I considered it, but didn't see the proper follow-up, so this is where my streak ends.

I can take some consolation in that Stockfish confirms that 20. Re3 is a decent move. At a depth of 21, as I write this, it is saying that White gets a < > advantage.

This said, 20. Re3 is still inferior to the one and only outright winning move - GM Kupreichik's 20th.

So, my streak ends at 27 straight-correct solves.

pickle does it allow in knightf5 at iron in e1 castle exploit ok in

be7 then o i reckon it erstwhile in qh8 victory in lights out black

again method too stodgy this Kupreichik gob engage ala all his

Beside the text 20 Qh8 Nxb3 21 Be7!, one must analyze the responses 20…d5 and 20…e5. These two moves do not make 21 Be7 feasible because black now protects the key d6 square from 22 Nxd6+.

Nevertheless, white is still winning. To answer 20…e5, white has 21 Bg6!, threatening 21 Ng7+. Black has to concede the exchange to avoid that possibility.

click for larger view

After 20…d5, 21 Bxd5 looks the most straightforward. I’m not sure what black should play here but I believe white can retreat the bishop and reinstitute the Be7 threat since the d pawn is now out of the way.

For example if 21…Na4, then 22 Bb3 and the original threat reappears.

click for larger view

The black knight is pinned which makes the e6 square weaker. However, any sacrifice on e6 seems to go nowhere.

Another weak spot is d6, which suggests 20.Bg3, but after 20... e5 21.Bxe5 dxe5 22.Qxe5+ Re6 (22... Ne6 23.Ba4) 23.Bxe6 fxe6 White doesn't seem to have made much progress.

A better option appears to be 20.Qh8 with a number of threats (Be7, Bf6-Bg7, Qxh7-Bxe6, etc.). For example, 20... a5 21.Be7 Rxe7 22.Nxd6+ winning the queen (after capturing the white knight or after a discovered check).

<So, my streak ends at 27 straight-correct solves.>

Wow, that's a heck of a lot more than my record!

I have a magic number. I hope to make that one someday. :)

<20. Qh8 d5 21. Rxd5 Rxd5 22. Bxd5 Qb4> Or try this Crafty link for clarity.>

I hope you do not mind the hard work I put on your plate. Its much better than glossing over a hollow Kupreichik win.>

22. Rxe6+ Kf7 seems ok. Yeah, it looks intimidating for black, but I dont see any major threats.

By the time I scrolled down to the kibitzing text area, I already knew my "solution" was wrong. My only consolation: I am apparently in good company.

For the record, here's what I thought of before I took the bucket of icewater in the face. My candidate moves included:

(1) 20. a4, to pin the rook. Unfortunately, after 20. ...xa4; 21. xa4, b5, the pin proves of short duration, and White has no useful continuation.

(2) 20. h8, to pin the bishop. I rejected this move because it unpins the knight and decentralizes the queen, which seems an undesirable tradeoff; I should, however, have examined it more closely, because the follow-up 21. e7 fits the week's interference theme. *Kicks self smartly in rump.*

(3) 20. xe6, to begin to crack the wall around the king. But there seems to be no real threat after the simple 20. ...fxe6, so the idea was soon abandoned.

(4) 20. g3. Nothing exciting or brilliant about this move; it's a case of bailing out for steady pressure in hopes of exacting helpful concessions. In any case, it's what I settled for. My proposed line went more or less as follows:

20. g3, e5;
21. xe5, dxe5;
22. xe5†.

Black has four replies:

(4.1) 22. ...e7??;
23. h8† with mate to follow.

(4.2) 22. ...d8?;
23. e8†, c7;
24. xf8 . If White can avoid losing to any unexpected shots, he's two pawns up and retains some pressure.

(4.3) 22. ...e6;
23. xe6, xd1†;
24. xd1, xe6 ∞. White has a rook, pawn and pressure for two bishops.

(4.4) 22. ...e6;
23. a4, d8;
24. xd7, xd7;
25. d5, c5;
26. xf7, f6!;
27. e8†!, c7;
28. e5† . White has rook and two pawns for two bishops, and thanks to the scattered black forces, retains a slight edge.