Question to math expert may be Ajedrecista

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
towforce
Posts: 12812
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK
Full name: Graham Laight

Re: Question to math expert, maybe Ajedrecista.

Post by towforce »

Ajedrecista wrote: Thu Jan 29, 2026 9:05 pm...But the OP asked about winning more game pairs, which introduces a subtle difference: an engine can win more engine pairs while losing more games than wins. How can be possible? It is very unlikely to be fair, but mathematically possible: please imagine that the engine A is only able to win game pairs 1.5-0.5 against the engine B, while the engine B only wins game pairs 2-0 against the engine A, with A winning 1.5-0.5 more frequently than B wins 2-0, but not far more:
Hmmmm...

If this actually happened over a large number of games, then there would be a problem with the Elo rating system!

The Elo rating system is a statistical distribution, and thus requires independence between events. If this was happening over a large number of games, then that independence would appear to be lacking. There could be reasonable explanations why this anomaly was happening - but none of these would be compatible with independence between games.

If you were seeing this kind of anomaly, then maybe probability theory might provide a better model than one based on statistics: this could provide insight into where the dependence (or "lack of independence") might be coming from. Maybe the players were somehow adapting to each other as they played?
Human chess is partly about tactics and strategy, but mostly about memory