with, at least, one of my components. Paper, perhaps the kind of warrior I can make better, was not effective against their deadly spl-spl-dat clear, so were stones/imp (Blue Funk and Juliet scored low with them); the only warrior scoring well was Agony but I was (am) not able to make one so good and older cmp-scan were not effective against PS and WT; fast bombers like HeremScimitar and his offspring Copyright notice, scored well but not well enough; maybe we are on the right way; O.K. I need a specialized warrior able to hit them where they are softer. Let's give a short look at the code and how it works. ;redcode-94 ;name Porch Swing ;kill Porch Swing ;Author Randy Graham ;assert 1 ;strategy Swing with a little wider range. ;strategy Now 80% bomb/scan with djn-stream once-thru STEP equ 14 ;guess Randy didn't published it EXTRA equ 4 DJNOFF equ 5000 ;still gues but I don't think it changes a lot dat.f #gate-10, step-gate+5 gate dat.f #gate-10, sneer-STEP+1 dat2 dat.f #gate-20, step-gate+5 dat1 dat.f #gate-25, step-gate+5 site2 spl.a #gate-30, step-gate+5 site spl.a #gate-40, step-gate+5 for EXTRA dat.f 0, 0 rof adder sub.f sweeper, sneer hithigh mov.i step, *sneer hitlow mov.i step, @sneer sneer sne.i @gate+STEP*6-1-EXTRA, *gate+STEP*3-EXTRA ;so we bomb step looper djn.b adder, gate swinger djn.b mover, {site step dat.f gate 18-swinger djn.b mover, {site 19-step dat.f gate djn -1, {wipe1 W.T. will start clearing gate+1 then gate+2, .. the end. It's finished. I only hope, after looking this dissection of a warrior, others will continue publishing their warriors ;-) I'll do. Soon coming "How to beat Marcia Trionfale" by Randy. -Beppe Bezzi bezzi@iol.it -------- Beppe Bezzi bezzi@iol.it Gaudeamus igitur.. From news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!oleane!tank.news.pipex.net!pipex!newsfeed.internetmci.com!howland.reston.ans.net!news.sprintlink.net!cs.utexas.edu!geraldo.cc.utexas.edu!usenet Wed Sep 6 17:29:32 1995 Article: 2204 of rec.games.corewar Path: news-rocq.inria.fr!jussieu.fr!univ-lyon1.fr!oleane!tank.news.pipex.net!pipex!newsfeed.internetmci.com!howland.reston.ans.net!news.sprintlink.net!cs.utexas.edu!geraldo.cc.utexas.edu!usenet From: jwilkinson@utexas.edu Newsgroups: rec.games.corewar Subject: Re: How to beat P. Swing and W. Thrice (and die in the try :-) -Long- Date: 6 Sep 1995 14:45:31 GMT Organization: The University of Texas at Austin, Austin, Texas Lines: 52 Message-ID: <42kc6b$cdf@geraldo.cc.utexas.edu> References: <429s4o$87u@mikasa.iol.it> NNTP-Posting-Host: slip-45-3.ots.utexas.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.2b5 (Windows; I; 16bit) ;redcode-94 ;name Porch Swing ;kill Porch Swing ;Author Randy Graham ;assert 1 ;strategy Swing with a little wider range. ;strategy Now 80% bomb/scan with djn-stream once-thru STEP equ 14 ;guess Randy didn't published it EXTRA equ 4 DJNOFF equ 5000 ;still gues but I don't think it changes a lot dat.f #gate-10, step-gate+5 gate dat.f #gate-10, sneer-STEP+1 dat2 dat.f #gate-20, step-gate+5 dat1 dat.f #gate-25, step-gate+5 site2 spl.a #gate-30, step-gate+5 site spl.a #gate-40, step-gate+5 for EXTRA dat.f 0, 0 rof adder sub.f sweeper, sneer hithigh mov.i step, *sneer hitlow mov.i step, @sneer sneer sne.i @gate+STEP*6-1-EXTRA, *gate+STEP*3-EXTRA ;so we bomb step looper djn.b adder, gate swinger djn.b mover, {site step dat.f From: graham@hal.mathcs.rhodes.edu (Randy Graham) Date: 13 Sep 95 20:27:08 -0500 Nntp-Posting-Host: hal.mathcs.rhodes.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 44 Well, nothing much of interest here. But, since I said I would eventually give the EQU lines for Porch Swing, here they are (the whole thing again, in fact). she fell off the hill a few days ago. I tell you what, with PSpace, Porch and Withers got eaten up, but Agony is still going VERY strong. Now there's a warrior everyone should be clamoring to see. Anyway, here she is (including the bug that kept here from perhaps scoring higher). Randy ----------------------------------------------------------------- ;redcode-94 ;name Porch Swing ;kill Porch Swing ;Author Randy Graham ;assert 1 ;strategy Swing with a little wider range. ;strategy Now 80% bomb/scan with djn-stream once-thru STEP equ 12 EXTRA equ 4 DJNOFF equ (-426-EXTRA) dat.f #gate-10, step-gate+5 gate dat.f #gate-10, sneer-STEP+1 dat2 dat.f #gate-20, step-gate+5 dat1 dat.f #gate-25, step-gate+5 site2 spl.a #gate-30, step-gate+5 site spl.a #gate-40, step-gate+5 for EXTRA dat.f 0, 0 rof adder sub.f sweeper, sneer hithigh mov.i step, *sneer hitlow mov.i step, @sneer sneer sne.i @gate+STEP*6-1-EXTRA, *gate+STEP*3-EXTRA ;so we bomb step looper djn.b adder, gate swinger djn.b mover, {site step dat.f