I have played out some variations when White fights by letting the Q go in the melee for supremacy. These all end in P dominated winning end game for Black.

I concluded, at best Qf1 is a speculative move, in that Black may not be able to find 37...Bxe4 pinning the N from controlling b2. (Its not an easy move to find and White may not have foreseen the development).

Fritz says Black is better, but I had the old "ghost Rook" thing going on, I thought that c3 was covered ...

so nuts.

Black to play for a win.

I'm liking 31...Qxc3! 32. Rxc3 Rxc3 33. Qg2 Rxc2 34. Qxc2 Rxc2 35. Kxc2 Bxd5+ 36. Kc3 Be3! 37. Ng3

[37. Kb4 Bf3 38. Rb1 (38. Rd3?! loses time as 38...Bc5+ 39. Kxb5? Be2 wins outright) 38...Bxg4 39. Ng3 Bf4 and Black's center pawns are very strong with the support of the Bishops.]

37...Bf3 38. Rd3 Bd4+ 39. Kb4 Bxg4 40. Kxb5 Bf2 41. Kb4 Bxg3 42. Rxg3 Bxf5 and Black's center pawns must win over White's b-pawn as the B controls b7, at the very least, while the King can now move freely into the game.

<Everyone> I'm officially 7 for 7 this week, guys!

Two weeks ago: 5.5 out of 7

Last week: 6.25 out of 7 (as in last Sunday's Tarrasch puzzle a draw was proven as what would occur with best play after Nxf7 double check, which is what I went with: The alternative to the game continuation, Ne6+)

This week: 7 for 7! Official!

I looked at this for 40 minutes, thinking I was missing something in the line 31...Bxe4 32.Qxe4 d5 33.Qxd5 Qxd5 34.Rxd5 Rxc3 35.Rd7+ Kf8 (35...Ke8 36.Nxf6+ ) 36.Rd8+ . I also looked at 31...d5 to threaten a fork, but 32.exd5 threatens his own fork. The only forcing move that seems to lead somewhere is the insane looking 31...Qxc3, threatening immediate mate.

A) 31...Rxc3 32.Rxc3

A.1) 32...Qxc3 33.Rxc3 Nxf4 34.exf4 looks interesting although I'm not sure where this leads.

A.2) 32...Qe2 33.Rxc2 Qxc2 34.Rxc2 Kxc2 35.Bxe4+ Kc3 36.Be3 looks like another interesting line.

31... Qxc3!

The threat of 32... Qxc2 leaves white with no real choice.

32.Rxc3 Rxc3 33.Qe2 (Qxc3 Rxc3 34.Nxf4 exf4 35.Rf1 f3 followed by 36...Re3 looks hopeless for white) Rxc2 34.Qxc2 Rxc2 35.Kxc2 Bxe4+ 36.Kc3

No really serious alternatives until this point. Black has obtained 2 bishops + 2 pawns for rook+knight, but the smoke has not quite cleared, because the white king can get to the black queen-side pawns.

36... Be3!

It's critical to keep the two bishops on the board to support the new passers. 36... Bh2? keeps the knight trapped (FWIW) but puts the DSB on a low mobility square. Now the bishops control the board:

A) 37.Re1 Bd4+ 38.Kb4 (Kd2 Bc3 is hopeless) Bf3 39.Kxa3 Bxg4 40.Ng3 h5! (Threatening Bc2. The immediate 41... Bc2 42.Rc1 Bxg3 allows 43.Rg1 getting the piece back) 41.Rf1 Kd7! (getting king off the e-file before pushing e-pawn) 43.Kb4 Kc6 44.a4 ba 45.ba d4 and the advance of the e-pawn will win.

A.1) 39.Kxg5? Bxg4 40.Ng3 Bf2 41.Rf1 Bxg3 42.Rg1 Be2+ 43.K moves Bxh4 with a winning material advantage.

B) 37.Kb4 Bc3 38.Rf1 (Re1? Bd2+) Bxg4 39.Ng3 Bc5+! 40.Kc3 (Kxb5? Be2+ ) b4 and white has no counter-play.

30.qe2 instead g5 31.qf3 maybe b4? Show an tell in must absolute

conviction over qxc3 rxc3 rxc3 defend looks in c2 exchange down

bishops pair for probing against out action h5 rookd1 level

material castle rag in back seventh in draw slight plus forge

black it often in extract or swindle underneath wave looked d6 e5

alike bridge. This pineapple goes down to whites queen placement

it fallow in ground rich picking after f1? Shave ready made cull

opening old c3 go enterprising but I got stuck to right draw-back

queen in dilemma again try in amostly it over mini trate in woold

"In general you should stop your analysis if:

- The sacrifice isn’t calculable and your judgment says the sacrifice is worth it (or not),
- The reward becomes less than the risk, or
- There are no more checks, captures, and threats and you end up behind in material."

Black to play (31...?) "Insane", even material.

This one truly earns it's difficulty level. After briefly looking at 31...Qxc3 (which I prematurely dismissed for some reason), I spent too much time looking at moves like 31...d5 and 31...Qb4 (which draws after 32. cxb4 Rxc2 33. Nxf4 Rb2+ 34. Ka1 Rcc2 35. R3d2 Rxa2+ 36. Kb1 Rab2+). After nearly giving up, I returned again to the former try, and noticed what I missed before.

31...Qxc3!! 32. Rxc3 (forced) Rxc3 33. Qe2

Or 33. Qxc3 Rxc3 34. Nxf4 exf4 35. Re1 Rg3, with a winning endgame advantage. Also, 33. Qf2 Rxc2! transposes, but not 33...Bxe4 as I first thought (with the idea of 34. Bxe4?? Rc1+ 35. Rxc1 Rxc1#) because of 34. Qa7+! Kf8 35. Qxa3 Bxc2+ 36. Kb2 Bxd1 37. Nxf6+ Ke7 38. Nd5+ with a draw.