Meanwhile cautious-minded scientists are disinclined to assert, even with the prospect of all these fascinating possibilities coming true, that electronic computers can veritably think or that they really give evidence of quasi-human intelligence. The experts express their opinions more carefully. Computers are said to "simulate human intelligence." They pretend; in other words, they put on an act. But after all there are many people, too, who content themselves with an imitation of human intelligence, and they get along nonetheless quite well. Simulated cleverness is enough for a brilliant performance in a chess game. Chess-playing automata are nothing but calculating machines which have been taught the moves each piece is allowed to make, and informed that the point of the game is to checkmate the king. Then it is left to the calculating mechanism, the arithmetic unit, to work out the possible countermoves it can make to each move made by its opponent, to think a few steps ahead, weigh all the pros and cons, and finally make the move which looks like it, being the successful one. Such a computer will I never be very imaginative; it will always play very coolly. A first-class chess-player accustomed to deceiving his opponent about his tactics will generally beat the computer. But a middling player will be worsted, because he cannot calculate and balance possibilities as quickly and thoroughly as the automaton can.