63 ( +1 | -1 )
Can you explain this ?
1.a3 is a really poor move, Chesslab database suggests that only 28% of such openings are successful and 44% result in defeat. However, 1.h3 is a successful opening and has 39% success with only 38% failure.
Why should the Kingside flank opening be far superior to the Queenside one ?
Whilst I have been known to play 1.a6 in serious games (for St George, Tony Miles and England !!) I doubt I will ever play 1.a3 although White statistics indicate that I should from time to time seriously consider playing 1.h3. Perhaps this shock move would be good in blitz games ?
23 ( +1 | -1 )
Pure percentage statistics can deceive. The ratings of the opponents in question also need to be examined. 1.h3 is sometimes played by strong players in correspondence games so their inferior opponent cannot "equalize" by Chessbase usage.
64 ( +1 | -1 )
Neither is a particularly wodnerful move; the game statistics simply aren't that meaningful. How many of those 1. a3 or 1.h3 games were played by strong masters against other masters of comparable strength, for instance? Not too many, I'd wager. Both moves probably lead to easy equality for Black.
I also don't see the shock value, myself. They're basically just pass moves when you get right down to it, so Black can just play 1... e5 or 1... d5 and while of course there will be some differences, it's really not all that far of a stretch to just apply elementary opening principles associated with 1. e4 or 1. d4 as White.
10 ( +1 | -1 )
.....may be that 1.h3 is more liable than 1.a3 to tempt Black into a rash, premature attack.
54 ( +1 | -1 )
is usually played with an idea to undertake an 'improved' Grob opening, while 1. a3 usually has no real merit after 1... d5, as in some variations the move is not really necessary. All this leads to a cyclical effect where you get people like Grob playing 1. h3 and people like Amateur playing 1. a3 :) Neither is really a serious opening (although I know Grob and a few others have a lot of success with 1. h3), so any unstratified statistics you find are likely to be inaccurate (unlike, say, statistics about 1. e4 c5, which is used by a huge population).