Showing posts with label Kosten. Show all posts
Showing posts with label Kosten. Show all posts

Wednesday, November 22, 2017

Jerome Gambit: Give Me A Break

I have maintained The Database of Jerome Gambit games to help me get a sense of how certain lines or moves succeed or fail - from a practical (i.e. over-the-board play) point of view.

I have used chess analysis engines (currently Stockfish 8) to get a sense of how certain lines or moves are "objectively" strong or weak.

It is a joy when both the practical and the analytical agree. It can be confusing when they don't. Take the following game. It left me a bit confused...

Petasluk - givemeabreak
5 0 blitz, FICS, 2017

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 



4...Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Kf6 



Including this game, Petasluk has reached this position 7 times according to The Database (which, I admit, need some updating). Before this game he was 2-1 with the followup 7.Qf5+ and 2-1 with 7.f4. This compares with statistics from the whole Database - White scores 80% (35 games) with 7.Qf5+ and 80% (10 games) with 7.f4.

So, from a practical point of view, 7.Qf5+ and 7.f4 are both good move choices, according to Petasluk's experience; and both are excellent choices from the experience of those represented by The Database.

7.f4 

So, no surprise here.

Yet the computer analysts, starting with Stockfish 8, leap directly for the straightforward 7.Qf5+ Ke7 8.Qxe5+ Kf8 9.Qxc5+ picking up the two sacrificed pieces.

Why the discrepancy?

7...g6

Well, here is one reason: the defender in this game is human, and humans make mistakes. The alternative 7...Nc6, withdrawing the attacked piece, allows White to only recover one of his sacrificed pieces:  8.Qf5+ Ke7 9.Qxc5+ Kf7, with pressure on the enemy King, but not full compensation for his offerings.

8.Qxe5+ Kf7 9.Qxh8 Black resigned 



Well, it looks like "Practical Experience 1, Computer Analysis 0".

But that is not the whole story. A number of Petasluk's games continued from the above "final" position:

9...Nf6 (9...d6 10.Qxh7+ Kf8 11.d3 Qf6 12.Qxc7 Be3 13.Nc3 Qxf4 14.Rf1 Qxf1+ 15.Kxf1 Bxc1 16.Rxc1 Black resigned, Petasluk - douthy, FICS, 2015; or 9...Qe7 10.Qxh7+ Kf6 11.Qxg8 Qxe4+ 12. Kd1 d5 13.Qh8+ Kf5 14.Qh3+ Kxf4 15.Rf1+ Kg5 16.Qg3+ Bg4+ 17. Rf3 Re8 18.c3 d4 19.h3 Qe2+ 20.Kc2 d3+ 21.Rxd3 Qd1 checkmate, Petasluk - kosten, FICS, 2013) 10.Qxd8 Black resigned Petasluk - skaks, FICS, 2015 and Petasluk - kaliz, blitz, FICS, 2014 Be7 11. Qh8 11... b6 12. e5 Bb7 13.Qxa8 Bxa8 14. exf6 Bxf6 15. O-O Be4 16. d3 Bc6 17. Nc3 Black resigned, Petasluk - TaccyChan, FICS, 2007.

That's 4 wins and 1 draw. More support for practical considerations.

But I expect the "Wait 'til next time!" bunch will want to look closer at 7.Qf5+.

(Where did the "extra" games come from? Petasluk - TaccyChan, FICS, 20017; Petasluk - Kosten, 2013 and Petasluk - douthy, FICS, 2015; all transposed from 6...Ke6.)

Monday, July 14, 2008

Beware: Mad Dog!


My buddy A B Hailey
likes to adopt the nickname "Mad Dog"
when he's playing chess.

He's partial to the Evans Gambit and the Sicilian Dragon.


He's been known to ball-bat my Blackmar Diemer Gambits and Latvian Gambits, too. (One of our games shows up in Tony Kosten's The Latvian Gambit Lives!)

Somehow "Mad Dog" got interested in playing the Jerome Gambit... Probably the bad influence of good friends, or something like that.

I like to refer to him as the "unluckiest Jerome Gambit player in the world."

You've seen his first effort in "Jerome Gambit Tournament: Chapter II," abhailey - peonconorejas, net-chess.com, 2008 (0-1, 20) -- just his luck to find someone with the calculating skills of a computer to overturn past theory and execute a nifty Queen sac!


Does "Mad Dog" give up? Of course not!

He comes right back with another Jerome:


abhailey - cruciverbalist
net-chess.com, 2008


1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+ Nxe5 6.Qh5+ g6 7.Qxe5 Qe7

Yippie...

By accident or design Black plays Whistler's Defense -- one of the numerous refutations of the Jerome Gambit.

8.Qf4+ Nf6

He, he, he...

One of the "problems" with a refutation, though, is you have to play it properly. After 8...Qf6 "Mad Dog" can finish his day playing "World of Warcraft." After the text, he's back in the game.

9.e5 Re8 10.d4 Bxd4 11.Qxd4 Qxe5+ 12.Qxe5 Rxe5+ 13.Be3 Nd5

Oddly enough, up until this point the game has been following Jerome - Jaeger, correspondence, 1879 which continued with the inaccurate 13...Ng4, but then Black simply outplayed White: 14.0–0 d5 15.Nc3 c6 16.Bd4 Re7 17.h3 Nh6 18.g4 b6 19.f4 Bb7 20.b4 Rc8 21.Na4 Re4 22.c3 Ba6 23.Rf2 Bb5 24.Nxb6 axb6 25.Bxb6 Rce8 26.Kg2 Ke6 27.a4 Ba6 28.Bc5 Kd7 29.Rb1 Rb8 30.Kg3 Rc4 31.Rc1 Ng8 32.f5 Nf6 33.Ba7 Rbxb4 34.Kg2 Ne4 35.Rf3 Rb2+ 36.Kh1 Rxa4 37.fxg6 Raa2 38.Bg1 hxg6 39.Re1 Nf2+ 40.Bxf2 Rxf2 41.Rg3 Be2 42.Kg1 Bf3 43.Re3 Be4 44.c4 Rh2 45.Kf1 Jaeger announced a mate in three at this point. 0–1

In our game, the attacker chomps on and does not let go.




14.0–0 Nxe3 15.fxe3+ Kg7 16.Nc3 a6 17.Rf3 b5 18.Raf1 Bb7


My analysis buddy Rybka sees the position as equal, and recommends a line where White forces the draw: 18...b4 19.Rf7+ Kg8 20.Rf8+ Kg7 21.R1f7+ Kh6 22.Rh8 bxc3 23.Rhxh7+ Kg5 24.h4+ Kg4 25.Rf4+ Kg3 26.Rf3+ Kg4 27.Rf4+ etc.

19.Rf7+ Kh6 20.Rxd7 Rxe3 21.Rxc7 Be4 22.Re7 Rd8 23.Rxe4 Rxe4 24.Nxe4

The extra piece and pawn are enough to win.

24...Rd4 25.Re1 a5 26.h3 a4 27.c3 Rd5 28.Rf1 Re5 29.Nd6 Re2 30.Rf2 Re1+ 31.Kh2 Re7 32.Nxb5 g5 33.c4 Kh5 34.c5 Re8 35.c6 Rc8 36.c7 h6 37.Rf7 Kg6 38.Rd7 Kh5 39.Rd8 Rxc7 40.Nxc7 Kh4 41.Rd6 g4 42.g3+ Kg5 43.h4+ Kf5 44.Rxh6 Ke5 45.Re6+ Kf5 46.h5 a3 47.bxa3 Kg5 48.h6 Kf5 49.h7 Kg5 50.h8Q Kf5 51.Qf6 checkmate

Later "Mad Dog" even "Jerome-ized" a semi-Italian opening by meeting 1.e4 e5 2.Nf3 Nc6 3.Bc4 h6 with 4.Bxf7+ -- but that is a story for another time,


(Mad dog, sheriff artwork compliments of Jeff Bucchino, "The Wizard of Draws")