This is not really a fair criticism of Larsen. Although this game is the one everyone remembers, Larsen actually scored 2.5/3 in his other three games. He scored a win and a draw in his other two games against Spassky, thus getting an even score in the three games they played. The Soviets then substituted Stein for Spassky in the last round, apparently fearing that Spassky, the reigning world champion, might lose that game, and thus the match, to Larsen. Larsen played very dubiously against Stein, who missed a forced win, but nonetheless Larsen won the game. Larsen's 2.5/4 on first board was not quite as good as Fischer's 3/4 on second board, but was nonetheless an excellent result.

I don't know when this analysis was written, but Fritz 10 found 11...h5 at only 13-ply (on my comp, in about 20 secs). Needless to say, 12...h4 and 13...hxg3 were found even more quickly.

And how's this for blasphemy? Fritz says Spassy did not play the best move in 14...Rh1!

Bent Larsen - Boris Spassky, Beograd 1970

click for larger view

Analysis by Fritz 10 (20-ply):

1. (-12.07): 14...Bxe3 15.dxe3 Rh1 16.Rxh1 g2 17.Bd4

2. (-7.31): 14...Qh4 15.gxf5 Qh1 16.Rf1 g2 17.Nc3 Qh4+ 18.Kd1 gxf1Q+ 19.Bxf1

Blasphemy indeed. To paraphrase David Bronstein: However, it is hard for chess lovers to agree. Moves like 14...Rh1 are not easily forgotten.

Your pawn structure puts too many pawns on white squares; meanwhile, the fianchettoed bishop on b2 will be hitting the dark squares, and these two don't go together well. With a Bb2 you want, ideally, to play moves like f4 that fight for the long diagonal.

There are also some tactical problems. Play it as described and you could be lost almost before you start. For example:

1.e4 e5
2.d3 Nf6
3.c4 Bc5
4.b3?
actually loses: Black plays 4...Bd4 and white must lose material.

OK, so you wouldn't continue with your b3 plan after 3...Bc5 -- but a plan that's so easy to derail is a dodgy one.

Also, when facing passive moves like 2.d3 most black players hit back fast with ...d5, often not waiting for ...Nf6 to support it (as it lets you play c4, controlling d5). What you really need with this pawn structure is a bishop on g2, hitting the long diagonal and the white squares: try g3 and Bg2 instead, and watch out for tactics. Your b3 plan leaves too many holes on the dark squares and fails to fight coherently for the centre.

Historically, btw, the b3 opening was used by Nimzowitsch in the 1920s: he used to play 1.Nf3 or 1.f4 (preventing ...e5 in either case, and hitting the dark squares); then playing 2.b3 to increase pressure on the a1/h8 diagonal.

Then, about 40 years later, Larsen began to use 1.b3 without preparation. Its main drawback is that black can play 1...e5 and form a solid centre, but Larsen showed that this could be undermined.

Ray Keene wrote a good book on both systems, called The Nimzo-Larsen Attack -- still the best on the subject, as far as I know.

But if you insist on playing 1.e4, be careful about fianchettoing - you can get in a nasty mess with an open centre and slow development.

click for larger view

14 ... h8-h1!!

click for larger view

<notyetagm: 14...h8-h1> I agree is the most off-the-wall, out of nowhere, amazing, beautiful move.

I especially enjoyed it while playing guess the move, because for me it was so unbelievable that I had to sit for many minutes trying to convince myself that it was even sane.

In fact a brain has to be ticking near the edge of insanity to conceive of such a move.

At any rate, I can only think that Larsen must have been almost as dumbfounded when Spassky played it. Is there any way at all to recover? Larsen certainly didn't find it.