Engine programming

[abulmo2]

(except for lack of underpromotion and legality checking, but pretty close...)

I would even say the lack of promotion. As the promoted piece is not displayed when played, some chess interfaces (cutechess-cli, for example) consider the move illegal and give the game to the opponent. micromax is unfortunately almost impossible to use into any tournament manager.

[hgm]

Oh, you are right! And the amazing thing is that you are the first one to ever complain about this.

This is completely unintentional. The micro-Max code I showed here, and on which the character count is done for determining Elo/char, is a stand-alone program, and doesn't print any moves at all: it shows what it played through printing a board diagram of the new position. And after a promotion the piece will certainly be shown as a Queen.

So it is just a consequence of an oversight in the WB-protocol driver that I fused to the stand-alone version to make automated testing against other engines (by me and by CCRL) possible. I always test under WinBoard, and this is very forgiving towards the format of the input move. If no promotion character is present in a Pawn move to the last rank, it defaults to promotion to Queen.

So WinBoard definitely is one tournament manager for which playing micro-Max is no problem at all, which is why I never noticed the non-compliant promotion move.

I wonder how CCRL managed to test micro-Max without ever running into this problem. I cannot imagine they only use WinBoard.

BTW, Fairy-Max doesn't suffer from this problem, and Fairy-Max 4.8 should be functionally equivalent to micro-Max 4.8. I think I fixed that for the benefit of engine-defined chess variants, for which WinBoard would have no idea what the default promotion piece should be. (The Queen might not even participate.)

[Michael Sherwin]

I wrote my first chess program with QuickBasic 4.5 long time ago : very weak, very slow
Then I switched to C , my préférence.
As bob wrote : it is your personnal choice ,

HI Daniel, If you still have that basic code you can compile it in QB64. It is supposed to be QB4.5 backwards compatible and compiles to C++ and then using MINGW compiles to a native exe. Just one note about speed-- DO loops run faster than for loops or while loops in QB64. It would be interesting to know how much faster it is.