|
History of Backgammon ProgramsFollowing last week's post on the computer software that analyze situations in football games in a similar way Snowie and GNU evaluate positions in backgammon, we got some emails asking to elaborate about "the way computers had changed backgammon". And in an unbelievable coincidence, netwalker blog had just published a post about the milestones in computer backgammon, computer bridge and go.
Netwalker review on computer backgammon begins in 1979 with the first defeat of a world backgammon champion, Luigi Villa, by the computer program BKG 9.8, a defeat that was ascribed more to the backgammon computer's luck than to its skill. The review ends in 1979 and ends in 1992 with the development TD-Gammon, the first backgammon neural network that achieved a skill level similar to the best human backgammon players.
No notable milestones have been marked from 1992 till our days, yet during that period of time backgammon computers programmed to analyze backgammon games, evaluate winning chances at any given position and solve cube decisions using statistical means, have become an essential learning tool of every backgammon pro and each backgammon beginner who aspire to become expert.
Interesting enough, the computerizing process of bridge and go was similar to the evolution backgammon programs underwent. At first, the game programs raised suspicions and even mockery (bridge master Zia Mahmood bet on £1,000,000 that he would beat any computer program but went back on his word after playing the GIB software) but eventually even the most arrogant human players got used to the idea that a computer program can become a world champion.
|
