Jun

18

Tinsley vs. ChinookI was having a discussion with a colleague on the topic of Chess vs. Checkers. Somewhere I had the impression that Checkers had been "solved" –that it is ultimately an elaborate version of tic-tac-toe, i.e. there is a well-defined correct move to make in every situation. Chess though is different, as I understood it–there is no known correct way of playing in every situation, either because it can't be known in principle or because the computers just haven't found it yet. Can someone set me straight on this topic? (Background: I haven't played chess or checkers in over 30 years, but I am quite good at tic-tac-toe.

Nigel Davies weighs in: 

As I understand it there is no 'solution' as such to either game and that with checkers in particular it is quite easy to make it considerably harder by playing on a larger board and with more pieces (one can also play 'big chess', though this looks somewhat artificial to my eye). With regard to board games being 'computer proof' it's also worth checking out Shogi and (especially) Go where computers are still rather mediocre compared to the best humans.

From the point of view of educating children all of these games are wonderful in that they can teach the young to falsify their own ideas. In order to play 'well' one must find out what's wrong with a move before playing it on the board.

One major consideration in the choice of game might be the number of opponents to be found. In the West at least I believe this is where chess shows to advantage.

Hope this helps. 

Pitt T. Maner III writes:

TinsleyDr. Schaeffer wrote an appreciation of one of the best checker players ever, Marion Tinsley, who actually liked the challenge of facing a computer (nicknamed Chinook).

After Chinook's first game against Tinsley in 1990, we started analyzing the game. Tinsley began recounting the history of the line we played, recalling games he played in the 1940's! The move sequences flowed easily from him without hesitation, sometimes annotated with the name of the opponent, date or place where the game was played! 1947 was as vivid in his memory as if it were only yesterday. The second facet to his play was his incredible sixth sense. A glance at a position was sufficient to tell Tinsley everything he needed to know. For example, in 1990 Chinook was playing Tinsley the 10th game of a 14 game match (won by Tinsley 1-0 with 13 draws). I reached out to play Chinook's 10th move. I no sooner released the piece when Tinsley looked up in surprise and said "You're going to regret that". Being inexperienced in the ways of the great Tinsley, I sat there silently thinking "What do you know? My program is searching 20 moves deep and says it has an advantage". Several moves later, Chinook's assessment dropped to equality. A few moves later, it said Tinsley was better. Later Chinook said it was in trouble. Finally, things became so bad we resigned. In his notes to the game, Tinsley revealed that he had seen to the end of the game and knew he was going to win on move 11, one move after our mistake. Chinook needed to look ahead 60 moves to know that its 10th move was a loser. In my experience with tournament chess and checker players, the sixth sense is experience. It is well-known how intensely Tinsley studied the game, analyzing anything from a Grandmaster game to a game between novices. His uncanny ability to know good from bad and safe from dangerous, is the direct result of all his hard work. Strong chess players have the same ability, but perhaps it is not quite as evident as it was with Tinsley .

Nigel Davies writes:

Seems like we get a whisker away from quite deep philosophical questions. My personal belief is that the goal of 'replacing humanity' in the cause of 'efficiency' is a deeply flawed one. It always feels to me like the attempt to show that computers can 'play' these games much better makes our attempts at self-improvement appear futile, an idea which many people will buy into. Is it too fanciful to suggest that they represent a 'greater goal' of being looked after by machines whilst humans have mindless 'fun'? Nigel Davies

David Hillman writes:

This is not unlike giving up the warm, tactile sensation of the paper page in a book for the slick plastic of a Kindle, or the daily newspaper's beautiful scent of cheap pulp and ink replaced by the netbook's display. The aromas of silicone and polymers do not mix as kindly with the scent of espresso wafting on the morning air. My own livelihood is derived from computer-based industrial productivity and efficiency systems, but my life is kept on a yellow legal pad with a #2 pencil. Balance, always balance. To paraphrase Queen, "we need it all and we need it now." The Deep Blue's, Chinook's, etc. may be wondrous, but there is simply no mineral nor petrochemical-based substitute for the hug of a happy child, for the lap of a caring spouse upon which to lay one's head at the end of a bad day, or for the twinkle in a grand-master's eye across the chessboard when he mates you in 6 moves.

Nigel Davies responds:

I don't think it's the same thing David. An analogy with having a kindle versus a book would be to play chess against a human via your PC. Having computers do the playing and trying to demonstrate their 'superiority' is more like having them write the books, and purportedly do it more efficiently than humans; fewer words for the same meaning perhaps, 'War and Peace' reduced to 10 pages.

Chris Tucker agrees:

I agree with you completely Alan, my point is just that programmers are not out to replace us completely (yet, anyway), but they are out to codify decision making. Games are a good place to do this because the rules and possible moves are very limited, even though the number of possible outcomes can be astronomical. The arena is structured and they can test and validate their ideas within this framework. The idea of game playing is much deeper, philosophically, (as Nigel suggests) than most care to admit. I will leave that bit for you two to explore. Machines that can replace the humanity of squaring off with an opponent do not exist, there are simply too many levels of interaction there. 

Nigel Davies replies:

Chris, there is no decision making in the programs or any attempt to replicate human thinking, they simply use brute force to analyze all the possibilities (with chess slapping in a primitive evaluation function) and the mathematical limitations of the games enable them to get away with it and 'win'. Perhaps when they started out the intention was to create 'artificial intelligence', but I don't see that this claim can be maintained given the route they have taken. Looks like an ego driven attempt to 'beat mankind' of the type which enables a car to go quicker than someone on two legs. 

Dave Bacon addresses the original question:

I believe Checkers on a standard sized board has indeed been solved. The reference is Science, Sept. 2007, Vol. 317. no. 5844, pp. 1518 - 1522.

“Checkers Is Solved” Jonathan Schaeffer, Neil Burch, Yngvi Björnsson, Akihiro Kishimoto, Martin Müller, Robert Lake, Paul Lu, Steve Sutphen

The game of checkers has roughly 500 billion billion possible positions (5 x 10^20). The task of solving the game, determining the final result in a game with no mistakes made by either player, is daunting. Since 1989, almost continuously, dozens of computers have been working on solving checkers, applying state-of-the-art artificial intelligence techniques to the proving process. This paper announces that checkers is now solved: Perfect play by both sides leads to a draw. This is the most challenging popular game to be solved to date, roughly one million times as complex as Connect Four. Artificial intelligence technology has been used to generate strong heuristic-based game-playing programs, such as Deep Blue for chess. Solving a game takes this to the next level by replacing the heuristics with perfection.


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search