Chess - humans and computers

 



If you’ve ever played chess against a computer, you know the feeling: One small mistake… and suddenly your king is in trouble. Play a computer at full strength—and watch it choose excellent moves instantaneously—and it can be brutal. 

Computers have been beating the best human players for decades, but it wasn’t always this way.

In 1770, Hungarian engineer Wolfgang von Kempelen unveiled a mysterious chess-playing machine, the Automaton Chess Player. It looked like a life-sized figure in Ottoman clothes, seated behind a cog-filled cabinet with a chess board on top.

The machine toured Europe, defeating famous chess hobbyists like Napoleon Bonaparte and Benjamin Franklin.

There was just one problem: Hidden inside the cabinet was a human chess player making the moves. The gears and cogs were just for show, distracting from a human-sized compartment within the machine.

The secret stayed hidden for more than 80 years, until the “machine” was destroyed in a fire in 1854. 

Real progress in computer chess wasn’t made until 1948. After his crucial work cracking German codes in World War II, mathematician Alan Turing turned to a different puzzle: chess.

Together with David Champernowne, Turing created the first chess program, called Turochamp. This algorithm was capable of receiving information about a position and choosing a move using a set of logical rules.

Turochamp was developed long before the invention of electronic computers, so all calculations had to be done by hand. One move could take more than 30 minutes!

Slow, but it proved something important: Clear rules and step-by-step logic could produce real chess moves.


In 1950, mathematician and computer scientist Claude Shannon made the case that chess was the ideal test for computer science. If it were possible for computers to problem-solve for chess, then computer science could be applied to other areas and new questions.

In his research to develop an algorithm for playing chess, Shannon had to calculate a rough estimate for the total number of possible chess games. Today this number is called the Shannon number: 10120 (a 10 with 120 zeros).  

However, Shannon didn’t want to encode every possible game into his algorithm. Instead, he wanted to consider only “reasonable” moves by giving the computer certain parameters about each playing piece, king safety, and the basics of good vs. bad strategy.

 In 1957, the IBM 704 was invented. This early computer weighed 9.7 tons and took around 8 minutes to go through 2,800 positions and choose the next move. This was a huge improvement from Turochamp. 

The 704 computer used eight preliminary parameters to select up to seven moves for deeper examination and calculate up to four additional moves for each of the seven possibilities. This same system of parameters and move analysis is still used in modern chess engines today—just much better and much faster. 

By the 1990s, computers were strong enough to challenge the world’s best.

IBM’s supercomputer Deep Blue faced world champion Garry Kasparov in two six-game matches. In 1996, Kasparov narrowly beat the machine, winning three, drawing two, and losing only one. But in 1997, after major upgrades, Deep Blue outplayed Kasparov, winning two games, drawing three, and losing just one to its human competitor.

Comments

Popular posts from this blog

IBM goes big to keep quantum computers cold

IBM - avances en mantener adecuadamente fríos computadores cuánticos.

Estructura multinivel de los sistemas de agentes de IA