The Most Unusual Move in Chess (or "How I won a beer in my lunch break")

| 16

A couple of days ago I posted on my Facebook wall about a game from the fourth round of the strong Hoogeveen chess tournament in the Netherlands. Watching the games live from my office, I saw Dutch Grandmaster Erwin l'Ami play the incredible 36.Bh8-a1!! against Polish Grandmaster Michal Krasenkow. Why was this Facebook-worthy, I hear you ask? Well, think about this move: The white bishop was located in one corner (statistically the worst place for a bishop) and moves, without capturing anything, to the only other square that is equally bad: the other corner. In fact, I seemed to recall Christian Hesse writing in his fantastic book  The Joys Of Chess that a bishop move from one corner to the other, without capturing anything, is statistically the least likely move to appear on the chess board.

I vaguely remembered him coming up with this result from an exhaustive database search of existing games, and I posted this claim on Facebook. However, it was immediately challenged by IM Simon Ansell, who claimed that "Surely bxa8=B is somewhat less common!". A quick flick through Hesse's book failed to uncover the relevant section I apparently recalled, which got me thinking that Simon might well be right. But what really got me interested was when he challenged me that "If you can find a game where bxa8=B+ was played, or even make up a game where it's plausible, I'll buy you a drink at the next 4NCL." Well, what should we say to that?

Unfortunately, ChessBase doesn't have the capabilities to allow its users to search for criteria such as this. There is apparently a funky command-based computer program called Chess Query Language that allows for more sophisticated searches of chess positions...but also requires a more sophisticated operator. That wasn't going to happen, at least not over my lunch break. So instead I decided to try and compose a chess study where 1.bxa8=B+ was the only solution. Why a chess study? Well, let's dissect just why this move is so unlikely to occur. It's not that it involves a promotion with a capture - unusual, sure, but not THAT unusual.

No, the real reason is the promotion to a bishop, with check. Promotions to a queen are of course the most common, as her majesty is the strongest available piece. Promotions to a knight aren't altogether uncommon because the knight can move in different directions to the queen, so there are circumstances in which it's preferable. But a queen can do everything a rook or a bishop can do, and more, so these "underpromotions" (promotions to anything other than a queen) are less common. Promotion to a rook can sometimes happens right at the end of the game in king-and-pawn endgames when getting a queen would give stalemate, but a rook allows Black's king enough wiggle room to ensure White can avoid the drawing trap and claim victory.

But a bishop? Well, it's also possible that it could be done to avoid stalemate, but it's a lot less common. And in this case it's an underpromotion with check, meaning that if we'd instead chosen a queen, it would still be check. That means it wouldn't have been stalemate anyway! So this makes things really tricky. Why would White ever want to choose a bishop, a strictly less mobile piece, instead of a queen, given that it can't be to avoid stalemate?

The answer came to me over lunch. White wouldn't do it to avoid stalemate, but she might consider it to enable stalemate - in this case, to get herself stalemated. Imagine a situation in which promoting to a queen will lead to a loss, because Black's army would still vastly outweigh White's forces. Perhaps White could instead take a bishop, then subsequently imprison it (fortunately, the corner is the easiest spot to incarcerate a bishop), and force a surprise stalemate to save the day.

Once the idea was in place, the rest was just a matter of finalising the details. Because the first move is a check (a cardinal sin in chess problem composition, but in this case obviously part of the task!), the study will never get published or win any awards. But then again, how many problem composers win a beer for their efforts?

I hope you enjoy it. Simple and uncomplicated, but cute nonetheless. That's why I'm calling it my Penny study.

(If you're having problems solving it, or you want to see the full analysis of the solution, you can check out