Is Blindfold Chess a Bad Game? Comments on Chabris & Hearst
1. Introduction
Researchers have focused on memory and decision making for many years. For example, chess has been the main focus of their research (de Groot (1946), 1978; de Groot & Gobet (1996). The main issue was the mystery behind Grandmasters' superiority. The one side claimed recognition was more important than evaluation of the pattern, but the other side insisted that it is crucial for Grandmasters to search for and evaluate possible moves.
Simon (1996) and Gobet are two of the most prominent proponents of pattern recognition theory. Simon conducted a study that supported their views. Their analysis also included Garry Kasparov's performance in simultaneous exhibitions. Their argument was that, although Kasparov had played very fast, the rating quality (performance), showed no significant difference in comparison to his usual performances. Holding (1985) was a proponent for the other view. He argued that pattern recognition is less important than searching for and evaluating moves. Chabris (2003) and Hearst (2006) presented results that demonstrated significant differences in the frequency of mistakes in classical and rapid play. Their conclusion was that reducing the time limit has a significant impact on the quality and enjoyment of the game.
The search dilemma vs. pattern recognition was the backdrop to which a particular research area emerged. Because of its extraordinary nature, blindfolded chess attracted many scientists. Blindfold chess allows a player to play without seeing the board or pieces. This seems to require exceptional memory skills for a common observer. Famous chess players, such as Pillsbury or Bourdonnais or Koltanowski, could simultaneously play between 30-45 blindfold games. This phenomenon quickly attracted attention, leading to many studies. Binet (1893,1894) discovered that skilled players couldn't imagine the physical properties of the game like the colors or the pieces. Such players preferred an abstract representation. Saariluoma (1991), investigated the memory use in blindfold chess. He discovered that chess players were unable to recall illegal random moves. This same author also studied blindfold chess decision making. Kalakoski (1997 and 1998) demonstrated that the use of dots to replace chess pieces had no significant impact on memory performance, in masters as well as amateurs. They concluded that practicing blindfolds improves chess skills. Grandmaster Jonathan Tisdall (1997) also believed the same. It was his ability to play blindfolded, that was cited as the reason he earned the International Grandmaster title. Campitelli and Gobet (2006) and Campitelli. Gobet. Williams. Parker (2007) provide more details about the blindfolded-chess phenomenon.
Chabris and Hearst(2003) did not find statistically significant differences between errors made in blindfold games and those made in rapid games. This was the reason we were attracted to this surprising result. This paper's first author is an International Grandmaster, which gave us the unique opportunity of using his expertise in this area.
2. Method
Similar methods were used to analyze the results of six Monaco events, 1993-1998. This time period was not the only one that was considered. The analysis also included the Monaco tournaments of 2002-2007. Chabris & Hearst also studied the impact of time pressure. Their fundamental conclusion was that reducing a time limit drastically affects the quality and enjoyment of the game. This paper will concentrate on the controversy surrounding the rapid-game blindfold. This is an approach that has a few advantages. Chabris and Hearst have reached an important conclusion regarding the classical versus rapid game issue. We strongly support this conclusion. It's also very difficult for us choose the classical games that we want to include in our analysis. It's difficult to decide which of the many games Grandmasters around the world play against one another. Because we believe this is a very important issue, we decided to only compare the quality of rapid and blindfold games.
In the years 1993-1998 and 2002–2007, 12 International Grandmasters were competing for the title at the Monaco tournament. The organizers selected players based upon their Elo (Eleanor 1986) chess rating. Round-robin format is used for the tournament. Grandmasters compete in two games. One game is called the "blindfold", in which the players cannot see the actual position. Their field of vision is reduced down to the chessboard shown on the computer monitor. The other is the "rapid", which allows the players to see the actual situation. The time limits are almost identical for both games. However, the blindfold game requires that the player type their move in the computer for an extra 10 seconds. In the second game, players with white pieces will receive black pieces. The tournament begins with the draw and the final decision on piece color. This system allows each player to play as many games as possible with white pieces. All these measures were taken to maximize the objective tournament results.
To identify the mistakes, we used the chess program Rybka 3 with AMD Athlon Dual Core Processor with 2.00GB RAM, using 10 ply deep search (Vasikrajlich, Larry Kaufman). Rybka is both the World Computer Chess Champion of 2007 and 2008, and the best chess program in the world. Additionally, it is the highest-rated program (Karlsson (2009)
This chess program has analysed every move in all our games and reported any "blunders." Blunders are moves in which the actual move was at least 1.5 times worse than the program's. In this analysis, it is irrelevant whether the opponent used the blunder. This research excluded all errors that could have an effect on the game's outcome. If the same side has a minimum 3.0 pawn advantage, even after a blunder is made, the error cannot be calculated. This criteria of 1.5 pawns was selected by computer chess researchers and Grandmasters who consider this sufficient to win the game (Hartmann 1990).
3. Analyse on the 2002-2007 Monaco games
Table 1 summarizes the results from analyzing our data with Rybka Chess. The table below shows that the rapid game lasted 46.87 moves. However, the blindfold game took five moves less, at an average of 41.94 steps. This is where our concern lies: the frequency of blunders in both of these situations. In the rapid games, there were 0.53 blunders per match to 0.67 in blindfold games. Also, we calculated the variable number blunders, which is (numbers of blunders/number of moves * 2)1. In rapid games, there were 5.84 blunders every 1,000 moves. Blindfold games were the most common, with 8.42 per 1,000 moves.
The results show that Grandmasters made greater mistakes in the blindfold game. One interesting fact is the number of moves in blindfolded games.
Next, we used the Kolmogorov–Smirnov Test to see if each condition's number of moves and number of blunders had a normal distribution. Gupta (1999) used the Kolmogorov–Smirnov Test to determine that neither of these variables were distributed normally (p’s.05). Therefore, we performed nonparametric testing.
Our next step was to divide the data into 396 pairs (a rapid game and a blindfold make a pair). A pair could be composed of the rapid and blindfold games played by Vallejo and Aronian in 2007.
The Wilcoxon Signed Ranks Test was used to determine the number and blunders per games, as well as the number and blunders for each condition (blindfold chess or rapid chess). Since it compares distributions that cannot be assumed to be normal, the Wilcoxon Signed Rank test was very common (Gupta et al. 1999). This is exactly why we used a nonparametric method to analyse the two samples. This method shows that the differences in moves between rapid games and blindfold games are significant, with p.001.
This finding is very easy to interpret. It is difficult to play a blindfolded game. Players have trouble staying focused. Blindfold games are less complex because of this. This paper focuses on the frequency of blunders. These results reveal a stark difference (p =.01), in the number of mistakes per game in different conditions (rapid, blindfold). This is a significant finding as it strongly challenges the Chabris/Hearst conclusion.
The data were further analysed by comparing both the number of mistakes per move. These results also revealed significant differences in the number of mistakes per move for the rapid and blindfold games with p.001
The paper discusses the importance of blindfold games and the possibility of potential blunders. Mantel-Haenszel Common Odds Ratio Estimate = 1.545 (95%CI 1.16-2.05, p =.003). This part is important because it evaluates relative risk for blindfolded games. This perspective has been a valuable tool in our research. Blindfold games are at a relative risk of 1.2391 (CI 1.0787-1.222). If the game is blindfolded, one mistake in a game under rapid conditions will result in 1.2391 errors.
4. Analysis of Monaco 1993-1998: Games
In this section, we used the same research method as the first. Table 2 shows the results from Rybka chess program. Table 2 shows that the rapid game was completed within 49 moves. The blindfold game took three moves longer, and the average was 45.84 steps. This is what we are concerned about in our research. The average number per game for rapid games was 0.644. Blindfolded games were 0.75. Playing Blindfolded
Komentar
Posting Komentar