COPElAND & PRINz | 333
If I were to sum up the weakness of the above system in a few words I would describe it as a
caricature of my own play.
In his lecture Kasparov tended to confuse the earlier and later versions of Turochamp. These
historical confusions, though, didn’t prevent him from wiping the floor with a re-creation of
Turochamp Mark 2 when he played against it on the stage at Manchester; the game is set out
later (Game 4). But Kasparov described the re-creation that he played—the ‘Turing Engine’
coded by the German chess software company ChessBase—as embodying the ‘first program
ever written, even before computers that could run it were invented’. But it was the Mark 2, not
the Mark 1, that ChessBase re-created. No recorded games with the Mark 1 survive, and all
that is known about the Mark 1 is Champernowne’s rather vague description quoted earlier.
What is more, although Turing and Champernowne created the Mark 1 at a time when the only
stored-program computer in existence was the tiny Baby, the Mark 2 dates from the period after
a computer capable of running it was installed in Turing’s Computing Machine Laboratory (the
Ferranti Mark I). What Kasparov should have said, if he wanted to be completely accurate, was
not ‘first program’ but ‘successor to the first version of Turochamp, written at the time that the
first big electronic digital computers went on the market’.
As Kasparov’s game with the Turing Engine highlighted, Turochamp Mark 2 is no
champion-beater. Nevertheless, the pioneering Mark 2 included much that has become stand-
ard art in chess programming:
• evaluation rules that assign numerical values, indicative of strength or weakness, to
board configurations
• variable look-ahead: instead of the consequences of every possible move being followed
equally far, the ‘more profitable moves’ are ‘considered in greater detail than the less’,
Turing explained
• heuristics that guide the search through the ‘tree’ of possible moves and counter-moves
• the minimax principle
• quiescence search: the search along a particular branch in the tree of moves is discon-
tinued when a ‘dead’ position is found—a position with no captures or other lively
developments in the offing.Turing also realized the necessity of using ‘an entirely different system for the end-game’. In
addition, he gave a prescient discussion of how to make a chess program learn from experience.
He advocated the use of what we would now call a ‘genetic algorithm’ or GA, saying:^16
[A]s to the ability of a chess-machine to profit from experience, one can see that it would be
quite possible to programme the machine to try out variations in its method of play (e.g. varia-
tions in piece value) and adopt the one giving the most satisfactory results. This could certainly
be described as ‘learning’, though it is not quite representative of learning as we know it.
GAs are now widely used in areas as diverse as codebreaking, hardware design, pharmaceuti-
cal drug design, and financial forecasting. As mentioned in Chapter 25, GAs are based on the
principles of Darwinian evolution—random mutation and survival of the fittest—and some
researchers even employ GAs as a means of studying the process of evolution itself. Turing’s
not-very-catchy expression for his invention was ‘genetical or evolutionary search’.
Arthur Samuel, an early pioneer of AI in the United States, independently reintroduced
the idea a few years later in his famous checkers playing program, a remake of the Strachey