Patterns for the Bxh7+ sacrifice?

One of the most common sacrificial methods to open up the castled Black King's defences is the bishop sacrifice on h7

But it's difficult to tell when this sacrifice will work. Anyone have a comprehensive check list, or pattern manual, to tell you if there is a reasonable (80%?) chance that it will work? 

Burgess suggests there are four main follow ups for white:

1. Ng5+ with the white Q ready to come to the h-file "if" Kg8.
2. White Q & R "quickly" coming to h file.
3. Q check forking K & other piece, winning material.
4. Further bishop sacrifice on g7, completely destroying the pawn cover.

He adds, if the aim is to force mate:

• It is "useful" to deny a black knight access to f6, or remove it.
• The presence of a black Rook on f8 and e7 "can help" White by blocking the King's flight squares.

But this is all a bit vague:

1. What if Kg8 doesn't happen?
2. What does he mean by "quickly"?
3. Forking and winning material seems good, firm, advice! But how often will you get that chance?
4. One sacrifice worries me, but two?!

Burgess illustrates Point (2) by the Greco Mate:

Some non-rhetorical questions on this that I'd love answers to:

• The White pawn on e5 eventually helps checkmate the King. Is it always an essential part of a Bh7+ sacrifice?
• Are White's black squared bishop, and Black's pawn on d5, essential to the proceedings?

watch some of Mikhail tals games. he does Bxh7 a lot so you can analyse to see if there are any patterns.

that is when I go for it, or when there is a fork. I am sure, I have missed some great opportunities, because didn't tried in positions with less, but I am afraid to try anything complicated in blitz games. In daily, I will go for it with much less (but will calculate first). The position you displayed, almost guarantees a check mate or huge advantage in material.

- The white pawn on e5, is of great importance, because it can stop an enemy Knight from defending the h7 square

- The black squared bishop is essential, because usually there is a queen on d8 and u need the bishop to protect ur knight in g5 (also it protects ur knight from the enemy King). Usually the best response for black is to sacrifice his queen for capturing ur knight (but u will be ahead in material).

I have won few games from such positions

If g5 is only defended once, eg black's Q bishop has been traded or is posted somewhere other than e7, and if Nf6 can be removed/relocated while your queen, knight and bishops work in tandom, it usually works well.

There is an book examining the consequences of the Bxh7+ sacrifice in its many manifestations titled Sacking The Citadel by Correspondence Senior International Master, Jon Edwards.  You can read John Donaldson's review here.  I own two copies so I can highlight and make notes in one.

Seeing some master games, involving the Bh7+ sacrifice, might be useful. Have you any to post? Here's one, Barva-Kis Hadjuboszormeny 1995, 1-0:

Thanks for the responses. Ging, could you post one of Tal's games?

#5 Irahran, thanks for that game, it looks to follow the standard Greco Mate pattern exactly, and illustrates the various options for the King, coming in or out. Burgess suggests 15. Qd3+ as a possible alternative to 15.Qg4!, but the latter is his first choice, as is the response 15... f5. He says if the King comes out it is very dependent on specifics of the position. He doesn't show any continuation, so it's nice to see your 16 to 26, though I don't know the theory well enough to say if it is "optimal" or not.

dk - ye sthe black squared bishop seems universal. Where else would it be? I wa sthinking maybe h4 could protect N on g5, but that may be just my fantasy...

bob - thanks for that reference, it looks very comprehensive! Maybe too much for 1400-1800 players?

Hi  here are some great articles on the topic by Silman - very instructive, understandable to us mere humans, and they helped me hugely on this topic. 

https://www.chess.com/article/view/the-classic-bishop-sacrifice-part-one

https://www.chess.com/article/view/the-classic-bishop-sacrifice-part-2

https://www.chess.com/article/view/the-classic-bishop-sacrifice-part-three

https://www.chess.com/article/view/the-classic-bishop-sacrifice-part-four

https://www.chess.com/article/view/oddities-in-the-classic-bishop-sacrifice

 Great question! +1

5:14 in this video. https://www.youtube.com/watch?v=w5oxUXJ25xE&list=PL55B2E0491289D108&index=34 4:00 in this video. https://www.youtube.com/watch?v=VQzIRqtAbB0&index=36&list=PL55B2E0491289D108 5:15 in this video https://www.youtube.com/watch?v=J2ZqPF3zmXk&list=PL55B2E0491289D108&index=42

Another good thread here:

https://www.chess.com/forum/view/game-analysis/when-to-offer-the-greek-gift

Spod has a nice summary in post 12 of that thread of Seriwan's thoughts on this matter, from "Winning Chess Combinations." 

Here's the first part of the game he mentions animated:

So we are left to perform the classic Greco Mate. A point to note here is that white doesn't even need to castle before going ahead....

There is no way around hard work - you will never play Bxh7+ based on some pattern alone, unless it's something identical to what you've seen before.

Like all tactical patterns, you need to actually calculate variations to see if it works in that specific position or not. 

Here's a rapid game I just played that almost attained the pattern:

Note the 10...f5 move that derailed my plan. How to get round that? I thought I'd experiment with a "remove the guard" type manoeuvre. But it didn't really pan out. I won the following blunderfest, but more by luck than judgement...

There are also several examples without a knight on g5.This is prone by: