Wednesday, April 20, 2005

Some more chess blabber

I just found a minor flaw in Chess 960. In some initial positions of the game, one or two of a player's pawns are initially unprotected (in most initial positions, including the normal chess one, all the pawns are protected at the start of the game). It's caused by the weird range of movement of the knight pieces, such that they don't protect the pawns in front of them. This gives white a very slight additional advantage over black. In normal chess, white already has the advantage of moving first, giving him the initiative at first. This flaw in Bobby Fischer's invention gives white the extra advantage of potentially being able to attack the unprotected pawns first. I'm still figuring out whether this flaw is worth addressing. And if it is, what's the most elegant way of addressing it?

Now on to other news. You might have read about my new chess variant. Or maybe not. Anyway, to recap, I created a game almost exactly like chess, but with one exception: when and only if a king is in check, the existing knights of the defending player could move to any square [n,n-1] away from their current position to remove the checking threat (note that the normal knight moves in a [2,1] fashion). I call this event in the game "la croisade" (meaning, "The Crusade"). I've been studying it for a while now, and I found that contrary to my previous thought, there is a way for la croisade to be used for checkmating. This can be done with two steps: (1) Place the knight such that it would threaten the opponent's king once a la croisade happens. (2) Check the king (with a different piece, of course) such that the only way to escape the check is by checking the other king. Here's a sample game (see the final position here):

1. d4 d6 2. c4 e5 3. Nf3 Qe7 4. a4 Na6 5. g3 Bg4 6. Bg2 exd4 7. Nxd4 c5 8. Nb5 Qxe2++

In the final position, the black queen makes the check. The king can't capture the queen because of the bishop on g4 protecting it. But if the white queen captures the black queen, the white queen would be checking the black king (which is conveniently positioned behind the black queen), thus allowing the knight on a6 to check the white king (this is what I call the "Crusader discovered check"), and thus, checkmate.

Notice that it is technically the queen that made the mating move, not the knight, so fears of unfair knight advantage is still quite unfounded. One would think that people playing my chess variant (which I'm temporarily calling "Crusader Chess") would be far more aggressive against knights than in normal chess, but really, it isn't such a big deal. I predict that the defensive role of la croisade would be a lot more prevalent in Crusade Chess games than the mating role, and that mates involving the augmented knight would be far less common than even en passant moves (which is pretty rare).

Lastly, I just recently discovered the site, which has a titanic searchable database of recorded games as well as player info (here's the page for the current youngest grandmaster, Magnus Carlsen). You could discuss individual chess games via the kibitz forums, and even answer regular chess problems. Enjoy. :-)