TalkChess.com

Steve: thanks for all your work on this project.

It's a win in 65, the winning move being 1 Bc6+, assuming the 50-move rule doesn't come into play. 1 Bc6+ A must so that White can get the other bishop and the knight out of danger next move. 1... Kb3 This move lasts the longest. With the other two moves, White plays Bd2+ and extricates the pieces. ...

The man was a top flight player in his day, a known and respected theoretician, and one of the great chess writers and authors of all time. His content was always fresh and interesting. He was also a gentleman who didn't blow his own horn nonstop like many do today.

R.I.P., Mr. Evans.

The solution was to change the search order to XPK. According to what I was able to read on the other forum, that searched first for a solution in which the defender's replies were limited (much as MaxMoves=1) does. Again, it took less than a second. I would imagine that reading the English.txt file...

I know Chest solves this fast... http://www.playwitharena.com I'm not a programmer, and I've only been using the UCI engines for a couple of months, so I don't understand all the parameters. I let Chest have a go with this problem at the default settings (except setting Hash=128) and waited 10 minu...

Some engines have problems with Kingside castling in Arena, which gives a Ke1h1 error. Not being a programmer, I don't know why this is, but I've seen other programmers talk about this issue.

Apologies if everybody is already aware of this.

This is an elegant problem. Unfortunately, it isn't one that any of the current engines are likely to solve, unless they can do so by brute force. 1 Qc6!! An obvious move, even if one doesn't see what White is going to do next, but the move was ignored by the engines, which were all enamored on 1 Qg...

There are six initial moves that mate in 20 from this position, so the engines are correct.

It's impressive for a new engine. I'm sure it will get better in the future.

It's not a matter of finding the "right" move. The difference in move choice is due to the evaluation of the position by the engine. I ran this position through various engines. I always used 64 bit and 2 CPUs when possible, and set the hash to 128 mb in most cases. According to most engines, the tw...