Cheating in chess has been an occasional subject for this blog, but the focus has generally been on the dangers present in Swiss system events.
Tennis star Lindsay Davenport, however,
has expressed concerns concerning the Association of Tennis Professionals' plan to introduce round-robin events in the 2007 men's season.
Perhaps it's worth visiting the question on here: should professional chess revert to the knockout system, at least for most events? That would take care of all the BAP business, if nothing else! (No more BAP comments, please.) Four arguments in favor of the round-robin come to mind, which I'll present and respond to - not equally persuasively in each case.
(1) I imagine that one historical consideration was the worry that the two best players might face in an early round, causing someone who might easily have come in second or at least a high place to finish near the bottom.
Reply: With the rating system, it's easy enough to create a seeding system that will render such a result unlikely.
(2) There's an argument from the fan's perspective: it's a lot more interesting to watch many games than just a few at the end of the event - especially if the leading fan favorites have been eliminated.
Reply: One nice way of avoiding this problem is to use the system employed earlier this year at the World Cup (and elsewhere), which is to keep all the players, or all the players after a certain number of rounds in the event, playing further knockout matches for their particular placement. Thus in a 16-player event, the 8 first-round losers will have their own knockout sub-event competing for 9th-16th places, those in lose in the second round as well compete for 13th-16th, etc. As many players as the organizers want will continue to play!
(3) The round-robin system allows one to overcome an "accident". Mikhail Tal was famous for his first-round losses, including most notably the USSR Championship that got him into his first, and ultimately successful, world championship cycle; and in the Candidates' Tournament that was the final stepping stone on the way to playing Mikhail Botvinnik for the title.
Reply: First, the seeding system makes it unlikely, though not impossible, that such an upset will take place. Second, there will be color equalization, so if a Tal loses to an unknown, he'll have a second chance, and that will give him a reasonable opportunity to overcome the "accident".
(4) The knockout system doesn't correlate as well as a double-round robin when it comes to matching up the winner with highest rated player. In other words, if the goal of an event is to reward the best player, and the best player is most likely the player with the highest rating, then the best system is the one that is most likely to see the highest-rated player win.
(By fair means, of course. This would rule out a pairing system that puts the top seed into the final, where he awaits the winner of, for example, a very long knockout tournament, and forces that player to change locations for the final match. It goes without saying that this could never happen, especially not for a world championship event in the late 1990s. Perish the thought.)
Reply: Hmm...one possibility is just not to worry about it, on the grounds that the other advantages of the knockout system outweigh this problem. A supplementary suggestion is to avoid knockouts for really major events, like national and world championship contests. Another possibility is to use a double-elimination system, but I leave it up to the statisticians to say just how effective that is.
At any rate, I'll finish here and open the floor to discussion. Should we listen to Davenport, or is the ATP properly following the chess world?