Greengard's ChessNinja.com

Game and Match

| Permalink | 15 comments

It's interesting that the Olympiad uses board points instead of match points. (I explained this a bit in an Olympiad report.) It makes things more dynamic and could encourage more aggressive play, because beating a team 4-0 is much better than 2.5-1.5.

One downside is that when combined with Swiss pairing it creates weird cascades. With 20-30 teams just a few board points apart, a single big match win changes everything and it cascades up the table as in an individual Swiss. This can really distort things if it happens in the final round. Jumps of 20 or even 30 places aren't unusual. For example, Malaysia busted up Lebanon in the eighth round. That big score put them up against Turkey, who duly crushed them 4-0. That shot #69-ranked Turkey up in the pairings, where they got stomped by Slovakia 3.5-0.5. That puts #32 Slovakia up against #9 Poland today. #5 India has #43 Canada (Canada wiped out a stronger Macedonian team to jump up, so it's not always a chain of mismatches.)

Another aspect of Swiss pairings is that in a long event the top teams will have faced each other long before the end of the event. But there are so many strong teams now it doesn't matter so much. That underdog Slovakian team includes veteran GMs Movsesian and Ftacnik!

The US Men came through 2-2 against Ukraine yesterday and now have Russia. Some reward. The US has a history of "playing up," doing well against the top teams while struggling against teams they should beat. They are so balanced they should be beating weaker teams on boards three and four consistently, but it never seems to work that way. The US doesn't have a 2700, but only Russia and Ukraine have stronger reserves.



Movsesian is only 25 years old.

I know, I've met him several times. Nice guy. I just meant experienced, not old. I think he's been a GM for eight or nine years already!

The point Mig is making is called "voting paradox", which means that different voting rules lead frequently to different results. The fact is that there is no method to elect the best team or the best player that is garanteed to give the best (whatever the definition of best) result.

What's the difference between board points and match points?

As a Canadian chessplayer, I am proud of our team and the results they have achieved this year. They are having an excellent tournament, showing that they can compete with many of the upper echelon teams, and no doubt their 43rd ranking is not reflected in their performance.


If the tournament is scored by match points, you get one point for winning a match whether you win 4-0 or 2.5-1.5

If the tournament is scored by board points, you get 4 for winning 4-0, 2.5 for winning 2.5, etc.

So with board points, you can jump up by beating your NEXT opponent 4-0 while the previous leader wins by only 2.5

I wonder if it would make sense to determine the final standings by game points (to encourage dynamic play), but to pair the teams according to match points (to avoid the cascades Mig describes).

Not to drudge up the ancient past Mig, but don't you find it interesting just how much effort went into getting Anna Hahn off of the US Women's team, only to replace her with a reserve who has played just two games, scoring a half-point? It boggles the mind that some people believed having her on the team would eliminate the USA from medal contention, while her replacement (Jennifer Shahade) has been reduced to a glorified spectator.

Good point, OK, definitely worth a mention. Shahade, like Hahn, is significantly lower-rated than the other US players and it's predictable she wouldn't see much action as long as the top three played well and as long as the US was in medal contention. (Zatonskih was on fire, so rotating wasn't likely.) Hahn's removal was about power and the prestige of the training squad, and only a tiny bit about Olympiad success.

I'm most disappointed for Jen. Despite all the silliness that went on, she is the current US women's champion and deserves more than two games with black, especially with her sharp style. Jen usually plays well against strong competition and would no doubt show her stuff with a few whites.

Zatonskih appears to have tired in the last few rounds after a terrific event. Her loss was after refusing a quick repetition draw with black, a curious decision perhaps made by the captain. It also came after two marathon draws in a row. It shows the danger of riding a fast horse too long. The coaches are experienced, but it's better to give rest before a bad result or two shows how needed it is.

I really liked Flaneur's pairing idea at first, but it doesn't work. It would mean the stronger teams would face much stronger fields for their points, a sort of artificial parity. Weaker teams could lose four or five matches but wipe out weak teams with big scores and finish high without ever facing a top team. I think. Running a model is probably necessary to verify, but that's what comes to mind.

I agree with Mig's points about the problems with using board points and a Swiss pairing system. Board points result in big fluctuations and unfair pairings, and they mean that only the final rounds are really decisive (at least for the secondary placed teams). The difference between 0-4 and 4-0 in the final round will be enough to send a team from a place around #10 to around #60 with the current scores. Or from #2 to #25. And of course, the leading teams meet up early in the tournament, so the final rounds are just about who can beat up the underdogs the most.

The best solution, IMO, would be to go back to the old system with preliminary rounds and finals. A suggestion with the current number of teams could be 16 groups of 8 teams each, from which the winners met in two semifinal groups, from which the winners again met in the grand final. All the other teams would meet in finals according to their placements. This would be a more fair system, and provide excitement for all teams throughout the tournament.

An even simpler way could be swiss pairings for the first 10 rounds, then the best 6 teams met up in a final round robin (perhaps taking the points with them).

Or, of course, the current system, using match points instead of board points. But that would probably spoil the excitement quite early on.

Board points reward wins. In a match point system, 1 win and 3 draws earns exactly the same score as 4 wins...so switching to a match point system would see a lot more cautious play.

In these days when so many people suggest we need to revitalize the fighting spirit in chess, I wouldn't tamper with one of the few events that legitimately encourages every player to play for a win.

One more thought...scoring by board points eliminates much of the possible friction between someone trying to get a medal on board 3 and a team captain trying to get a match win, since it changes the risk/return ratio for the team overall.

Xereos is correct that Canada has outperformed their ranking. If the teams were arranged by performance rating going into this final round, Canada would move up 19 places! This represents the worst pairing luck in the Olympiad. They have had to play UP 11 of the 14 rounds. On the other hand, Norway has had the best luck. They are 17 places higher than performance would predict, and have played DOWN 11 of 14 rounds.

On a positive note, the current top 10 include the 10 best performers.

Does anyone know what happened to light a fire under the US teams (both men & women)? They seemed to play the first few rounds like they were going through the motions, making a lot of draws against weaker teams/opponents and then took off. I know they essentially benched their weak links, but that didn't seem to explain the whole story. Polgar was especially impressive, finishing 5.5/6.

Dunno about fire, but having experienced (re: Soviet trained) coaches and captains is no small thing. I thought they blew it by not using Shahade at all, but it turned out okay in the end. Impressive stamina by Krush and Zatonskih. Polgar we know is in a separate weight class from most of her opponents (dangerous using that term about a woman player!).

After a certain point fairly early on, successful teams end up facing weaker opponents because they've played the top teams already. Susan's four wins in a row to finish were against an average rating lower than her opponents in rounds 2-7, for example. But still very impressive of course. Remember that this was her first international experience in a long time; she had to get warmed up.

Twitter Updates

    Follow me on Twitter



    About this Entry

    This page contains a single entry by Mig published on October 26, 2004 7:53 AM.

    Look, a Kitty! was the previous entry in this blog.

    Bring Back the K's? is the next entry in this blog.

    Find recent content on the main index or look in the archives to find all content.