The origins of chess are shrouded in mystery. Board games were certainly played by the ancient Egyptians and Romans, but nobody knows the rules of these games or whether they were the ancestors of any games played today. The earliest date to which a definite precursor of chess can be traced is about AD 600. Travelling from India via the Arab world to Europe, chess has exerted its peculiar fascination over a wide range of cultures. This universal appeal, stretching for a span of 1400 years, surely indicates that chess taps some deep-rooted elements of the human mind. Chess almost certainly had its origins in a type of war game, and the sporting or competitive element is still one of its most seductive features. The urge to compete is undoubtedly a fundamental part of the human psyche, and while one may argue as to how much of this urge is a result of upbringing and how much is genetically based, the fact is that it exists and is likely to do so for the foreseeable future.
However, chess does not appeal only to the basic desire to win. It also exerts its pull on another important psychological element, the desire to detect patterns and to impose order on chaos. A human playing chess depends not only on memory and ability to calculate sequences of moves, but also on pattern recognition. This often manifests itself on a subconscious level. Somehow, deep within the brain, all the games of chess one has ever seen have made a mark, and the position on the board at any given moment is compared with all these pre-existing patterns. A close match will result in a sudden “feeling” that one knows what the correct plan should be. A skilled human player will know when to trust these mysterious hunches and when to show a healthy scepticism. Often, when a grandmaster is asked why he played a particular move, he will struggle to explain exactly how he came to a decision. It may appear as though he is trying to keep his secrets to himself, but more often it is simply because he does not himself understand the subconscious processes that led him to play a particular move.
This type of process is one of the highest mental faculties of which human beings are capable. By contrast, purely mechanical reasoning, based on calculation, is less distinctive. Suppose, for example, that a businessman makes decisions about whether to invest money in a particular endeavour purely on the basis of financial calculation. Then he might just as well be replaced by a computer, which would be faster and perform the calculations with less chance of error. A real talent for business goes far beyond routine calculations, into the area of judgement and intuition. Just as in chess, a talented businessman will have a lifetime’s experience in his subconscious and this will enable him to make a reasonably accurate decision, even when the information available would result in an “insufficient data” error from a computer. Indeed, his main problem will probably be to try to convince his colleagues of the correctness of his decision. When playing chess, one is alone. There are no colleagues or meetings. One has total authority to implement one’s decisions, and conversely one has to accept full responsibility for the consequences, good or bad. This, too, is one of the appeals of chess. There are few areas in life where decisions can be made without consultation and at a moment’s notice, and there are few areas where the effects are visible so quickly.
The history of chess is one of very gradual development. The rules have evolved over the centuries but have been more or less static for the past 500 years. For half a millennium, chess knowledge has gradually accumulated, but today’s grandmasters still find that much about the game is totally baffling. Now, however, after 1400 years, chess is facing a new challenge from the computer. It is far from clear whether the strongest chess-playing entity on the planet is based on carbon or on silicon. It often seems remarkable to non-chess players that a human has any chance of beating a computer at chess. This arises because of the common misconception that playing chess is all about calculating sequences of moves. The logic then is that since computers are much better at calculating than humans, a computer should normally beat a human. The fallacy here is that, as mentioned above, calculating moves is just one part of playing chess. The mysterious processes of the subconscious are just as important, and these cannot, as yet, be programmed into a machine.
In the 1970s, a human vs computer contest was just a joke, as the poor machines stood no chance at all. Even though they could calculate hundreds of times faster than a human being, the result was a foregone conclusion. At the time, in the artificial intelligence departments of universities all round the world, a great deal of effort was expended on chess programming. The reason was that it was felt that a computer would only play chess successfully when it had been programmed to think like a human being, at least in the limited area of chess. And if it was possible to do it in chess, then why not in other areas of human mental activity? Now, however, the academic interest in chess programming has more or less vanished. Nobody ever succeeded in programming a computer to think like a human being, and all the progress in chess programming has been made by concentrating on what computers do best – calculation. Over the past twenty years the speed of computers has increased enormously and now they can calculate millions of times faster than a human being. What the human does elegantly and with little effort by subconscious pattern recognition, the computer does by “brute force”, i.e. by examining millions of possible continuations and finding the right move by an exhaustive analysis probing many moves ahead. Today’s computers are so fast that the fact that this method is very inefficient doesn’t matter – the computer has gigahertz to spare and can afford to waste 99% of its time looking at irrelevant variations that a human would never consider, provided the remaining 1% is spent on the critical lines.
In 1996 the human World Champion, Garry Kasparov, faced the world’s leading computer, Deep Blue, in a six-game match. Most commentators imagined that Kasparov would see off the silicon challenge with little difficulty. There was palpable shock when Kasparov lost the first game. Perhaps Kasparov had underestimated the powers of his opponent; in any event, Kasparov played much better in the remaining games and ended up winning by 4–2. However, Kasparov fared less well in the 1997 rematch and the number-cruncher came out ahead by 3½–2½. Since then, the machines have steadily advanced, and Deep Fritz’s 4–2 win against Vladimir Kramnik in 2006, in which the then world champion failed to win a single game, indicated that the computers had pushed ahead of the top humans. After this, interest in man-machine contests faded since it was apparent that an interesting contest could only be obtained by handicapping the computer in some way.
The lessons of computer chess may have wider implications. Even though the efforts of computer scientists to make computers think like human beings have not been very successful, perhaps this should not be a cause for regret. In chess, at any rate, computers have achieved their current level of performance precisely by not thinking like human beings. The result of this has been a new perspective on chess which has not been provided by 1400 years of human development. Because the processes by which computers select a move are totally different to those employed by humans, computers sometimes come up with an astonishing move which a human would never even consider. Many times in recent years a computer has played a move which the watching humans have dismissed as the result of a programming bug, only for further analysis to reveal the concealed idea behind the move. In other words, the differing perspective of the computer has provided insights which would probably never have been obtained by human analysis. The simple fact is that nobody knows what human beings are missing, and this applies not only to chess.
The limitations of human thought must be there, but little can be said about them because it is impossible to measure the unknown. First of all, our brains are undoubtedly limited by our evolution. Intelligence evolved as a survival characteristic in a world very different from the one we now inhabit. The slow million-year crawl of evolution may have fitted us well for the world in which humanity first appeared, but by comparison the social and technological changes taking place today are lightning-fast. The fact that slightly intelligent apes have developed world-destroying weapons does not imply an ability to manage that same world in a positive way. Moreover, there are probably culturally-based limitations. Human development tends to build gradually from one generation to the next, and only very rarely does it suddenly shoot off at a tangent. Perhaps thousands of years ago there was a fork in the road of knowledge; humanity sped off along one branch, never suspecting that the other existed.
In the absence of any external perspective, we cannot tell how limited our thinking processes are. If we suddenly make contact with extra-terrestrial intelligences, will their thoughts prove to be along utterly different lines to ours, perhaps even to the point of making communication impossible, or will there be enough common ground to make meaningful contact? Nobody can say. However, the progressive development of computers suggests that one day we may be able to create another viewpoint ourselves. Just as the primitive computers of today have shown us new concepts in chess, perhaps one day their successors may show us what we are missing elsewhere. And the first clue will have been provided by a game from 6th-century India.