Top 3 Match

Sep 24, 2010, 1:21 AM |

The discussions following Yelena Dembo's banning have been various and boring at parts, with a lot of technical, mathematical detail, mainly disregarding moral and legal aspects. Anyway, in order to understand some of the numbers, I decided to apply top 3 match to the otb games I'm going to play during the season 10/11. With club championship and serial play I should be able to collect about 20 games or 700+ moves. I know this is not a scientifically acceptable method. I'm doing this mainly for myself for the numbers to take on a more concrete meaning.

Game 1


I ran a top 3 match with Rybka 3 (human) 256 Mb Hash 30 sec per move: (50 moves 4 was book)

Rybka 3, first choice,     28/46 = 61%

Rybka 3, second choice, 35/46 = 76%

Rybka 3, third choice,    38/46 = 83 %

I think my near GM results is due to some part of the game being very forced. In this case there was a pawn promotion race, where 7-8 moves were forced. My opponent had 58, 67 and 76 %. But it was well played and only when in beginning time trouble he finally lost it with move 47 and 5 minutes left on the clock. A relatively fast game - 20 moves/1h + 1h for the rest of the game. I had like 45 minutes left on the clock. On the other hand an otb game, with no books and no touching pieces should generate lower correspondence with prefered engine moves. Getting near 90% seems impossible though. In retrospect I would maybe have played 2 moves differently, especially the last light piece swap, but it would not get me past 90%.

According to Steve Collyer, The Main Cheat Hunter on

"What do the results mean?

After years of evidence gathering, both with pre-computer CC World Championships and also the highest quality OTB GM games, it is know that only the very best players can get near results of

Top 1 match = 60%

Top 2 match = 75%

Top 3 match = 85%

using full methodology.  Results for a single game may be much higher than these.  It is the average over time that is the critical factor.

If you have found a suspect with total average stats like these, you should at least report for engine use and let site Admin take it further."

Game 2

This is a really sad sight. I was trying to play something totally out of character, going for a hard tactical game - I thought;) I was but a small fish against a shark this time. We leave book at move 5. Then it's wild but with c5 and/or 0-0-0 at move 14 it would have been an open ended story. I get seriously lost and after 14 Bb5+ it's curtains. So be better prepared or do not play this type of game :)

I ran a top 3 match with Rybka 3 (human) 256 Mb Hash 30 sec per move: (18 moves 4 was book)

Rybka 3, first choice,     5/18 = 28%

Rybka 3, second choice, 7/18 = 38%

Rybka 3, third choice,    9/18 = 50%