Chess

LAST WEDNESDAY I reported on the recent Petrosian Memorial tournament in Mos-cow, where the 10 gentlemen must surely have set a record by drawing a staggering 42 of 45 games (at an average of 26.78 moves a game).

Intrigued, I spent a few hours investigating further (the results appear below). I looked at a few recent tournaments and, with the help of Bob Wade, a selection of the century's great events.

In the following list, the first figure after each tournament refers to the percentage of games drawn, the second refers to the average number of moves per game:

St Petersburg 1909: 32.75 40.77

New York 1924: 34.55 48.04

New York 1927: 60.00 39.20

Nottingham 1936: 40.00 40.76

AVRO 1938: 57.14 39.73

World Champ 1948 38.00 42.52

Zurich Candidates 1953:

56.19 39.00

Piatigorsky Cup 1963: 50.00 42.04

Santa Monica 1966: 61.11 40.78

Montreal 1978: 57.78 36.62

Reykjavik World Cup 1988: 62.09 33.95

Linares 1990-1999: 46.13 43.16

Hastings 1998-9: 55.56 40.44

Katrineholm 1999: 42.22 39.64

Malmo 1999: 46.67 37.42

Wijkaan Zee 1999: 54.95 35.81

The total number of games was 2062 with 48.59 per cent draws and an average length of 40.77 moves. You would need much more data to do anything statistically significant, but I certainly wouldn't want to inflict that on you, my readers.

There are some interesting points, however, especially relating to Senor Rentero's Linares tournaments, of which I've lumped together the eight between 1990 and 1999 (there wasn't one in 1996). Sr Rentero has a legendary hatred of draws and stipulates in many contracts that (unless there's a decisive result) players must go on until at least move 40 on pain of losing appearance money: as you can see, this has led to slightly longer games and a lower percentage of draws.

I also examined Chessbase of Hamburg's complete Mega '99 database, which contains just over 1.1 million games, including many open tournaments. This bumps up the proportion of decisive games, with just 34.26 per cent drawn, while the average length is 38.41 - lower than most of the closed tournaments listed, presumably because a lot of open games are mismatches that end quickly and decisively.

Reshevsky v Geller

Zurich 1953

To counterbalance any implied criticism of draws, a nice finish from one of the great Candidates tournaments: Geller soon drew after 53 ...Rf3+! 54 Ke2 Rxg3; of course, 54 Kxf3 would have been stalemate.