Apr0305   Autoreparaturwerkbau: Demolition on h6 AND g7 ;) 

Jun3020   stacase: That took about two seconds. 

Jun3020   CaliWest: Much easier than yesterday 

Jun3020   siggemannen: What about 21...g6 though 

Jun3020   newzild: 21. Bxh6 wins a pawn, as 21..Qxe2 runs into 22. Bxg7+ Kg8, when both 23. Qh6 and 23. Bxf6+ mate quickly. 

Jun3020   Brenin: Much easier than yesterday. If 21 ... g6 then 22 Rge1 or Bd3, and White is two Ps up with a much better position. 21 ... Ng4, blocking the gfile and putting both White Bs en prise, looks like the best defence for Black. 

Jun3020
  agb2002: White is a pawn ahead.
The weak black castle suggest Rxg7 and Bxh6.
In the case of 21.Rxg7 Kxg7 22.Bxh6+ Kh8 23.Qg5 Rg8 25.Qh4 Nh7 26.Qh5 Qxe2 wins for Black. In the case of 21.Bxh6:
A) 21... gxh6 22.Qxh6+ Nh7 23.Qg7#.
B) 21... Rg8 22.Bxg7+ Rxg7 (22... Kh7 23.Qh6#) 23.Rxg7 Kxg7 24.Rg1+ B.1) 24... Kf8 25.Qh6+ Ke8 26.Qh8+ Qf8 27.Qxf6 wins decisive material (27... Bxh2 28.Rh1 Bd6 29.Rh8). B.2) 24... Kh7 25.Bd3+ Ne4 (25... Kh8 26.Qh6+ Nh7 27.Qg7#) 26.fxe4 wins decisive material (26... Be5 27.f4). B.3) 24... Kh8 25.Qh6+ Nh7 26.Qg7#.
B.4) 24... B(N)g4 25.fxg4 wins two pawns at least. C) 21... Qxe2 22.Bxg7+ Kg8 (22... Kh7 23.Qh6+ Kg8 24.Qh8#) 23.Bxf6+ Kh7 (23... Kf8 24.Qh6#) 24.Rg7+ Kh8 25.Qh6#. D) 21... g6 22.Rde1 wins a second pawn. 

Jun3020   Brenin: As in yesterday's POTD <Kiuttu vs Giri, 2005>, the defender could have disrupted a deceptively easylooking attack by throwing a Knight onto an important line of attack (16 Ne4, onto the a8h1 diagonal yesterday, 21 ... Ng4, onto the gfile today). In both cases, disaster could have been averted, with relatively minor material cost. Mondays and Tuesdays are not as straightforward as they used to be. 

Jun3020
  chrisowen: Elkpacks rekeysja keyihuff bh6jaffa honejaff qe2jaffa eonejaff junks it i 21g6
kestrels nagjaffa coojamma gonejaff keyihuff 2 3 it ok keyojaff junks it go
22.Bxg7+ kestrels honchoja maffkeys gonejaff rebudums rematers rekeyers keyijaff 4 it ok keyojaff eonejaff elkpacks rekeysja vitallyz its peektons keysajam keysjaff withgack laddyons its pecktons keysjaff mugwithy dullyons its ikeyhuff bh6jaffa honejaff coojamma 20Kh8 honejaff vitojaff keyojaff 2 3 it ok keyojaff coojamma 20Nh5 honejaff bondhoof 0 1 it ok keyojaff keysijam umsumums gikeyjam flickjam etcetera jdumjaff keypucky elkpacks rekeysja bh6jaffa honejaff? 

Jun3020
  Predrag3141: All I could figure out was that 21 Rxg7 doesn't seem to work, and 21 Bxh6 gives White a pawn and an obvious positional edge (Black's exposed king). 6second Stockfish analyses confirm that the edge for White slightly exceeds a pawn and that there is no forcing line to get more. For example, 21 Bxh6 Rg8 22. Be3 Nh7 23 h4 Rbe8
If 23 h4 is the best you got, even in just 6 seconds, it's a positional game. 

Jun3020   TheaN: The main question here is 21.Rxg7? or <21.Bxh6>. It's rather obvious the former doesn't work well because White's queen isn't going first. After 21....Kxg7 White can choose between two evils but they come down to the same, most consistent would be 22.Bxh6+ Kh8! + and Black will eventually reload over the gfile. After the text, Black's in trouble. Obviously the game line 21....Qxe2? loses instantly to 22.Bxg7+ Kh8 (Kh7 23.Qh6+ Kg8 24.Qh8#) 23.Bxf6+ Kh7 (Kf8 24.Qh6#; any B 24.Qg5+ Kh7 25.Qg7#) 24.Rg7+ Kh8 25.Qh6#. However, Black doesn't have options a plenty. Defending, 21....Nh5, fails on 22.Bxg7+ with 23.Qh6 and mate soon. 21....Rg8 is a bit more sturdy (probably should have looked at it a bit better); as Black has two defenders on g4 White should cut his losses and retreat 22.Be3 or perhaps more consistently reload 22.Bd3, both about +3. If White insists on pushing through as per <agb>'s line 22.Bxg7+ Rxg7 23.Rxg7 Kxg7 24.Rg1+ Ng4 25.fxg4, 25....Qf6! ± +1.5 might complicate matters as Black threatens Bf4, Qxf2 and Bxh2, so it's not just two pawns and White lost quite some initiative. After <21....g6> White has a few ways to continue but it's not done just yet. Though I concur 22.Rge1! +, shifting the play, is White's best, I enjoyed the line 22.Bg5. I correctly pictured 22....Qxe2 23.Bxf6+ Kh7, but now played the flawed 24.Qd4?! Qxf3 25.Qh4 Qh5 26.Qxh5+?! (Rd4! ±) gxh5 27.Rf7+ Kh6 ⩲:
click for larger view
As there's too much in for White also (Nb4, h2). Unless White plays 28.Nc6 ⩲, Black can simplify to opposite colored bishops a pawn down. After 28.Rxf7? Bg4! ∓ Black's even better. In this line, White should swap 24.Qxe2 Rxe2 25.Rd4! which does the same but without Qxf3; Black's forced to 25....Re5 26.Bxe5 +. The last paragraph is not incredibly relevant though, after 21.Bxh6 g6 + White pretty much got what he wanted. 

Jun3020   mel gibson: That was easy however although Stockfish 11
started out with the same answer it changed its first move to:21. Rxg7
(21. Rxg7 (♖g1xg7 ♔h8xg7 ♘b4c6 ♘f6e4 ♖d1g1+ ♔g7h7
♗e2d3 ♖e8g8 ♖g1xg8 ♖b8xg8 ♘c6xe7 ♖g8g1+ ♕d2d1 ♖g1xd1+ ♔c1xd1 ♗d6xe7
f3xe4 b5b4 e4e5+ ♔h7g7 a3xb4 ♗e7xb4 ♔d1e2 a4a3 b2xa3 ♗b4xa3 ♔e2f3
c7c6 c2c4 c6xd5 c4xd5 ♗a3b2 ♔f3f4 ♗b2c3 ♗e3c5 ♗c3d2+ ♔f4f3 h6h5
♗c5e7 ♗d7g4+ ♔f3e4 ♔g7g6 ♔e4d4+ ♗g4f5 d5d6 ♗d2a5 ♗d3xf5+ ♔g6xf5
♔d4d5 ♔f5f4 d6d7) +5.15/37 47)
score for White +5.15 depth 37
In the actual game 
following the bad reply by Black it's mate in 4. 

Jun3020
  chrisowen: I rebuke h8 maybe it is h5 in no? 

Jun3020   goodevans: Today's puzzle has generated a lot of interest even if it is flawed. <mel gibson>'s post begs the question what the intended solution was actually supposed to be. I guess finding <21.Rxg7 Kxg7 22.Nc6 Ne4 23.Rg1+ Kh7 24.Bd3 Rg8 25.Rxg8 Rxg8 26.Nxe7 Rg1+ 27.Qd1 Rxd1+ 28.Kxd1 Bxe7 29.fxe4> would be a big ask for a Tuesday, especially as we would also need to consider <21...Bxb4 22.axb4 Kxg7>. So it looks as though the simpler <21.Bxh6> was the intended solution after all, even though it yields only a moderate advantage against best defence. I'll throw a spanner in the works by suggesting that <21.Bd3> is about as good as that move since it retains both white's extra pawn and his positional plus. I wonder if <21.? white to play> will ever reappear as a Friday puzzle with <21.Rxg7!> as the correct solution. 

Jun3020
  Predrag3141: <mel gibson: Stockfish … changed its first move to: 21. Rxg7 … +5.15> Below is a one hour analysis after suggesting 21 Rxg7. Like your analysis, it's 37 ply but starts a ply later and the score drops by 2.18. 1) +2.97 (37 ply) 21...Bxb4 22.axb4 Kxg7 23.Rg1+ Ng4 24.fxg4 Kg8 25.g5 h5 26.g6 f6 27.Bc5 Qg7 28.Bd3 Ra8 29.Bd4 a3 30.bxa3 Rxa3 31.Qf4 Ra6 32.Bc3 Rd6 33.Be4 Bg4 34.Kb2 Be2 35.Re1 Bc4 36.Qf5 h4 37.Rg1 Ra8 38.Ra1 Rf8 39.Qh5 Qh8 40.Qg4 Qh6 41.f4 Qh8 42.Ra7 Qg7 43.Qxh4 Bxd5 44.f5 Bxe4 45.Qxe4 Qd7 46.Qh4 Qg7 2) +3.59 (36 ply) 21...Kxg7 22.Nc6 Ne4 23.Rg1+ Kh7 24.Bd3 Rg8 25.Rxg8 Rxg8 26.Nxe7 Rg1+ 27.Qd1 Rxd1+ 28.Kxd1 Bxe7 29.fxe4 Kg7 30.e5 Be8 31.f4 f6 32.Kd2 b4 33.axb4 Bxb4+ 34.Ke2 fxe5 35.fxe5 Bh5+ 36.Kf2 Be8 37.Kf3 a3 38.bxa3 Bxa3 39.Ke4 Bg6+ 40.Kd4 Bb2+ 41.c3 Bxd3 42.Kxd3 Kf7 43.Bxh6 

Jun3020   Hercdon: So should we just say that 21. Rg7 and Bxh6 are both good solutions and leave it at that? Rg7 is our Easter egg after Bxh6 was actually played 

Jun3020
  NM JRousselle: Isn't 6... h6 the correct move? 

Jun3020
  gawain: Like yesterday we have a combination that gives one side a rather modest edge. 21 Bxh6 g6 and White settles for a two pawn advantage and a slog toward eventual victory. Is 21 Rxg7 really any better? 

Jun3020   saturn2: How about 21 Nc6 forcing black to give up the bishop pair? Black can win back a pawn. ...Bxc6 22. dxc6 Bxh2 but opens white the h file for further attack. 

Jun3020   Brenin: When I suggested 21 ... Ng4 as a response to 21 Bxh6, I had no time for analysis, so I hoped that one of the usual hard workers on CG would investigate. No such luck, so here goes with some probably suboptimal lines: A) 22 fxg4 Qxe2 23 Qxe2 Rxe2 24 Be3 (trapping the R? No, here comes the cavalry) f5 25 h3 f4 and White is 2P up. B) 22 Bxg7+ Kxg7 23 Rde1 Qf6 (threatening Bf4) 24 Kb1 Bf4 25 Qd1 Bxh2 26 Rg2 Bd6 27 fxg4 Bxb4 28 axb4 Re4 and White is 2P up. C) 22 Nc6 Qxe2 23 Qxe2 Rxe2 24 Nxb8 Nxh6 25 Nxd7 Rxf2 and after losing a P White will be the exchange up. Conclusion: 21 ... Ng4 is no worse than some of the other responses suggested, and is a lot better than the move Qxe2 played. 

Jun3020   RandomVisitor: Adding to the collection of computer analysis posts: click for larger viewStockfish_20062920_x64_modern:
<59/110 5:49:32 +6.03 21.Rxg7 Bxb4 22.axb4 Kxg7 23.Rg1+ Ng4> 24.fxg4 Kg8 25.g5 h5 26.g6 f6 27.Bc5 Qg7 28.Bd3 Ra8 29.Kb1 a3 30.b3 f5 31.c3 a2+ 32.Ka1 Ra6 33.d6 Rxd6 34.Bxd6 cxd6 35.Kxa2 Qf6 36.Rg5 h4 37.Rh5 Re7 38.Kb2 Kf8 39.Qh6+ Kg8 40.Qf4 d5 <59/93 5:49:32 +3.30 21.Bxh6 Ng4 22.Nc6 Qxe2> 23.Qxe2 Rxe2 24.Nxb8 Nxh6 25.Nxd7 Rxf2 26.Rdf1 Rxh2 27.Rh1 Rg2 28.Re1 Kh7 29.Ne5 f6 30.Nf7 Bh2 31.Nxh6 Bf4+ 32.Kb1 gxh6 33.Re4 Bd6 34.Reh4 Rd2 35.c4 bxc4 36.Rxh6+ Kg7 37.R6h4 f5 38.Rxc4 Rf2 39.Rxa4 Rxf3 40.Rc4 Rd3 58/87 5:49:32 +3.04 21.Bd3 Qf8 22.Kb1 Rbc8 23.h4 Bc5 24.Rg2 Bxe3 25.fxe3 Qd6 26.Rdg1 Rg8 27.c4 Rce8 28.Ka1 bxc4 29.Bxc4 Qb6 30.Re1 Nh5 31.Qc3 Qd6 32.Bd3 Re7 33.Rc2 Be8 34.Rcc1 f6 35.f4 Qb6 36.Qd2 Qd6 37.Rg1 g6 38.f5 g5 39.hxg5 Rxg5 40.Rh1 Kg7 
