Modern engines sometimes find incredible resources and rarely, if ever, make short-term tactical miscalculations in middlegames (their opening and endgame skills are not at such a high level, however).
Occasionally I come across an old game where instinctively I disagree with, or at least find surprising, the positional conclusions of the annotator(s).
Usually I am the one who is mistaken, which is why I will be particularly interested in what Stockfish14.1 and Komodo12.1.1 have to say about the following from 500 Master Games Of Chess.
The algebraicised notes in italics are by Savielly Tartakower and Julius du Mont.
Jacques Mieses - Mikhail Chigorin
Deutscher Schachbund (Hannover) 1902
Danish Gambit
The exciting vicissitudes of this game can be divided into three phases, from the point of view of Black: stopping the enemy's first onslaught; building up a close but defendable position; watching all openings, and, at the right moment, launching the counterattack.
1.e4 e5 2.d4 exd4 3.c3 dxc3 4.Bc4 cxb2 5.Bxb2 Qe7?!
This defence is not without logical foundation. The queen surveys the critical sector, prepares for casting on the queenside, and in addition threatens to win one of the bishops by 6...Qb4+.
The engines are distinctly unimpressed, preferring 5...Nf6, which was an earlier favourite of Chigorin's and has been played by Magnus Carlsen, and 5...Bb4+, which Tartakower and du Mont describe in analysis to another game as "not as promising as it looks." It is possible 5...Qe7?! was a novelty - there is no earlier example in ChessBase's 2022 Mega database. The move immediately struck me as suspicious as it slows Black's kingside development, but perhaps this is not important if the black king finds safety on the queenside.
6.Nc3 c6!?
This defends the b5 and d5 squares from a knight incursion and prepares the tempo-gaining ...d5.
7.Qc2
Two years later, against the same opponent at Cambridge Springs, Pennsylvania, Mieses played 7.Nge2!? The game continued 7...b5!? (7...Qb4 can be countered by 8.Bb3) 8.Bb3 a5 9.Rc1!? Na6 10.0-0 Nc5?! 11.Nd4 Nxb3 12.Nf5!? Qe6 13.axb3, after which the engines reckon White has much more than enough compensation for two paws (but 0-1, 54 moves).
7...d6
Suicidal would be 7...d5 8.Bxd5 cxd5 9.Nxd5 and wins.
The engines prefer 7...d5!? to the text, as long as Black does not grab the bishop, eg 8.Bxd5 Nf6 9.Bb3 Na6, when Stockfish14.1 gives White a slight edge although Komodo12.1.1 calls the position equal.
8.0-0-0?!
The engines reckon this is a mistake, preferring 8.Nf3 Be6 9.Bxe6 fxe6 10.Rd1!?, when they rate White as having full compensation for being two pawns down.
8...Be6
 |
| How should White proceed? |
*****
*****
*****
*****
9.Be2!?
At the crossroads. Instead of giving up territory, the imaginative 9.Nd5 maintains the initiative.
The authors do not explain how White should continue after 9...cxd5, eg 10.exd5 Qg5+ and ...Bf5, or 10.Bb5+ Nd7 11.exd5 Qg5+ etc. The engines reckon objectively best is 9.Bxe6, but it is understandable Mieses thought such an exchange would help Black's defence.
9...Nd7 10.Nf3 Nc5
Not 10...0-0-0 as yet, because 11.Qa4 and White gets going.
The engines reckon castling is fine, and if 11.Qa4 then 11...Kb8.
11.Rd4 0-0-0 12.Rhd1 Qc7
Slowly but surely Black proceeds with his development. 12...Nf6 would be premature, on account of 13.e5.
The engines prefer 12...Nf6 to the text, meeting 13.e5 with 13...Ne8 or 13...Nd5, in each case claiming a winning advantage for Black.
13.Na4 Nd7!? 14.R1d3?!
This natural-looking move makes White's position worse, according to the engines, although their suggestion of 14.Bc4!? Bxc4 15.Rxc4 Ngf6 16.Kb1 hardly inspires. It seems White has nowhere near enough compensation for being two pawns down.
14...Ngf6 15.Rc3 Be7
The "hedgehog" position, which Black has obtained, is cramped but solid, his pawns being particularly strong in the defence of the king's position.
16.Rb4
With the threat of 17.Ba6, which is, however, prevented by Black's next move.
The point being 17.Ba6 bxa6 is met by 18.Rxc6.
16...Nb8
Also meeting the threat are 16...Nc5 and 16...c5.
17.Nd4 d5
This counterthreat in the centre provokes the crisis.
18.Nxc6!?
Being already two pawns to the bad, and threatened with the loss of a third, White, by the offer of the knight, tries to alter the normal course of things.
18...Nxc6 19.Ba6
 |
| How should Black proceed? |
*****
*****
*****
*****
19...Qf4+
As neither 19...Kb8 20.Rxc6, nor 19...Kd7 20.Rxb7, nor 19...bxa6 20.Rxc6, nor 19...Nxb4 20.Rxc7+, nor, finally, 19...Bd7 20.Bxb7+, is admissible, the "relieving check" in the text is the only move, which, however, saves the situation and refutes the hostile combination.
The engines show 19...Kb8? 20.Rxc6? loses to 20...Qf4+ 21.Kb1 Bxb4. However they reckon 20.Rxb7+ Qxb7 21.Rxc6 leads to a draw by repetition after 21...Qb4 22.Bc3 Qa3+ 23.Bb2 Qb4 etc. They also reckon Black is better after 19...Bd7? 20.Bxb7+? Qxb7 21.Rxb7 Kxb7, but that White wins with 20.Rxb7, eg 20...Qxb7 21.Rxc6+.
The authors do not mention 19...Bxb4?, which the engines reckon leads to equality after 20.Rxc6 Ne8!? 21.Bxg7!?, eg 21...Ba3+ 22.Nb2 Rg8 23.Rxc7+ Nxc7 24.Be5 Bd6 25.Qc6 Bxe5 26.Qxb7+ Kd7 27.Nd3! Bd6 28.Bb5+.
20.Re3
Or 20.Kb1 Bxb4 21.Rxc6+ Kb8, and the whole of White's game collapses.
20...Bxb4 21.Qxc6+ Qc7 22.Bxb7+ Kb8 23.Be5
His final trump.
23...Qxe5 24.Rb3 Qc7 0-1
