Material: Up a P. The White K has one legal move, to d7. Both Black and White have diagonal batteries: Black, a Qc6-Bd5 battery trained on Pg2 and Rh1; White, a Qe3-Bb6 battery. The position has some symmetry, but White has castled Q-side, with Rd1 attacking Bd5. The Qe3 is on the same file as Ke8, separated only by the Black Pe6; Qc6 is on the same file as Kc1, separated only by the White Pc2. The Rd1 effectively pins Bd5, because of the threat 1.d7+ 2.d8=Q. The asymmetries in the position favoring White, beside the Ks, are Pf4 and the passer Pd6, restricting Ke8. White pieces are well posted with the exception of Bf1 and Rh1, which need to be activated. The alignment on the e-file suggests a central break on the d- or f-files, to exploit White’s lead in development. The position is tense, because Black can counterattack with 17.Bxd6 or 17.Rc8. Development with tempo is critical.

Candidates (17.): d7+, Rxd5, Bxb5, f5, c5

17.Rxd5 Qxd5

Candidates (18.): Be2, c5

18.Be2 (threatening 19.Rd1 and 19.Bf3, skewering Ra1 20.Bc6+)

The response 18…Rc8 transposes unfavorably at best into a line where Qd5 moves. The Black Qd5 has 3 options: (1) counterattack, (2) capture, or (3) retreat along the rank.

(1) 18…Qxa2

Black’s only developed piece is now out of the action.

19.d7+ (threatening 20.Rd1 21.d8=Q)

19…Kxd7 20.Rd1+

threatening 21.Bf3 22.Qd4, then 23.Bd8+, 23.Bc6+ or 23.Qd7+ depending on Black’s response.

(2) 18…Qxd6 19.Rd1 Qe2

Other feasible options essentially transpose into other lines, or worse.

20.Bf3 (threatening 21.Bc6+ and 21.Bxa1) Rc8

21.Qd4 (threatening 22.Bd8 Rxd8 23.Bc6+ Rd7 24.Bxd7+)

(3) Retreat along the rank just misplaces Qd5 for defense in the lines above.

1.Rxd5
1.d7+

1.Rxd5 Qxd5 2.Bxb5+ Qxb5 3.Rd1 Bxd6 4.Rxd6 o-o only leaves white down the exchange while black has open lines of attack on the white king. Hold on, that Bxb5 is interesting. Lets add to our candidate moves.

1.Bxb5
1.a4

1.Bxb5 axb5 2.Rxd5 Qxd5 3.Rd1 and if queen moves and does not capture at d1 white will have 4.d7+ followed by 5.d8(Q)+ which will lead to mate with a bishop recapture at d8. So 3...Qxd6 4.Rxd6 Bxd6 5.Qd4 and it's Q+B vs R-R-B, and I'm not so sure about this.

1.a4 looks to be to slow and unnecessary, while 1.d7+ throws away the pawn without improving the position.

After my discussion on computer analysis with <MostlyAverageJoe> in the kibitzing to E Mortensen vs J Rodgaard, 1980, I should give the reasoning behind my computer methods. (1) To differentiate the quality of 2 moves, I fix the analysis at 15 plies to avoid the unfair accrual of advantages if one line is given a deeper analysis than the other. The depth is a compromise: it is generally (but not always) enough to demonstrate a stable advantage, while permitting other members of my family access to our home computer. (2) To avoid destroying information, I give the full computer line, which is obviously inferior near its end. With those prefatory remarks, here is the analysis.

The position after 18...Qxa2.

click for larger view

Toga II analysis of my move 19.d7+:

[ply 15/73, time 07:46, value +5.72]

19.d7+ Kxd7 20.Rd1+ Ke8 21.Qe4 Qa1+ 22.Kd2 Rd8+ 23.Bxd8 Qxb2 24.Ba5 Bd6 25.Bc3 Qa3 26.Bf3 Qa4 27.Qa8+ Ke7 28.Qxh8 Qxf4+ 29.Ke2 Qc4+ 30.Rd3 b4 31.Bxg7 Qxc2+ 32.Rd2

Toga II analysis of Nunn's move 19.Qf3:

[ply 15/59, time 01:09, value +3.99]

19.Qf3 Qd5 20.Rd1 Qxf3 21.d7+ Ke7 22.d8Q+ Rxd8 23.Bxd8+ Ke8 24.Bxf3 Be7 25.Bc6+ Kf8 26.Rd7 Bc5 27.Kd2 f6 28.c3 b4 29.Bc7 bxc3+ 30.Kxc3 Be7 31.Bd8 Bc5 32.Kc4 Be3

I disagree. At least 18...Qxa2 counterattacks, where other alternatives quietly put Black in a box to await his fate.

<<Jim Bartle> wrote: Nunn gives long variations for other 18th moves for black, but doesn't particularly criticize Qxa2.>

Psychologically, 18...Qxa2 was the toughest move to anticipate, because without 18.Be2, Rh1 appears at risk.

19. d7+ Kxd7 20.
Rd1+ Bd6 21. Rxd6+ Kxd6 22. Qc5+ Kd7 23. Qc7+ Ke8 24. Qc6+ Ke7 25. Qb7+ Kf6 26. Bd4+ e5 27. fxe5+ Kf5 28. Bg4+ Kxg4 29. Qe4+ Kg5 30. Be3+ Kh5 31. g4+ Kh4 32. Bf2+ Kh3 33. Qf3+ Kxh2 34. Qg3+ Kh1 35. Qg1#

Your solution is a Schrodinger's cat, half dead and half alive.

Toga II 1.3.1 evaluates 17.d7+ as a draw.

[ply 15/48, time 05:41, value +0.09]

17.d7+ Qxd7 18.Bd3 Bd6 19.f5 Qc6 20.Bd4 0-0-0 21.fxe6 fxe6 22.Rhf1 Rd7 23.g3 Bxa2 24.Be4 Bd5 25.Bxd5 exd5 26.Kb1 Kb7

<<>My only problem is that this is pretty straight forward - where is the "insane" aspect?>

In fact I found 18.e2 (and was pleased to discover that it was the correct move) which sets up the zugswang ... <Nova485:> certainly qualifying as a challenge, albeit the fireworks are in the notes.

I must respectfully disagree with your disagreement and with your assessment of Qxa2. The move is easy to anticipate -- after all, the a2 pawn is always a weak spot after castling long. Furthermore, losing Rh1 is a very small concern for white - it takes black Q far too out of play. And Rh8 is facing the same kind of risk anyway.

Look again at the position after 17.Rxd5 Qxd5 Be2 (which is about as far as I was 100% sure the solution would go):

click for larger view

It is quite clear that the d-pawn is a major threat here once it gets supported by Rd1. At this point I fully expected black to neutrailize this threat - the only question was whether the correct move would be Bxd6 or Qxd6.

Tangential note: in my attempt to solve the puzzle, I looked at 18...Qa2 19.Rd1 as a possible continuation. Even though it is highly sub-optimal compared to 19.Qf3, it also wins, and gives white more advantage than he would have afer 18...Bxd6.

Addendum regarding the risk to Rh1: Hiarcs tells me that 23...Qxh1 is a big blunder, allowing forced mate. A better response would be:

23... Qa5 24. b4 Qd8 25. Bd3 (followed by 25...Bxc5 26. bxc5 or perhaps 26...g6 or Kg8 to get the K some breathing room).

Unlike Nunn :-) I was aware that 19.Rd1 would've been the correct continuation, since <crwynn> posted it a couple minutes before my response (after noticing it I actually took out a part of my post to avoid duplication). And Rd1 is quite easy to spot without computer help, following the common rule about rooks belonging on open files.

I will avoid the obvious riposte :)

To explain more carefully, I was giving a statement about my psychological difficulties during analysis. Apparently two GMs did not think my fears about 18...Qxa2 entirely unreasonable. That said, I found everything you wrote about the objective merits of the various moves quite sound, as usual.

Fritz differs from Black on move 18. ...Qd5xd6

Time 9min Depth 47

John Nunn - John Fedorowicz, It (ca

Analysis by Fritz 11:

1. (1.74): 17.Rd1xd5 Qc6xd5 18.Bf1-e2 Qd5xd6 19.Rh1-d1 Qd6-b8 20.Be2-f3 Bf8-e7 21.Bf3-c6+ Ke8-f8 22.Rd1-d7 Be7-h4 23.Bb6-d4 Kf8-g8 24.Bd4-e5 Qb8-c8 25.Rd7-c7 Qc8-b8

2. (0.39): 17.d6-d7+ Qc6xd7 18.Qe3-d4 Qd7-e7 19.Bf1-d3 Qe7-f6 20.Qd4xf6 g7xf6 21.Rh1-e1 Ra8-b8 22.Bb6-d4 Bf8-d6 23.g2-g3 Ke8-e7 24.b2-b3 Bd5-f3 25.Bd3-e2

(, 31.03.2008)