Anybody know the percentage of master or grandmaster level games where neither side castles? I had a game recently, of no particular importance, but noticed that nobody castled.
Just curious.
On a personal note, its very rare in my games.
10% have opposite-wing castling.
Or, maybe 10% of games where both players castle have opposite-wing castling.
I cannot remember precisely the context, although I remember the 10%.
Currently have 6,422,446 games in my copy of Mega Database 2015 from ChessBase.
In 1,797,835 games both players are rated >= 2200.
In 69,009 games out of the 1,797,835 neither side castles.
Based of these numbers, 3.8%.
Regards,
John
Thanks. I’m not surprised that the percentage is under 4%, considering that castling is often done in the opening to early middle game stages of play. I have on rare occasions, as with most players, castled in the end game stage for a real zinger, in which the opponent long though I was beyond being able to castle. 
Although it doesn’t mean I won the game, it does have a bit of a shock value when played so far into the game.
In about 1970 I was directing a CCA event in Florida when John Nardandrea pulled off …0-0 mate. The white king had walked up to f6 in a late middlegame / early endgame. …Rf8 alone would not have been mate because then white could have played Kg7.
I saw it coming a few moves ahead, and started staring at the players’ scoresheets (discreetly, I hope) to make sure castling was still legal, just in case the opponent objected. When it happened, the opponent exclaimed “That’s still legal?!” but did not object. I got a nice laugh out of that one.
Bill Smythe
Famous games? Tal-Botvinnik game 1 of the 1960 World Championship came to mind.
At a Microsoft Chess Club event circa 2006, a fellow giving a simul told an opponent - “I’ve been waiting 20 years for a chance to do this!”. He played OOO for mate.
Would Rd1 alone also have been mate?
Bill Smythe
If the answer is “no,” Black king would have to be on d3.
Also true if Rd1 is mate & castling long isn’t.
Problemists, is this a mate-in-three theme? Or helpmate in two with two solutions?
Or g1 or h1.
Bill Smythe
duh
The bolded sentence overlooked epaulet mates on the d-file with black having a knight on a2 or a bishop on a3 or b2 that is blocked from retreating to the d-file. It also overlooked the possibility of the QR or the K having previously moved. In those cases Rd1 is mate and 0-0-0 is illegal.
I was assuming both moves were legal. But Bill Smythe already noted the real problem with my false claim.