This is a very important advantage, but reigning champion Viswanathan Anand is no stranger to winning from behind: he was able to turn around matches against Bulgarian Veselin Topalov in 2010 and against Israeli Boris Gelfand in 2012.
If Anand manages to win two of the remaining six games, the championship will come down to a number of tie-break games with increasingly faster time limits.
So what are the differences between traditional chess and fast chess? Are the same skills required to be good at both modalities?
Chunks and templates
Here is a snapshot of the template theory: when chess players study and play games, their brains produce memories of typical configurations of pieces.
These memories are available for future use, and, depending on their size, they are called chunks (configurations of three to four pieces) or templates (more than 12 pieces). Templates are also associated with typical moves, typical strategies and other information related to that configuration of pieces.
Pattern recognition occurs when experienced players observe a position in a game. The templates most similar to the current position get automatically activated together with the most typical moves.
The player engages in search only if the typical moves don’t seem to work in the current position. In this context, search consists of trying to anticipate the opponent’s responses to the player’s move.
A typical anticipation would be “if I move my bishop to g5, he may play the pawn to h6, threatening my bishop, then I would move my bishop to h4 …” and so on.
The theory indicates that the position gets updated in an internal structure called the “mind’s eye”, and pattern recognition is applied recursively over the positions represented in the mind’s eye. This, obviously, takes time.
In fast chess, there is little time to execute search processes; therefore, players mostly play their games in pattern recognition mode.
As indicated earlier, the template theory enunciates that pattern recognition also participates in search processes. Therefore, for the template theory there is little difference between normal chess and rapid chess.
On the other hand, theories that consider search as a very different process to pattern recognition (such as Holding’s SEEK theory suggest that normal chess and rapid chess require very different skills.
In a 2007 study of Argentine chess players, Professor Gobet and I found that the correlation between the rating in normal chess and the rating in blitz chess (consisting of only five minutes per player for the whole game) was .89. Given that 0 indicates no relation between variables and 1 indicates perfect relation between variables, .89 is a very high correlation.
These data provide some support for the template theory. This result was confirmed in another study this one by Bruce Burns from the University of Sydney, who found correlations between .78 and .90 in an Australian sample and a Dutch sample.
In essence, a person good at traditional chess will very likely also be good at fast chess.
Halfway through the World Championship
Since my last article on the World Chess Championship match, Anand and Carlsen have played four more games.
The first of these two were very long, and, unexpectedly, the players with black pieces obtained clear advantages. However, the defensive resources of both players were superb, and both games finished in a draw.
In game 5, Carlsen again playing with white pieces was not able to obtain advantage in the opening. However, a slight inaccuracy of Anand’s led to an endgame with a small advantage for Carlsen.
Carlsen’s endgame ability is his main strength; he tests his opponents with ingeniously disguised tricks. Anand could not defend perfectly and lost the game.
Game 6 was similar to game 5. They arrived to an endgame with a small advantage for Carlsen, in which Anand could not resist Carlsen’s constant pressure, and eventually lost the game.