Backgammon Cheating

Cheating

Using computer to aid online play

From:   Patti Beadles
Address:   pattib@mauve-new.rahul.net
Date:   27 January 2003
Subject:   Re: Jellyfish a way of cheating?
Forum:   rec.games.backgammon
Google:   b140j9$c64$1@blue.rahul.net

> I would like to know whether anyone believes that jellyfish is a near > perfect programme and whether it can be used to input moves and > therefore cheat on other sites. I don't wish to do that myself but it > is something that I was discussing with a partner the other day. Jellyfish is better than most human players in theory, and perhaps better than almost all of them in practice, since it will never overlook a play. Yes, it could in theory be used to cheat online. It's a relatively slow process to do it manually, though.

Peter Schneider writes:

1. Two acquaintances of mine who I know from my Hamburg time tried the following simple setup: One of them played on FIBS while the other one was operating Snowie on a second computer, entering the dice and opponent's moves and telling how to move. Even at 3-ply, it reportedly is as fast as if you think yourself. If you play for significant amounts of money, you'll want to think a little then and when, won't you? Nobody can tell whether your buddy feeds Snowie instead. 2. Since gnubg is available in source code, it's no problem to program a FIBS client which moves gnubg's moves without any manual intervention except the kibitzes. Examples are the gnubg bots playing on FIBS and otaku's bots. gnubg is probably stronger than JellyFish. 3. Some sites which live from money games claim to check games with Snowie to detect whether one side actually moved the bot's moves. Well, gnubg allows you to set a "noise" factor. That means, it plays generally very strong but introduces a second or third best move from time to time, as any human would do. This is afaik not distinguishable from (still very strong) human play. For all these reasons, I can only strongly discourage anybody from playing online for significant amounts of money, except against good friends. And who plays good friends for significant amounts of money? Regards, Peter aka the juggler

Bob Ebbeler writes:

It is far better to introduce the errors yourself, often using the most "human" of the choices displayed, but at other times having to make up moves on your own to appear credible.

Did you find the information in this article useful?

Do you have any comments you'd like to add?