Saturday, October 9, 2010

Dubai 1986 our best performance!

Hairulov have a flawed argument -Dubai 1986 our best performance? - regarding one of my statements in my olympiad review.

There are so many other factors that can be considered when determining which Olympiad we did the best.

Hairul argues that there are many more strong countries especially after the break-up of the Soviet Union. The case seems to be that because of addition of more strong teams a 42nd placing in 1986 is not worth a 42nd placing in the present. This is a fundamentally flawed argument.

The fact is that there are also weaker countries participating now than in 1986. A quick glance shows that Nepal, Monaco, Korea  ,Chinese Taipei, Aruba  ,Jersey ,Namibia  ,Malawi , Ethiopia      ,Guernsey   , Surinam , Macau ,Mozambique , Madagascar ,Cameroon ,Sao Tome and Principe , Sierra Leone, Burundi  ,Rwanda were not at Dubai and arguably they are weaker than us.

Then lets look at rating performance. Dubai performance was 2337. There is another team that bettered this which is Turin 2006 which performed at 2405 - just 68 points better. But have we considered rating inflation? I would say a rating in 1986 is worth another 100 points today. Which would make the Turin performance below Dubai.

A world championship candidate in 1980 would have a rating in the mid 2600s. For many years only Fischer could exceed the 2700 barrier. Today there are quite a number of 2700 players and the barrier is now 2800. Read this article regarding rating inflation. and you might start thinking that my 100 points is too conservative.

IM Lim Yee Weng also mentions that the Turin team had excellent results winning against stronger teams like Canada 3.5-0.5 (!)  and Switzerland 2.5-1.5. However, a tournament is not an single match but a series of matches. I am not talking about best performance in a single round. I already mentioned that Malaysia assembled the strongest team ever in 2006. Luck also plays a part as in Turin they were unlucky to meet two strong teams in the last two rounds and could only manage a single point from the eight games.

Finally Hairul poses the question, how would the  Dubai 1986's team final ranking be if they were to face the post -Soviet & Yugoslavia countries. Let me ask a question, how would Morphy do in a match with Fischer? Or Capablanca in a match with Carlsen? Morphy was the best compared with players in his era and similarly Capablanca was the best in his time. You cannot make this comparision and you cannot compare how a team in 1986 will perform in a tournament in 2010. The only comparision is how they performed against in the company of their peers. And that actually is what the Olympiad is about.


The Chess Ninja said...

Don't you think you're being a little unfair here Jimmy? First you compare the ratings of people in 2006 and in 1986 for inflation. Then at the end, you said you cannot compare how the players in 1986 would fare against the players in 2006.

To define rating inflation properly before I expound my argument, let us hopefully agree that rating inflation means that a player with rating X in 2006 is not as strong as a player with rating X in 1986. I got this information off the "rating inflation" link you provided. I hope it is fair.

Now, focussing on the reference to "playing strength", I would like to bring your point up about not being able to compare players from different generations. You must compare them against their peers. So how can you say that a player rated X in 1986 is better than a player rated X in 2006? So then, your argument to deduct the 100 points in performance from the 2006 team is flawed.

Perhaps it would be fairer to look at the percentage difference between their performance rating and their actual average rating. So for Team 1986, their average rating was 2271 with a performance of 2320. That means they performed 2.2% higher than their rating. Team 2006 was rated at 2331 and performed at 2405. That means they performed 3.2% above their rating. It seems to me, Team 2006 has a better performance :P

Don't you agree that this percentage change theory accounts for rating inflation? Besides, following your rating inflation argument, the Malaysian team's rating would be inflated as well. Hence, using rating differences would reduce this effect of inflation as you call it. After all, as you suggested, you must compare the team with their peers.

This is a purely academic argument and in no way meant to belittle your achievements in 1986.

Sources for rating performance:

Jimmy Liew said...

Let me answer you point by point.

"Then at the end, you said you cannot compare how the players in 1986 would fare against the players in 2006."

This kind of scenario have been argued many times before, I already gave some scenarios. And the reason is nothing to do with rating. Its the number of games played since 1986 , advances in computers that give current players the kind of access and knowledge not possible in 1986. That is why I said "peers" who had access to the same resources which more or less makes the playing field level .

"Team 2006 was rated at 2331 and performed at 2405. That means they performed 3.2% above their rating". If you measure performance that way, then I agree with you that their performance is better than 1986. It is a measure of how well that particular team did against the expectations of the ELO rating system.

If we were to use this measurement the best performing team of all time will be 1994 Moscow. Ironically, it was also the year we had the worse result in our history of participation. Perhaps you see why it makes no sense to use this method of performance. The olympiad rankings are not decided by looking at the team rating performance.

Is it not measuring ourselves against the rest of the world that is the objective of the Olympiad?

Post a Comment