This received the "Most Interesting Game" prize at the <30th USSR Championship 1962>.

I find unbalanced positions fascinating, so I thought that I would take a second and deeper look at this game. Several have commented that they thought that Tal had all the winning chances but at first glance I disagreed. While at the final position material is roughly even, usually the 3 pieces, if properly coordinated, will typically overwhelm the queen and be able to pick off the opponent's pawns one by one. It all, of course, depends on the position.

At d=28 <Critter 1.6a> evaluated the position as roughly even, [-0.03], after 51.Re4+ Kd6 52.Kg2 Kc7 53.Kf3 Qg5 54.Bd3 Qf6+ 55.Ke3 Qh6+ 56.Kf3 Qh7 57.Ke3 Qh3 58.Kf3 Qd7 59.Ke3 Qg7 60.Nf5 Qg5+ 61.Ke2 Qg2+ 62.Ke3 Kd8 63.Rh4 Qb2 64.Re4 Qc3 65.Ke2 Qb3 66.Rg4 Qe6+ 67.Kd2 Qa2+ 68.Kc3 Qa1+ 69.Kc2 Qa2+

click for larger view

Neither White nor Black has been able to make any progress with respect to moving their pawns and a draw by repetition is the likely result.

At d=26 <Houdini 1.5a> also evaluated the position as roughly even, [-0.06], after 51.Re4+ Kd5 52.Kg2 Qb2+ 53.Kf3 Qf6+ 54.Ke3 Kc6 55.Bd3 Qg5+ 56.Kf3 Kc7 57.Ne2 Qf6+ 58.Nf4 Kb7 59.Re6 Qh4 60.Re8 Qh1+ 61.Kf2 Qc1 62.Re7+ Kb8 63.Re4 Qb2+ 64.Ke3 Qf6 65.Nd5 Qg5+ 66.Nf4 Kb7 67.Bc4 Qd8 68.Ke2 Qd7 69.Bb5

click for larger view

Again, neither White nor Black has been able to make any progress with respect to moving their pawns and Black's pawns are tightly blockaded and there doesn't seem to be a way for Black to get his pawns moving. So a draw by agreement is the likely result.

Finally, at d=26, <Rybka 4.1>, which I consider the strongest endgame engine, evaluated the position as completely even, [0.00], after 51.Re4+ Kd5 52.Kg2 Qc2+ 53.Kf3 Qb3+ 54.Kf4 Qxg3+ 55.Kxg3 Kxe4 56.Kf2 Kd4 57.Bb5 c4 58.Ke2 Kc3 59.Ke1 Kb4 60.Kd2 Kb3 61.Kd1 Kc3 62.Kc1 Kb3 63.Kd1 Kc3 64.Kc1 Kb3 65.Kd1 Kc3 66.Kc1.

click for larger view

In contrast to Critter and Houdini, Rybka chose a Q for R+N exchange in an attempt to stir up winning chances with B vs. 2P and the more active king.

But this approach doesn't seem to lead to more than a draw either. If 59.Kb3 instead of Kb4, White still draws after 60.Kd2 Kb4 61.Kc2 a6 62.Bxa6 Kxa4 63.Bxc4 and the bishop can always give itself up to prevent the b-pawn from queening.

click for larger view

And if 60.c3+ instead of 60.Kb4, White sill draws after 61.Kc1 since Black can't force his pawns through; 61...a6 62.Bxa6 Kxa4 63.Kc2 Kb4 64.Bd3 b5 (zugswang) 65.Bxb5 and the c-pawn falls.

click for larger view

<ughaibu> speculated that if there were no a-pawns Tal would have played on from this position:

click for larger view

That's possible, since in that case it would seem that Black had all the winning chances. So this is what the same 3 engines came up with:

<Critter 1.6a>: [-0.69], d=26: 51.Rc4 Qa1 52.Kg2 Qb2+ 53.Kh3 Kd7 54.Ne4 Kd8 55.Kg4 Qg7+ 56.Kf3 Qf7+ 57.Kg2 Qg6+ 58.Kf2 Qf5+ 59.Kg1 Qf4 60.Ra4 Qf3 61.Ra8+ Kc7 62.Re8 Qg4+ 63.Bg2 Qd7 64.Ra8 Qd4+ 65.Kf1 c4 66.Ra7+ Kc6 67.Nc3+ Kc5 68.Ne2 Qb2

click for larger view

Black's initial king moves seem very passive but it was able to advance its king later on. Black also managed to advance it's c-pawn one square and seems to be able to advance its b-pawn as well; e.g. 69.Rc7+ Kb4 followed by 70...Kb4.

click for larger view

In fact, starting from this position Critter evaluates the resulting position at [-1.41], d=24 as Black is able to advance its pawns further and may have realistic winning chances.

<Houdini 1.5a>: [-0.52], d=26: 51.Rc4 Qa1 52.Kg2 Qb2+ 53.Kf3 b5 54.Re4+ Kd5 55.Nf5 Kc6 56.Re6+ Kd7 57.Re4 Kc7 58.Ne3 Qb3 59.Re5 Kb6 60.Rf5 Qc3 61.Rf8 Qe5 62.Rf5 Qd4 63.Nd5+ Ka5 64.Ne3 Qd6 65.Rd5 Qc6 66.Ke4 Qg6+ 67.Rf5

click for larger view

Black has managed to advance its b-pawn and transfer its king to the q-side but not much else. Still, Black may have some winning chances. Starting from this position Houdini evaluates the resulting position at [-0.77], d=24. Better, but no clear winning chances for Black.

<Rybka 4.1>: [-0.40], d=26: 51.Rc4 Qa1 52.Kg2 Qb2+ 53.Kf3 b5 54.Re4+ Kd7 55.Nf5 Qb3+ 56.Kf4 Kc6 57.Be2 Qa2 58.Ne7+ Kb6 59.Nc8+ Kc7 60.Na7 Qf7+ 61.Ke3 b4 62.Nb5+ Kd7 63.Bc4 Qg6 64.Re5 Qb1 65.Rd5+ Ke7 66.Kd2

click for larger view

Rybka either made a mess of this ending as Black or played inspired defense as White. It managed to advance Black's pawns somewhat but they are thoroughly blockaded and don't seem able to advance any further. In fact, starting from this position Rybka evaluates the resulting position at [-0.00], d=26 evaluating that Black has nothing better than a draw by repetition after 66...Qb2+ 67.Kd3 Qb1+ 68.Kd2 since otherwise Black loses both pawns.

So without the a-pawns Black's winning chances improve, but not all that much, so I wouldn't criticize Tal if he had also agreed to a draw if there were no a-pawns in the position.