07/26/2006
Playing Chess
It was the ultimate lesson in course management. I’m starting this column after Tiger’s emotional win at the British Open. To have hit so many fairways right down the middle on a course filled with so many unpredictable bounces was a testament to Woods’ mastery of his craft. Having myself golfed on two British Open courses, St. Andrews (carded a 114) and Carnoustie (109), I can appreciate even more Tiger’s achievement.
I’m also a hacker at chess. Our 13-year-old grandson and I have played a few games recently. It took him half an hour to convince me that I could not checkmate him even though I had a knight and my king versus his lone king. (He later found in a chess manual the statement that even if I had two knights I wouldn’t necessarily have been able to checkmate him.)
Do grandmasters in chess and golfers of Tiger’s caliber share something in common? Are experts born or made? Philip Ross, in “The Expert Mind” in the August Scientific American, writes about studies on grandmasters and how they become experts. The article pictures Tiger Woods as a child prodigy before becoming an expert in his field. In last Sunday’s (June 23) New York Times Magazine section, Professor David Karp of UC Berkeley wrote about a related subject, the effects of nurture and nature on IQ.
Ross cites an exhibition in 1909 by Jose Raul Capablanca of Cuba. He was playing 28 amateur players at once, going from one to another around a circle, taking only two or three seconds to make each move. These 28 matches were part of 168 successive matches he won on a tour showcasing his expertise. When asked how he could see so many moves ahead, he said he only saw one move ahead, but that was always the correct one!
So, the question: is expertise in chess at the highest levels, due to some innate characteristic, perhaps a special ability to foresee the ultimate consequences of a given move? Or is the expertise due more to intensive training and experience? The evidence favors the latter. How have researchers reached that conclusion? The nature of chess and its rating system has provided the raw material for studies over the past 70 years or so. The system tracks the performance of chess players and a statistical scale has evolved that can predict with a substantial degree of accuracy the results of matches between players of different rankings. A player whose opponent rates 200 points below him or her on the scale will beat the opponent 75% of the time.
For example, the famed Russian grandmaster Gary Kasparov has a rating of 2812 while Jan Timman, the 100th-ranked player, has a rating of 2616, essentially 200 points lower. Kasparov will win 75% of his matches against Timman. The median rating is about 1200 for tournament chess players and if a 1200-rated guy plays a 1000-rated gal the guy will win 75% of the matches. Enter the Dutch chess master and psychologist Adriaan deGroot, who took advantage of an international tournament back in 1938 to carry out an interesting experiment.
De Groot asked the tournament players to look at a board on which the pieces were arranged in positions matching those in an actual tournament game. They were allowed only seconds to look at the board and then were asked to reconstruct the positions. He found that the grandmasters were much better in recalling the positions than the novices. Not much of a surprise, I imagine. However, since then a number of similar studies have been made and the Scientific American article contains data that shows something more surprising.
One set of data is essentially the de Groot experiment in which the players examined for only 10 seconds a board with positions matching those in an actual tournament game. Players rated below 1600 only could recall about 4-5% of the positions correctly. Players rated above 2350 recalled over 20% correctly, 4 times more than those rated below 1600. Players rated in between the two extremes fell neatly in between.
However, in another experiment, the players looked for 10 seconds at a board on which the pieces were arranged randomly, not from a tournament game. This time the lowest rated players got about 2-3% of the positions correct while the players rated above 2350 only got about 5% correct, a major fall from the 20% they achieved for an actual tournament situation. Conclusion? These grandmasters aren’t gifted with unusually sharp memories; they have filed in their memories a wealth of tournament positions that enable them to draw on these memories rather than trying to memorize individual pieces in a short period of time.
In 2001, workers at the University of Konstanz in Germany ran brain-imaging studies on subjects playing chess on a computer. The results showed how much activity was going on in areas of the brain related to analysis and to long-term memory. The lowest rated players spent 40-65% of their brain activity in analysis. The highest rated players, rated 2500-2600, spent only about 20% in analysis and about 80% in searching long-term memory! Those thousands of hours studying and playing chess have created immense files of board positions over the years.
The implication is that experts are made not born. Would Tiger have become last Sunday’s Tiger without his late, beloved dad and the thousands of hours spent preparing for that moment? I think it’s unlikely. Tiger and his dad have counterparts in the chess world. A Hungarian, Laszlo Polgar, decided some decades ago to rear his three daughters to become chess experts and schooled them in the game, assigning as much as 6 hours a day of homework. Today, one is an international master and the other two are grandmasters. One, Judit, is 30 years old (Tiger’s age, I believe) and is ranked 14th in the world!
So, you can make an expert. Can you make a genius? I suspect the answer is no. However, Karp, in the Times article, discusses IQ studies and nature vs. nurture. I was taken with two studies Karp cited. Identical twins might be expected to have nearly the same IQs, if IQ is genetically determined. Psychology professor Eric Turkheimer reexamined studies done some 30 years ago and found that identical twins raised in wealthier families had nearly identical IQs. Not so for identical twins raised in the poorest families; the IQs varied as much as the IQs of fraternal twins, which are not clones of each other as are the identical twins.
Another study on adopted children in France showed a similar result. The children, 4-6 years old, were generally abused or neglected as infants and had been in various foster homes or institutions before their adoption. Their average IQ was only 77, close to retardation. The children were tested 9 years after being adopted. Those adopted by farmers and laborers increased their IQs to an average of 85.5. Those in middle-class families had IQs averaging 92. Those in wealthy homes jumped up to 98, nearly the average IQ for the general population.
Not all wealthy homes contain good parents. However, on balance, the wealthier environments provide the resources to fully develop the innate genetic capabilities. These and other such studies have spurred the movements to provide more preschool education to all children to help level the playing field.
Back to chess. Yesterday, I did sneak my queen deep into my grandson’s territory and a relatively quick checkmate resulted. However, he handily defeated me in a game he suggested where we just used only our pawns. To add insult to injury, he also took three out of four games of pool the day before. Old Bortrum is no expert!
Allen F. Bortrum
|