The American Checker Federation

I have been tinkering around over the past 10 days with Cake, and made some changes to make it able to solve Brian's problem. Below you can find my latest version's output - on a single search thread on my laptop it finds the solution in just 30 seconds. It will not show the pitches on the PV, becau...

The leading computer programs have been optimized to play the best possible "normal" game of checkers. They have not been optimized to solve problems such as this one. If we wanted our engines to be able to solve such problems, it would very likely be possible - but it would hurt their per...

Thank you for your kind words Alex - and good luck in your upcoming match!

best regards
Martin

Ed Gilbert and I have both published new versions of our checkers engines, KingsRow (1.19d) and Cake (1.89f). In addition, Ed has updated CheckerBoard (1.76) to keep track of game history better so that the evaluation of the engines is more accurate in analysis mode if a repetition is possible. You ...

Not sure this is a good idea. Anyone could just easily figure it out with a program in a matter of seconds. In any case, I will take a stab at it. 1.5 days of computation with Cake 1.89e with full 8pc database gives +4 with no convincing line. Looks like this is a nice problem to confuse the progra...

Dear all, back in 2002 at the computer checkers world championship in Las Vegas, Vic Habgood showed me something which I think was called the Whilter smother, where one side - let's say it was white - ended up with pieces on 6,7,10, and a king on 2 vs red pieces on 1 and 3. I didn't write down how t...

...and for good measure, I recompiled Cake Sans Souci (1.04k) for the current 64-bit platform. Back in 2003 this was probably the strongest checkers engine around.

Cake 1.89d - Cake Sans Souci 1.04k: +73 -0 =215.

Amazing!

Small correction: the last "traditional" version of Cake was 1.85, and I repeated the match and now it's not quite as bad, but still extremely lopsided:

Cake 1.89d - Cake 1.85: +42 -1 =245 (1s/move, 8pc databases, modern laptop).

I have released an updated version of Cake; 1.89d now finally uses the full pattern-based approach that Ed Gilbert brought to English checkers about 2 years ago. To give an idea on what kind of difference this makes, here's a match result against Cake 1.81, the last published "standard" ve...

Hi John, the 288 games that you get like this at such time controls will mostly be draws. The last time I tried this I got +5-6=277 for my current development version of Cake vs KingsRow 1.19b (5s/move, 8-piece databases, no books of course). With such low numbers of decisive games, a lot is just pu...

Hi John,

when you say "wins" or "loses", what kind of matches are those? (Openings, number of games, time controls, hardware)?

best regards
Martin

Hi John, I do still have some of those old engines, but they are all compiled for 32 bit, so you would need to have a 32-bit CheckerBoard install to use them. For some of them, I still have the source code, and could maybe even make a modern version which would run in a 64-bit environment. Then agai...

I didn't see this mentioned anywhere in this forum, so thought I would post it: A bit more than 2 years ago, Ed Gilbert introduced a machine-learned evaluation function in KingsRow, making it far stronger than any other checkers program. I have done something similar for Cake (but on a lower level -...

Hmm, no luck... - anyone else interested in the books? First one to send me a mail with his address gets the books!

cheers
Martin

Aloha! many many years ago, George Miller was so kind to send me all six volumes of Derek Oldbury's Checkers Encyclopedia - since I have long stopped working on my checkers programs, I would like to either return the books to George, or pass them on to someone who has a better use for them than me. ...