What do you think? Good? Bad? Ridiculous?
It seems to me that with the move 1. a3, White is saying to his opponent, “I will agree to play Black, and I believe that in any opening you might choose to play, I will be able to find a variation in which a3 is a useful extra move.” It’s a little bit like playing the game of GHOST, in which you say a letter and then your opponent says a letter and your objective is to force him into completing a word. The “strategy,” such as it is, is simply to have such a good knowledge of vocabulary that whatever he says, you can come up with a response that gets him into trouble.
I remember reading once, long ago, that 1. a3 works pretty well against the response 1. … e5, because then White can play 2. e4 and go into a double e-pawn opening. After 2. … Nf6 3. Nc3, Black’s options are already restricted. He can’t go into a Ruy Lopez reversed because of White’s 1. a3. This is the first success. Not only that, if he plays a Scotch reversed, 3. … d5, White can play the variation 4. ed Nxd5 5. Qh5!? This basically forces Black into sacrificing a pawn. With colors reversed the gambit is sound, but if he wants to follow the main variation, Black needs to play … Nb4 at some point. Here … Nb4 isn’t available because of White’s 1. a3.
And the litany goes on. If Black instead opts for an Italian Game formation with 3. … Bc5, White can go into a Two Knights Defense reversed, with 4. Nf3. Now many of Black’s normal options are taken away by the pawn on a3. For example, the normal move 4. … Ng4 runs into 5. d4 ed 6. Na4 and oops, Black can’t play 6. … Bb4+.
So for people like me who normally play double e-pawn games as Black, 1. a3 starts to look like a really good option. But what if Black plays 1. … d5? Well, here too there are some interesting possibilities. White can play 2. d4, and now it’s questionable whether Black can get away with the Queens’ Gambit reversed, 2. … c5, because White just takes the pawn. After 3. dc e6 4. b4, it seems as if White might be able to just hold the extra pawn. If so, Black might have to play a more stodgy move like 2. … Nf6 or 2. … Bf5, going into a London System reversed.
In fact, I think that this may be the true argument against 1. a3. Instead of trying to play as if he were White, Black should instead continue to play like Black. He should just try to equalize with solid, unspectacular moves. That includes variations like 1. a3 d5 2. d4 Bf5, and it also includes even more modest variations like 1. a3 g6 or 1. a3 Nf6.
And if worst comes to worst… Black can always play 1. a3 a6!? and pass the onus of the first move back to White! Or he could wait a move, and play for example 1. a3 e5 2. e4 a6.
To satisfy my curiosity, I looked through ChessBase today to see if there are players who adopt 1. a3 on a regular basis. Most notably, I found three Magnus Carlsen games… but on closer inspection, two of them were blitz games and one was a blindfold game from the Amber tournament (against Ivanchuk, and Carlsen lost). It seems as if 1. a3 might be suited to Carlsen’s style — “go ahead, play any opening, I don’t care” — but even he hasn’t used it in a full-length tournament game.
There are some 2300-2500 level players who do play it often. Ivanko Krecak, a correspondence player, plays 1. a3. Eric Prie and Anatoly Sidenko (the latter also a correspondence player) likewise have used it in many ChessBase games, although they prefer the move order 1. d4 d5 2. a3.
Would I ever play 1. a3 in a tournament game? Well, I wouldn’t rule it out. I think that the main reason I have never done so is that if I played 1. a3 and lost, I would feel as if I had given away the advantage of the first move for no reason.
P.S. I noticed on Wikipedia that 1. a3 is called “Anderssen’s Opening,” because Adolf Anderssen used it three times in a match against Paul Morphy. A pity, because I was going to call it the Mahgnimrib Opening (Birmingham reversed). My friend Mike Splane experimented with the Birmingham (1. e4 a6) for quite a long time, although I think he eventually became disenchanted with it.
