Cognitive Daily

The NCAA basketball tournament is down to the wire now, with only four teams left. How will the players respond? Will they be able to perform under the incredible pressure from the other teams, the coaches, and most of all, the fans filling in their tournament pool brackets?

Today’s headline might be just a tad misleading. While psychologists may not be able to predict the results of the games, they are beginning to develop a better model of what’s going on inside highly trained athletes’ minds while they play. André Didierjean and Evelyne Marmèche of the Université de Provence just completed a study which looks at how professional basketball players remember diagrams of basketball plays (“Anticipatory Representation of Visual Basketball Scenes by Novice and Expert Players,” Visual Cognition, 2005).

There is a significant body of evidence suggesting that experts build mental pictures in a different way from novices. The classic expert/novice study was conducted in 1973 by William Chase and Herbert Simon: they found that expert chess players could remember board positions much better than novices. However, if the position on the board did not represent an actual game, the experts were no better than novices. So it seemed that, rather than simply having a better facility for remembering chess pieces, the experts were more readily able to “chunk” common patterns found in chess games. More recent research suggests that rather than chunking, experts are probably using “templates” in which most elements are constant, but a few key variables may be changed.

Basketball play diagrams involve patterns similar to chess pieces on a board, but real basketball plays unfold much faster than chess games. Didierjean and Marmèche suspected that experienced players would mentally represent basketball plays as moving, not just static diagrams. They designed a study to compare how novices and professional basketball players remember plays. Since Michael Jordan and Shaquille O’Neal don’t typically vacation in Aix-en-Provence, they studied professional players from the National League, the top women’s league in France (for some reason, France doesn’t have an American League).

Didierjean and Marmèche designed two experiments to test how players’ memory differed from nonplayers. They had a professional coach design 10 basketball configurations, all of which were reasonably expected to occur in real games. A typical diagram looked like this:


They then asked the coach to diagram a configuration that might naturally result from such play, such as this:


The task for participants was to study one diagram for 5 seconds, then compare it to another diagram which appeared one second later. They were instructed to indicate whether the two diagrams were the same or different. Half the time, the diagrams were the same. The other half of the time, the second diagram was either one which naturally followed the initial diagram, or one which naturally preceded it. Experts—professional players—were much better at determining that plays which did not follow sequentially were different, but they were no more accurate than non-players at determining that the diagrams that naturally followed in basketball play were actually different. Furthermore, they took nearly a second longer (an eternity, in basketball time) than non-players to perform the task.

Their second experiment was even more telling. This time, they asked a new set of participants to memorize 10 basketball diagrams that were displayed for 8 seconds each. Next, participants were shown diagrams and asked if they had seen them in the memorization phase. The results were as follows:


“Novices” were players who had played in clubs, but not professionally, and formed an interesting intermediate group. But note, most importantly, when the professional players were shown an image that naturally followed a play they had memorized, 60 percent of the time they believed they had seen that image before, even though they had not. Non-players effectively saw no difference between the diagrams that naturally followed or naturally preceded the diagrams they were shown.

It appears that professional players have learned to anticipate what will follow diagrams they are shown, and build that anticipation into the memory of those diagrams. So, later, when asked if they have seen a diagram they had anticipated, they will behave as if they had seen it before—even though in fact they had never seen it!

So, next Saturday night when you’re watching the Final Four and it seems like a player anticipates a play before it even happens, that may be exactly what he is doing. Now, if only those psychologists could help me anticipate how to fill in my Final Four brackets….


  1. #1 Brian
    March 29, 2005

    new scientist has an article today about ten unsolved mysteries in science in general. i thought it’d be neat to read something similar for the cognitive science psych field… just an idea :)

  2. #2 Dave Munger
    March 29, 2005

    That’s an interesting idea, Brian, but not one we at Cognitive Daily feel we are qualified to address. Our focus here is on the really cool research that has been done. You might suggest the idea to Chris over at Mixing Memory, but I don’t know if he’d touch it either.

  3. #3 Isabela
    March 29, 2005

    Closely related, and possibly of interest, Prof. David Eagleman at UTexas Medical School in Houston (teaching a class jointly offered at Rice University this semester) is publishing a book called ¡°Ten Unsolved Questions of Neuroscience¡±. I don’t know when it’s coming out- makes me wish I’d taken the class so I could have gotten a pre-printing of it.

    ¡°Ten Unsolved Questions of Neuroscience¡± by David Eagleman & Patricia Churchland (MIT Press 2005)

  4. #4 Brian
    March 29, 2005

    thanks for the tip Isabela i’ll definitely keep an eye on that space