Sunday 21 December 2008

The Langer Principle

While listening to Australia go down to yet another historic defeat in the cricket (for those that haven't heard, South Africa easily chased down the 414 runs Australia set as a 4th innings target), Justin Langer (former Australian Open) made an interesting comment about players stepping up to Test Cricket.
His theory was that a player should really dominate one level before moving onto the next. In his example an average of 50 runs per innings was a benchmark for batsmen, so if you want to move from First Class to Test Cricket, this is what you should already be doing at the First Class level. Similarly a move from 2nd grade to first grade requires the same kind of performance.
Now for a lot of us, the level we play chess at is either forced upon us (eg by the club we play at) or self selecting (by which level of tournament we choose to enter). For example, the Doeberl Cup has been run in sections for over 20 years, and most players are happy to stick to their own sections (ie 1800's play in the Under 2000 section etc). But there has always been a small group who wanted to play at the highest possible level. When I asked why the reply was often "To improve my rating". Basically they accepted that they would lose 5 of their 7 games, but a positive result in 2 games may result in a ratings increase.
I've always thought that this was the wrong way about improving your chess (or even your rating), unless you had already outclassed players at your current level. In much the same way as Justin Langer suggests, it is the ability to beat players rated below you, or at the same level as you, rather than the odd win against higher rated opponents, that is a better sign post for improvement. And more importantly, the skill set required to win an event with 6 or 6.5/7 is both different, and more important, than the skill set required to win 2 games (and lose 5).

7 comments:

Anonymous said...

Shaun,

Unfortunately or fortunately, that is what the recent research tells us, see for example http://www.sciam.ru/2006/11/psyhiotrya.shtml.

It is also what famous coaches recommend. They say that the best tournament for your development is such where you are expected to score about 35-40%. If one wants to improve, every game has to be a really tough encounter.

Vladimir

Anonymous said...

Shaun

If there was a big open in Sydney, with Topalov, Anand, Carlsen, but there was an U/2200 section, would you play in the U/2200 section ?

Garrett.

Shaun Press said...

Me, I'd probably be an arbiter. But if I had the chance to play in the top section I'd probably choose the top section. If I was coaching a 1300 rated player who wished to play in the event, I would strongly suggest they play at their level.

TrueFiendish said...

"Ratingcentrism" is putting the cart before the horse. Surely the only reason to play in a tournament where you are likely to be regularly beaten is to improve your ability to play chess. If you do improve, your rating will increase.

Anonymous said...

I really don't care about the research but care about enjoying my chess and perhaps getting some friendly tips after the game.

I'm rated in the low 1500's and had a recent win against a 2000. However if I play in an u1600 event I could play someone rated 850 and have a lot of spare time on my hands before the next round.

Anonymous said...

I'm going to disagree with you here, Shaun, both as to the importance of the "domination" phase and the enjoyment factor. As somebody who took up chess late and has always had a very large variance in the consistency of my results, I get a lot more enjoyment out of playing in slightly outclassed situations than the reverse. I would also say I learn a lot more from the more difficult events. I've always admired the ability of people like Ian Rogers to consistently dominate local events and I agree there's a definite skill involved in that, but, personally and, I gather from talking to some friends in the chess playing community, it's motivating to have to rise to the occasion and really knuckle down from move one against stronger opposition.

Kevin Bonham said...

I think the difference between cricket and chess in this regard is that cricket is exclusively a team sport. A player who has not really dominated first-class cricket might very well learn more and improve faster from playing against world-class opposition, but even if they do learn more the problem for their team is that their poor performances while they're climbing the learning curve will harm the team's results.

This downside clearly doesn't apply in chess (except if you pick people for Olympiads not on merit but to blood them for their future promise, which the ACF specifically avoids but some other countries do). Still, I agree that the test of a rating as a sign of playing strength should be your ability to hold that rating at your own level. Not only is there some evidence that upsets occur more frequently than the rating curve says it should, but some players are particularly good at pulling off upsets (the downside being that they are unreliable against the rats and mice). If weaker players deliberately select many events where they might get upsets for ratings purposes then they may be inflating their own rating slightly, and doing so in a meaningless fashion.