Frequently asked questions about the match:
What is Othello? How to play the game?
These questions are answered here.
How does your event compare to the Kasparov-DeepBlue match?
Similarities
- Chess and Othello are in the same abstract class of games: players have perfect information about the current position, their legal moves, and the move history. There is no information hidden and no chance involved.
- The current World-Champion is playing against one of the strongest computer programs.
- The match result will tell whether the strongest programs are better than human players.
- The search depth of Logistello and DeepBlue is comparable: both look about 13+
half-moves ahead excluding search extensions.
Differences
- Chess is much more popular than Othello. However, Othello is very popular in
Japan where most World-Champions came from.
- Othello is simpler than chess - at least for computers: in a
typical Othello position the player to move
has about 10 legal moves and the game is over after 60 half-moves; in chess
there are about 35 moves available in average and a game can last much longer.
- Thinking ahead in Othello is harder for humans than in chess because moves can
alter the position drastically and missing a disc-flip can be disasterous.
- Logistello is running on a plain PC whereas DeepBlue features hundreds of special-purpose chess chips which makes it much faster and much more expensive: Logistello visits about
160,000 positions/sec in midgame when running on a PentiumPro/200MHz PC whereas DeepBlue
searches more than 200,000,000 positions/sec.
- Logistello taught itself to play Othello. It has extracted game knowledge by analysing
millions of example positions and has built an opening book by playing against itself.
The author doesn't dare to interfere with Logistello's opening
choice because he is a lousy Othello player. On the other hand
in the DeepBlue team chess experts were of great help for tuning the evaluation
function and preparing openings. By now it's not clear to the public if or which machine learning techniques were used to tune parameters of the evaluation function
automatically.
- There are hundreds of Logistello's games available enabling humans to prepare against
the program. Furthermore, all "secrets" of Logistello have been published but
virtually no information about DeepBlue was available to Kasparov for preparation.
- Prize money for the chess event was $1,100,000 - ours is $4,000.
If the match ends in a draw (3 to 3), do you plan to have an additional game?
If the match ends in a draw, how to divide the award?
How will Mr. Murakami play the game? Will he face the computer
moniter and use keyboard to enter his moves? Or will he sit in front
of the Othello board, and you will place Logistello's moves?
We'll play on an ordinary Othello board. An operator has to place and flip discs for
Logistello.
If either one wins first 4 games, do you still play the remaining
games?
Yes - though unlikely - the end-result could be 4-2 which is very different from
6-0.