Author: Gian-Carlo Pascutto
Date: 02:40:03 07/16/02
Go up one level in this thread
On July 16, 2002 at 05:08:57, Russell Reagan wrote: >I think there are quite a few good ways to deter cheating in an online event. >You could require that the engine have been made public for 6 months prior (or >some amount of time), and check it against Crafty or other strong open source >engines. This is totally unreasonable for commercial or experimental engines. Moreover, an engine can and will change A LOT in just 6 months. Not being able to play with your latest and greatest would be a major pain. >Surely working together we could devise a nearly automated system for >such a thing. Maybe comparing binary strands in the executable, comparing move >selection, and so on. I'm sure there are people who would be able to offer lots >more methods who have done this kind of thing in the past. >As for someone firing up Fritz and competing, I think it wouldn't be >unreasonable for the organization running the event to have copies of all of the >decent commercial engines, and compare outputs. It would be very easy to cheat and not be detected by such a naive scheme. >I think that if someone is dedicated and talented enough to create a modified >engine and work on it enough so that there are no similarities between it and >it's original engine that they either modified it enough to call their own, or >they are talented enough at code hacking that they are going to be able to cheat >in person as well as online. I'm sure there is some talented person out there >who could hack Fritz and use it at an event in person. The problem is that in a real event you have a just as talented person in front of you keeping a very close eye to what's happening on the screen. >Do you think that if the organizing group had some method of detecting clones >that there would be any more room for cheating in an online event rather than in >an "in person" event? If some dark horse comes out of nowhere to win, don't you >think people will start to wonder and look into it? Sure, but at that point it's already too late. >Regardless of all of that, >they would still have to compete int he semi-final match and finals match to win >anything. There could be more rigorous testing done in person, and if a cheater >is caught in person, then the next highest finisher gets bumped up and takes his >or her rightful spot. Again, too late. Moreover, you could simply not cheat in the live games. Qualifying for the semis might be enough. >Perhaps you could simply throw out all of those games and >reseed. Some acceptable method could be developed for handling this. I just >think that if someone is talented enough to either modify the source enough or >the executable enough to make it not detectable from a binary comparison for >similar binary patterns, and to alter it enough to make it play unique moves, >then that person won't be caught online or at WCCC or anywhere else. There are many ways to cheat which don't involve something as complex. For example, overriding your engines time-usage. Happened in the latest CCT! If done at crucial moments it can be decisive. Try pulling it off in a live event. Much harder. -- GCP
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.