Errare humanum est
To err is human. This certainly applies to chess players. Errors, including horrible howlers, are fairly common in chess. Most players, including world champions, blunder every so often, and the technical literature on chess provides some spectacular examples. In the game Petrosian against Bronstein in 1956, played in the Amsterdam qualification tournament for the world championship, future world champion Petrosian was winning but did not see that his queen was attacked. When Bronstein took her, Petrosian resigned on the spot. In the game Alekhine against Euwe, played in the world championship return match in 1937, both players overlooked a simple combination for three moves in a row. More recently, during the 2017 Wijk aan Zee tournament, world champion Magnus Carlsen missed a checkmate in three moves in his game against Anish Giri and let his lucky opponent escape with a draw. Worst of all, in his match against computer programme Deep Fritz, world champion Vladimir Kramnik overlooked a mate-in-one and was painfully checkmated.
Relatively little scientific research has been carried out on errors in chess. De Groot’s thesis has some mentions in passing and a few studies have used blunders, typically identified by chess computers, as a dependent variable. The only detailed qualitative analysis trying to understand the mechanisms leading to errors was carried out by Pertti Saariluoma. He induced errors by choosing combinations that could be found only by identifying two themes in close succession. As expected, most of the errors were made with the move that linked the two themes. Similar results were obtained with strategic positions and endgames. Saariluoma concluded by arguing that working memory overload explains errors only in part, and that it is necessary to develop more sophisticated explanations based on pattern recognition, planning and problem restructuring. This study explored an unchartered topic and called for more experiments; for example, one could try to induce specific errors as a way of testing different theories of chess skill.
By contrast, there are numerous discussions of errors in chess, and how to avoid them, in the practical chess literature. Alexander Kotov’s Think like a Grandmaster and Nikolai Krogius’s Psychology in Chess, which both present instructive and thought-provoking examples from grandmaster play, are good examples of such books. While their discussions are based on observations, anecdotes and a fair amount of speculation – explanations are often based on introspection, hindsight and a posteriori accounts – the described errors are real.
Kotov and Krogius provide several explanations for errors in chess. Some of the explanations are specific to chess, whilst others apply to other endeavours as well. Let us start with the domain specific ones.
Automatisms are important in chess, since they allow one to make decisions rapidly. Some consist of very simple conditioned actions, such as “if your opponent takes one of your pieces, take it back immediately”, or “if a piece is attacked, move it away”. Others are subtler and directly link to the chunking mechanisms discussed in chapters 2 and 3. For example, given a certain pawn structure, it is often a good idea to place one’s knight in front of an isolated pawn. The beauty of chess is that it is a game of exceptions, and even the best procedural knowledge will be incorrect every so often. If one follows one’s intuition without doubling-checking variations concretely, there is always a risk that some tactical blow will be overlooked.
Krogius spends a great deal of time with what he calls “chess images”. The term is maybe not the best one, as these images contain not only visual information on the location of pieces, but also more conceptual information such as the evaluation of the position. In some of Krogius’s examples, images really refer to what I have called schemas or templates in previous chapters. Chess images are very useful in most cases, as they provide much information. They can also produce powerful negative effects, however. For example, facing an unexpected move by the opponent, a player will often still think using the mind-set provided by a chess image that was suitable just one move ago, and might miss a new opportunity. Several examples will be provided below.
Another class of errors, which Krogius calls “retained images”, concerns the case where, when anticipating a sequence of moves, a piece taken by the opponent somehow still remains present in the mind’s eye. The difficulty thus resides in correctly updating the location of pieces or their disappearance in the mind’s eye. For example, with this type of error, a player would defend in his calculation against a bishop that actually is not on the board anymore. In a related type of errors, which Krogius calls “inert images”, the player automatically carries a positive evaluation reached at some point of the game on to the following moves and is essentially on an automatic pilot mode. As a consequence, the difficulty of winning the game can be underestimated and one’s sense of danger blunted. So, for example in the game Petrosian against Korchnoi in the 1963 Moscow tournament, Petrosian had had a winning position since the opening. Still winning after 34 moves, he overlooked a simple combination and played a careless move, which lost immediately. This type of error is related to overconfidence, which I shall discuss towards the end of this chapter. It is also related to the Einstellung effect that I described in Chapter 3. In this respect, Krogius reports a very interesting comment by Petrosian: “Personally, I am convinced that if a strong master does not see such a threat at once he will not notice it, even if he analyses the position for twenty to thirty minutes”.
Many errors are due to attention lapses. A typical example is the case where one player misses an obvious threat, such as in the example above where Petrosian lost his queen in one move. To some extent, it could be argued that at least some of the errors that Krogius imputed to chess images are in fact attentional errors. In some cases, it seems that the effort required for anticipating moves wipes out from working memory important information, such as an attack against one’s queen.
It might come as a surprise for non-players, but chess can be a very emotional game. To begin with, most players have their bêtes noires, opponents against whom their score is much less than expected by their relative objective strength. For example, Soviet grandmaster Efim Geller was the bête noire of Bobby Fischer, with five wins, three defeats and two draws, in spite of Fischer’s superiority. Similarly, American grandmaster Hikaru Nakamura, currently number 8 in the world, struggles against Magnus Carlsen, having lost 12 games and won only one, with 21 draws. Krogius reports that he analysed 80 games between 10 pairs of players where one player was a bête noire for the other. He found that the players on the losing side committed more obvious strategic errors and tactical mistakes against their bête noire than against other players. Krogius argues that a negative emotional state strongly diminishes these players’ vigilance and nervous resistance, although it could be argued that, in addition to this, there might be a profound incompatibility in style. For example, in the case of Fischer and Geller, Geller was known for his very dangerous, uncompromised attacking style. Interestingly, bêtes noires are not transitive.For example, ignoring draws, Tal had a lifetime score of 8–5 against Bronstein, and Bronstein had a lifetime score of 7–4 against Keres. However, rather than Tal dominating Keres, as predicted by a transitive relation, it is Keres who led by 8–4!
The opponent’s behaviour is another important emotional factor that might cause distraction and irritation. The noise he is making when drinking his coffee, the way he is looking at you, his aftershave . . . pretty much anything might exasperate you. In the 1959 candidates’ tournament, Hungarian grandmaster Pal Benko accused future world champion Mikhail Tal of hypnotising him, and decided to wear dark eyeglasses when playing against him. (This is another example of a bête noire, as Benko had lost the last five previous games against Tal.) The game ended in a draw, which is all what Tal needed to qualify for the world championship match against Botvinnik. After the game, Tal quipped: “When I want to win against Benko I win; when I want to draw, I draw!”
Emotional hatred and accusations of hypnosis were taken to new heights in the 1978 world championship match in Baguio between Viktor Korchnoi and Anatoly Karpov. Karpov included in his team Vladimir Zukhar, who at the time was a renowned parapsychologist and hypnotist in Soviet Union. Korchnoi complained and started wearing reflecting glasses. (This is only one of the many bizarre incidents that single out this world championship.) Whether or not hypnosis really played a role, I doubt it. But it is true that Korchnoi – although admittedly playing some brilliant chess at times – made terrible mistakes in some games. For example, in the 17th game, he committed a string of errors which turned a winning position into a drawn position, in which he blundered and overlooked a simple checkmate in three moves.
A final source of emotional pressure is one’s position in the tournament and the possibility of achieving an important result. In 1988, I was playing the tournament of my life in Biel and was very close to obtaining a grandmaster norm, a key requirement for becoming a grandmaster. I would have been the first Swiss player to do so. My opponent was former women’s world champion Nona Gaprindashvili. After some energetic and precise play in the opening and early middle game, I obtained a winning position against her unconventional defensive system. However, at the crucial moment, I missed a simple move that meant winning straight away, progressively lost the thread and ultimately had to resign. There is no doubt in my mind that my defeat was due to the fact that I could not handle the pressure of being so close to a grandmaster norm. How to regulate emotions is of course an important topic in other sports as well, and is indeed a standard topic in sports psychology.
Many errors are caused by a lack of knowledge. At weaker levels, this can be due to not knowing typical openings, standard tactical and strategic patterns in the middle game and common manoeuvres in endgames. As noted in an earlier chapter, masters and grandmasters often win games by waiting for mistakes, and lack of knowledge is at the core of many of them. Mistakes due to insufficient knowledge also occasionally occur at the top level. For example, in the first game of the world championship held in 2010 in Sofia between Veselin Topalov and Viswanathan Anand, world champion Anand could not remember the move order of a variation he had prepared in depth. He chose the wrong order and got thrashed by his opponent, who had remembered the brilliant combination that refuted Anand’s move.
Competitive games are played with a clock, and thinking time is limited. The limit is strict and overstepping one’s thinking time means defeat by “losing on time”, which is not uncommon. An extreme case is grandmaster Friedrich Sämisch, one of the leading players in the 1920s, who lost every single game on time in two tournaments in 1969, when he was 73 years old. The exact time limits vary from tournament to tournament, but there has been a tendency towards a decrease of thinking time over the years, partly to reduce the duration of games and make them more exciting. Currently, many tournaments use time controls of 90 minutes for the first 40 moves, followed by 30 minutes to finish the game, with each player receiving an increment of 30 seconds after each move.
Before the 1990s, it was common for each player to have 2 hours 30 minutes for 40 moves, and then 1 hour for each sequence of 20 moves. Games used to be interrupted (“adjourned”) after 5 hours of play, and resumed later in the day or the next day. This break was meant to allow players to rest and eat. In practice, it also allowed players and their seconds to analyse the current position. There used to be an interesting literature describing the excitements and pitfalls of analysing adjourned games, and about how to do it properly. However, with the advent of powerful computers, which pretty much find the optimal sequence of moves for both sides, it became clear that adjournments were becoming meaningless as the human factor simply disappeared. The practice of adjourning games progressively disappeared in the mid-1990s, with the added advantage that games are now played in one go, which is much more enthralling from the spectators’ point of view.
Whatever the exact time controls, it is common for some players to think for long periods in the opening and at the beginning of the middle game, when they face critical decisions. As a consequence, they have little thinking left at the end, and are in time trouble, or in zeitnot, a German word commonly used by chess players. The exact definition varies from player to player. For some, having less than 1 minute per move is considered as time trouble; for others, having to play 10 moves in 5 minutes is nothing special. Players vary considerably as to whether they are likely to be in time trouble. Some players are in zeitnot in nearly every game, and are used to playing 20 moves in 2 or 3 minutes. Others systematically avoid any time shortage.
During time trouble, short-term tactics dominate over long-term strategies, intuition comes before look-ahead search and concrete ideas take precedence over abstract considerations. Simple heuristics are used, such as “simplify positions by trading pieces off”, “restore material balance rather than play for the initiative” and “if possible, postpone making important decisions until the end of the zeitnot”. Because decisions have to be made rapidly and calculations are superficial, blunders occur often. In general, the advice is to not be in time trouble. As world champion Alekhine wrote in criticising one of his moves: “A horrible move, and in my opinion the fact that White was in time trouble when he made it is no more justification than the claim of a law breaker that he was drunk when he committed the crime”.
Alekhine’s comment might be too one-sided. There is actually an interesting trade-off with respect to decision-making. Not spending enough time on critical moments in the game might lead to playing inferior moves, losing the advantage in a superior position or not defending a difficult position properly. Thus, being in zeitnot might be a worthwhile price to pay. On the other hand, many players simply cannot make up their minds, sometimes because they want to find the optimal move rather than to play a move that is good enough. An interesting question, which to my knowledge has not been studied, is whether there is a link between being addicted to time trouble in chess and procrastination in life outside chess.
Many tragedies have happened in zeitnot, with winning positions being destroyed in one move. My game against grandmaster Lubomir Ftacnik in the 1984 Biel tournament is a good example of this. After a complicated game where my opponent progressively outplayed me, we were in a mutual time trouble. My opponent made a terrible mistake five moves before the time limit, overlooking a simple checkmate in two moves. His blunder is also a good example of an error caused by a conditioned reflex. He thought that attacking my rook with a pawn would force my rook to move away, while in fact moving my queen created an unstoppable checkmate threat.
This example illustrates the case where being in zeitnot is a rational decision. My position had vastly deteriorated in the last 10 moves, and was clearly lost. Thus, the only practical chance was to create a situation – a reciprocal time trouble – where the logic of chess could be perturbed by random factors and by putting psychological pressure on the opponent. While the choice to be in time trouble was rational from my point of view, it was a mistake from the point of view of my opponent. I hasten to add that such an analysis is easy with hindsight, but much harder to make during the excitement of the game. In fact, I found myself in Ftacnik’s unhappy situation in several of my own games, losing a winning game because of blunders caused by lack of time.
How to behave when one is in time trouble? Should one try to exploit or not the opponent’s time trouble? These are important practical questions, and much ink has been spilled on them. Krogius answers the first question by advising to keep one’s concentration and avoiding distracting thoughts. If one has a plan, one should follow it; else, the best option is to use waiting tactics, since there is simply not enough time to come up with a sound plan. He also recommends trying to find the goal of each opponent’s move, even by consciously asking questions such as “What is the threat?” Finally, he advocates using auto-suggestion techniques such as verbal commands, the efficacy of which has been documented in sports. As for Kotov, his main advice is simply to avoid getting into zeitnot. If this not possible, the main thing is to remain calm and to play as if one were not in time trouble. Basically, to do the same as normal, just quicker. According to him, this is what top players are doing.
There is some disagreement about how to handle an opponent’s zeitnot. On the one hand, both Krogius and Kotov agree on a few points. One should control one’s nerves, stay calm and evaluate the position objectively. In particular, one should not get excited or irritated by the opponent’s behaviour. Note that this is not always easy, as players with few seconds on their clock are sometimes in a state of extreme agitation if not outright panic. Both also advise against complicating the position for the sake of it, and against speeding up one’s play to prevent the opponent from thinking during one’s own time, as this is likely to lead to errors. Where they disagree is whether anything at all should be done to capitalise on the opponent’s shortage of time. Krogius recommends calculating a sequence of four or five moves, which if possible change the nature of the position (e.g. moving from a middle game to an endgame), and to play them quickly, in the hope to take one’s opponent unaware. Kotov disagrees, and points out that this is likely to lead to mistakes on the part of the player who has sufficient time. Rather, he endorses world champion Vasily Smyslov’s advice of leaving the board and coming back only after the opponent has played a move. Thus, Kotov’s general recommendation is to fully ignore one’s opponent’s time trouble.
Tiredness, overconfidence and habituation
This section briefly discusses a few additional sources of errors. An obvious one is tiredness; after a long and difficult game, or after several days or weeks of playing in a tournament, one is bound to become tired. International master Charles Partos, who coached Biel in the Swiss team championship, used to bring bananas and chocolate after 4 hours of play to replenish players’ glucose stores and thus provide energy for the brain – a sound application of nutritional principles. The long-term advice to avoid tiredness is to be fit physically, and many world champions such as Boris Spassky, Bobby Fischer, Garry Kasparov and Magnus Carlsen practiced several sports and were in top physical condition when they won the title.
Another standard source of error, not only in chess but in other sports as well, is overconfidence. With overconfidence, vigilance goes down, attention becomes relaxed and possible dangers are ignored. The chess literature contains many examples, even at the top level, where the proximity of victory led to overconfidence, a decrease in the quality of play and often errors that turn a win into a draw or, worse, a loss. Related to overconfidence is the desire to win a game by a flashy combination rather than by prosaic methods. As objectivity gets lost, the “brilliant” combination often turns out to be flawed and the opponent escapes with a lucky draw or even a win.
I have once been victim of an interesting kind of mistake, which I have not found reported in the chess literature, and which might be related to the mechanism of habituation. In animal psychology, habituation is a kind of learning where a response to a stimulus decreases or even stops after repeated exposure. In my game against former world champion Boris Spassky at the Reggio Emilia tournament in 1983, I chose a very sharp opening and had the advantage after the opening. For several moves, I considered a pawn sacrifice in the centre in order to initiate a direct attack against Spassky’s king. However, the idea never worked satisfactorily, so I progressively directed my attention to another central break. Ironically, at the very moment where I gave up on my initial idea, it would have in fact been the winning move.
How to avoid errors?
Chess coaches have provided many recommendations for avoiding errors. Some of them are rather obvious and simply aim to eliminate the source of the errors I have discussed. For example, one should be aware of automatisms, control one’s emotions, make sure that one’s knowledge is up-to-date, relax and get enough sleep before an important game and avoid time trouble. Some more general recommendations have also been made, which tend to take action against mind-set effects. A piece of good advice is to look at one’s position from the point of view of the opponent. This is meant literally: one should every so often physically stand behind one’s opponent and have a look at the position. This is actually not bad advice – maybe in a more figurative way – for many other avenues of life where one is likely to hold preconceptions.
Kotov encourages players to use Blumenfeld’s rule, named after Soviet master Benjamin Blumenfeld, who was also a psychologist. The idea is to make sure that one does not miss the obvious, even after having spent dozens of minutes analysing a position. After having selected a move, one should write it down on one’s score sheet, before actually playing it. (In tournaments, players are required to write down their moves and their opponents’, as an official record of the game. This can be useful, for example, to ascertain whether the required number of moves have been played before the time control.) Kotov emphasises that care should be taken to write the move neatly and clearly. He argues that, by doing so, one goes from the future possibilities of the game back to the here and now. A variant of this advice is to stop looking at the board and stare at some distant object in the room. Then, one should look again at the board, with fresh eyes, so to speak. This should make it possible to look at the position as it is on the board right now. The advice is now to stare at the position for about a minute, through the eyes of a novice, and ask very basic questions such as: is my queen attacked? Is there a direct threat against my king? According to Kotov, this double-checking procedure drastically decreases the risk of committing a blunder.
This is very good advice. Having used it in my chess career, I can vouchsafe that it saved me from an embarrassing oversight more than once. Unfortunately, one important part of Blumenfeld’s rule is not possible anymore. The International Chess Federation has changed its rules, and moves must be written down on the score sheet only after they have been executed on the board.
- Fernand Gobet is Professor of Decision Making and Expertise at the University of Liverpool. He is a chess international master, and played numerous times for the Swiss national team.
BPS Members can discuss this article
Already a member? Or Create an account
Not a member? Find out about becoming a member or subscriber