Article 2295 of rec.games.corewar: Path: hellgate.utah.edu!dog.ee.lbl.gov!agate!howland.reston.ans.net!cs.utexas.edu!not-for-mail From: d91andiv@und.ida.liu.se (Anders Ivner) Newsgroups: rec.games.corewar Subject: Killer instinct Date: 16 Nov 1993 20:03:17 -0600 Organization: UTexas Mail-to-News Gateway Lines: 289 Sender: daemon@cs.utexas.edu Message-ID: <9311162204.AA25033@styr16> NNTP-Posting-Host: news.cs.utexas.edu (First of all, my local newsserver is down, so I'm posting though rec-games-corewar@cs.utexas.edu, hope everything works OK.) One week ago, I was completely pushed off the hill for the first time since IMPire fell off. This really pissed me off (Watching the struggle on the hill is quite addictive :-), and the following nights, more than a few hours were spent testing and discarding ideas. Anyway, my work paid off, and after a few days of glory at the top, the details are about to be revealed: (tadaa!) - Anders Ivner presents - Killer instinct - (Hey, I'm about to give you the source to a warrior that holds a clear 1:st place on the hill, I think I _deserve_ to waste bandwidth with my bullshit :-) *** The missing strategy lines: ;strategy Behave like a bomber against scanners, ;strategy like an anti-vampire against vampires ;strategy and like an anti-imp-paper against everything else. This kind of strategy has been mentioned several times in the past, P. Kline has something similar in Vagabond (I think), J. Layland's FlyPaper is also close (but it uses a small scanner, instead of a bomber against scanners), there were articles in TCWN on a related topic, and every now and then there's an article here in r.g.c complaining that it's impossible to make a successful and 'intelligent' warrior under the '88 standard, with 'adaptiveness to the environment' as a criterion of intelligence. *** How it was implemented: loop add const, 1 mov <-99, 2 const jmz loop, ; Mon, 17 Jun 1996 21:39:29 +0200 (MET DST) Received: from su2-1.ida.liu.se by und.ida.liu.se (5.65b/ida.minimaster-V1.0b6d5) id AA08348; Mon, 17 Jun 96 21:39:28 +0200 From: Anders Ivner Received: by su2-1.ida.liu.se (5.x/ida.slave-V1.0b6d6S2) id AA09900; Mon, 17 Jun 1996 21:39:27 +0200 Date: Mon, 17 Jun 1996 21:39:27 +0200 Message-Id: <9606171939.AA09900@su2-1.ida.liu.se> To: damien.doligez@inria.fr Subject: killer intinct Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: R I wondered why 'killer intinct' didn't appear on mt olympys, and then I saw that I hadn't published the step-size for it. Kind of silly, really. Anyway, the step is 3044 so the equ should read step equ 3044 (I'm curious as to how it would do :-) /Anders