News in Nature that researchers at the University of Alberta have made their checkers program, Chinook, unbeatable. For any possible move in a game played against it, it can look up the best move. The best anyone can do against it is force a draw.
However, the best it will be able to do against a knowledgeable human opponent is also a draw, since all a checkers master has to do is find a way to draw against it, and just play that game whenever they face it. So paradoxically, by finding the best moves, it has made itself weaker — by becoming predictable.
The researchers expect chess to be solved within the next 50-60 years.
Random thoughts about the potential for chess being solved…
Add new pieces? David Gerrold, in his books about Earth being invaded by alien spores, had a modified chess with new pieces such as a Time Traveler, where you would take a piece off the board for a specified number of turns… always thought that was a cool idea.
Change the board? The popular 80s computer game Archon had pieces that gained or lost power depending upon the color of the square they were on.
Hack the database? Spock once found the Enterprise’s computer had been hacked when it stopped playing a perfect game of chess.
Make the computer an insane killbot? In 2001: A Space Odyssey, HAL 9000 enjoys playing chess against people. This being six years ago, of course, chess hadn’t been solved (researchers, after all, expect it to take until 2070!). You’d think HAL could have solved it in his spare time… you know, when he wasn’t plotting to kill everyone, and after all, how long could that take?
Â
1 thought on “Checkers Solved!”
Comments are closed.
“want my advice? let HAL win”
“why? you don’t see humans get upset when they lose”
“that’s cause humans don’t throw people out of airlocks when they lose. The HAL 9000 series is known to do that.”