Yesterday, I happened to come across some analysis on a new idea in the KG. It was extra unexpected as it was posted at the Chesspub forum in a thread about

*a new book on the Petroff*, due next year.

Here's a direct link to the analysis in the thread (For easy viewing, you can copy the analysis, paste it into Notes and save it as PGN).

A closer inspection revealed that the new idea had already been played twice in corr. So, to write my first blog on the dear old King's Gambit, here's a look at what happened in those two encounters, intertwined with most of the analysis & comments provided by Har-Zvi and Gupta at chesspub, plus some remarks by me and my engine.

**1.e4 e5 2.f4 exf4 3.Nf3 d5 4.exd5 Nf6 5.Bc4 Nxd5 6.0-0 Be6 7.Bb3 Be7 8.c4 Nb6 9.d4 Nxc4 **Ronen Har-Zvi and Ankit Gupta: 'The theoretical line that supposedly gives equality for Black.'

**10.Nc3**
By the way, 10.Bxf4 is met by 10...Nb6! 11.Bxe6 fxe6 12.Qe2 Nc6 13.Nc3 Nxd4! as I mentioned in the FKG 5-6 years ago, though by now it almost feels like in another lifetime...

**:/**
** 10...Nb6**
[

10...c6 Shulman - Onishuk, 2003

11.Re1 "N With the idea of Rxe6 and pressure on the c4 square." - Ronen Har-Zvi and Ankit Gupta:

11...b5 12.a4! "Black's position comes quickly under fire. It's hard for Black to activate his pieces and finish develop in a manner that will not give White a small edge." - Har-Zvi and Gupta.

However, nothing is new under the sun - unless you don't have a database of corr-games. ;)

And to not refer to corr nowadays, is not just snobbish anymore, but quite illogical, as in both OTB and in Corr, players get most of their theoretical inspiration from the use of engines. So these days, a chess book that doesn't refer to correspondence games will age at the rate of a vampire exposed to sunlight.

1. e4 e5 2. f4 d5 3. exd5 exf4 4. Nf3 Nf6 5. Bc4 Nxd5 6. O-O Be6 7. Bb3 Be7 8.

c4 Nb6 9. d4 Nxc4 10. Nc3 c6

11. Re1 b5 12. a4 b4 13. Rxe6 fxe6 14. Bxc4 bxc3
15. bxc3 O-O 1/2-1/2 B, Baroin (2223) - A G, Savage (2405), theme corr 2006. To accept the draw so early looks like a peculiar decision since thematic events are unrated, but maybe white wanted to save his findings for another tournament...?]

**11.d5 Bg4 12.Bxf4 **(For what it's worth, 12.Qe1 0-0 offers the extra option of 13.Ne5)

** 12...0-0 **Fedorov - Goloshchapov, Ohrid 2001.

** **
**13.Qe1!? **A check with my database, showed that this new idea had already been tried twice in corr chess.

So here's a look at what happened in those encounters, plus most of the analysis provided by - Har-Zvi and Gupta, plus some remarks by me and my engine.

**Now, we have already come to a fork in the road. **

**A) 13...Na6 **or** B) 13...Bd6**

**A)** **13...Na6 ****14.Qg3 Bh5 **
[14...Bf5 15.Rad1 Nc8

* (15...Bd6 16.Nh4 Bg6 17.Nb5 *"If you look at this position it's actually very hard to find a good move for Black. Black is quite cramped and White has several options for improving his piece placement." - Ronen Har-Zvi and Ankit Gupta.

*) *16.d6! "The purpose is to open diagonals for White's pieces. Very common thematic sacrifice in the IQP" - Ronen Har-Zvi and Ankit Gupta. 16...Nxd6 17.Ne5 Sacrifices against f7 are in the air. 17...Be6 18.Bh6 Bf6 19.Bc2 "Black will likely have to acquiesce to g6 and sacrifice the exchange". - Ronen Har-Zvi and Ankit Gupta.]

** **
And now a 2nd fork, should Ra1 go to the open e-file or position itself behind the d-pawn? Both make sense, and lead to similar positions, so it's a decision one could base on general feeling or a large amount of analysis...

**15.Rae1 **Diagram This was played in the most recent game of the two.

**15...Bg6 16.Ng5 Bd6 17.Nce4 Nc8 ** [17...Bxf4? 18.Rxf4

* (18.Qxf4? h6 19.Nf3 Bxe4 20.Rxe4 Nc5) *18...Nxd5 19.Bxd5 Qxd5 20.Nxh7! Qd4+ 21.Kh1 Qxb2 22.Nxf8 Rxf8 23.Qh3 Bxe4 24.Rfxe4 g6 25.Re8 and white should be winning.]

**18.h4!? Nc5 19.Nxd6 Nxd6 20.Be5 **Diagram

**20...f6 **This obviously invites trouble on the white squares, but the engine says "0.00"...

[A safer defence might be: 20...Nxb3 but maybe 21.h5!? works, if you look into it properly? Something like this:

* *21...h6 22.Nh7 Nd2 23.Nxf8 Nxf1 24.Rxf1 Ne4 25.Qg4 Qxd5 26.Nd7 Rd8 27.hxg6 Rxd7 28.gxf7+ Rxf7 29.Rxf7 Kxf7 30.Qxg7+ etc. The alternative would be 21.axb3 Ne8 22.Qf3 Nf6 23.Bxf6 Qxf6 24.Qxf6 gxf6 25.Ne4 which looks holdable for black after Bxe4.]

** **
**21.Ne6 Nxe6 22.dxe6 Qe7 23.Bf4 a5 24.Bd1 Be8 25.Rf2 a4 26.Rd2 Rd8 27.Bc2 **Diagram

White has a passed centre pawn deep in enemy territory plus all his long range pieces left on the board to back it up. Still, the engine believes the extra pawn makes it equal, somehow...

**27...Bc6 28.h5 Nc8 29.Rf2 Rd5 30.h6 g6 31.b3 Nd6 32.bxa4 Rc5 33.Bc1 **Diagram

Here, the engine finally starts seeing a white advantage, which grows and grows...

** **
**33...Rc4 34.Bb3 **Diagram

Finally, the bishop is back where he belongs!

**34...Rc5 **(34...Ne4 35.Ba3! Nxg3 (35...Qxa3 36.Qxc7) 36.Bxe7+-)

**35.Ba3 Rg5 36.Qd3 Rh5 37.Rd1 Rxh6 **(37...Rd8 38.Qd4+-)

**38.Bxd6 cxd6 39.Qxd6 Qxd6 40.Rxd6+- **Diagram And now, black can safely resign, but for some reason he keeps on "fighting". It seems that following the moves suggested by the engine is easier than taking the consequences of its evaluations...

The game ended:

**40...Kg7 41.a5 Rh5 42.a6 Kh6 43.e7 Re8 44.Re2 Ra5 45.axb7 Bxb7 46.Bf7 Raa8 47.Bxe8 Rxe8 48.Rd8 Bc6 49.Re6 Ba4 50.Rxe8 Bxe8 51.Rxf6 Kg7 52.Rf8 **** 1-0 **A, Krzyzanowski - R, Tacke-Ungruh, corr ICCF 2009.

And now over to the analysis in the chesspub thread, which chooses to put the rook behind the advanced pawn instead.

**15.Rad1** Diagram

**15...Bd6 16.Ne4 Nc8 17.Nfg5!** Diagram

**17...Bg6**
(17...Bxd1 18.Bxd6 Nxd6 (18...cxd6 19.Nf6+ Kh8 20.Nfxh7

(RR 20.Qd3 g6 21.Bxd1 (>Qh3) also wins - TJ.) 20...Bh5 21.Qh4 "And Black's position is completely lost." - Ronen Har-Zvi and Ankit Gupta.) 19.Nf6+ Qxf6 20.Rxf6 Bxb3 21.Qd3+/- - Ronen Har-Zvi and Ankit Gupta.)

(Btw, 17...Bxf4 18.Rxf4 Bxd1? also runs into 19.Nf6+ )

**18.Bc2** "Maneuvering the bishop to this new diagonal is very logical as it had no realistic prospects on the closed a2-g8 diagonal." - Ronen Har-Zvi and Ankit Gupta.

**18...Bxf4 19.Rxf4 Nd6 20.Nxh7!!** **Nxe4 **
(20...Bxh7 21.Nf6+ Qxf6 22.Bxh7+ Kxh7 23.Rxf6 gxf6 24.Qh4++-; 20...Bxe4 21.Bxe4 Nxe4 22.Rxe4 Kxh7 23.Qh3+ Kg8 24.Rh4 f6 25.Rh7 (>Qh5) 25...Qe8 26.Rh8+ Kf7 27.Qh5+ Ke7 28.Re1+)

**21.Qxg6!** **fxg6 22.Rxf8+ Kxh7 23.Rxd8 Rxd8 24.Bxe4** "This resulting endgame slightly favors White, even with the IQP because of the weakness of the c7 pawn, and the doubled g-pawns." - Ronen Har-Zvi and Ankit Gupta.

What, a queen sac with no diagram??

Well, actually, it turns out that the more pedestrian

**21.Bxe4!** Diagram

is much stronger, though less spectacular, e.g:

**21...Kxh7** * *(21...Bxh7 22.Bxh7+ Kxh7 23.Qh3+ Kg8 24.Rh4)

* ***22.Qh3+ Kg8 23.Rh4 Qxh4 24.Qxh4 Bxe4 25.Qxe4+-** - TJ.

So, black must retrace his steps, and try to find out where to improve. Which brings us back to the fork at move 13 and the first recorded instance of 13.Qe1!?

**B) 13...Bd6!?** Diagram This strikes me as the most logical, since Na6 + Bd6 as above, spend two moves to deal with the pressure on the h2-c7 diagonal. If Be7-d6 is necessary anyway, it makes sense to play it at once, allowing Nb8 the option of going to d7 instead.

**a) 14.Bxd6!?** might be most promising, though it's not easy figuring out exactly what's going on... E.g:

**14...cxd6** * (14...Qxd6? 15.Ne4 *+/-

*) ***15.Qg3 Bxf3 16.Rxf3 Nc8** * (16...N8d7 17.Qxd6 Nc8 18.Qf4 Qb6+ 19.Kh1 Nd6 20.Raf1 *and white has an obvious edge.

*) ***17.Ne4 Nd7 18.Raf1 Ne5 19.Rf6 *** **(19.Re3 f6) ***19...Qb6+ 20.Kh1** Diagram

Black has held on to the extra pawn, but paid by giving up the harmony in his position. So white has at least compensation for the pawn, perhaps even an edge if the engine can be trusted, e.g:

**20...Qb5** (20...Rb8 21.Qh4 Qe3 22.Bc2 h6 23.R6f2 Kh8 24.Ng3 Qg5 25.Qb4; By the way, the "natural" 20...Kh8?? loses to 21.Qh3! (>Ng5) e.g. 21...Qd4 22.Ng5 Qd3 23.Nxf7+ Rxf7 24.Rxf7 Ne7 25.Qxd3 Nxd3 26.Rxe7)

* ***21.Bc2 h6 22.R6f2 **- this will of course need a lot more work, so see it as a place to start your own investigations...

**b) **14.Ne5 is another option, I kind of like the look of it, but the computer isn't all that impressed... But if you got the time, it might be worth a closer look. 14...Bh5 15.Qg3!?

* (15.Ne4 Na6 16.Rc1) *;

And now what happened in the game, which of course follows the engine's main line.

**c) 14.Qg3 Bxf4 15.Qxf4 Bxf3 16.Rxf3** Diagram

(Btw: Just before posting this, it struck me that

**15.Qxg4!?** may be worth a much closer look, as opposite coloured bishops usually favour the one with the initiative.)

At first glance, white looks obviously better, but black can get his knight to block the d-pawn just in time.

**16...Nc8!** (16...N8d7 17.d6 cxd6 18.Bxf7+ Kh8 19.Qxd6)

** **
**17.Ne4**
(17.Rh3 Nd6 18.Ne4 (18.Nb5 Nxb5 19.Bc2 is a brutal draw, according to my engine.) 18...Qe7 19.Bc2 h6 20.Nxd6 Qxd6 21.Qf5 g6 22.Qd3 Nd7 23.Rxh6 Qb6+ 24.Kh1 Qxb2 25.Rf1 Qg7 26.Rh3 Rae8 and black is alright, though white is probably also alright, i.e he should be able to hold a draw.)

**17...Nd6 18.Raf1 Qe7 19.Re3 Nd7 20.Ng5 Qf6 21.Qg4 Qg6 22.Qxd7 Qxg5 23.Qe7 Qh6 24.Rh3 Qd2 25.Qxc7 Qd4+ 26.Kh1 Rfd8** Diagram

**27.Qc3?** (27.Qc2 White should keep the queens on and look for ways to trade the d5 pawn for something. At least, it won't be easy for black to make headway without allowing counter-chances.)

**27...Qxc3 28.Rxc3 Rac8 29.Rxc8 Rxc8** Now the odds are firmly in black's favour. The passed pawn is securely blocked and a long term weakness, while white has no counter-play. You can call me biased, but I don't want to look too closely at what now happened to white.

The game ended thus:

**30.Re1 b5 31.Kg1 a5 32.Kf2 Kf8 33.a3 g6 34.Ba2 Kg7 35.Re2 Kf6 36.h3 h6 37.Kf3 Rc1 38.g4 Rd1 39.Kf2 Rd3 40.Kg2 a4 41.Bb1 Rxd5 42.h4 Nc4 43.Ba2 Rd4 44.Kg3 g5 45.Bxc4 bxc4 46.hxg5+ hxg5-+ **Definitely time to resign, but hope springs eternal...?

** 47.Kf3 Rd3+ 48.Kf2 c3 49.bxc3 Rxc3 50.Re4 Rxa3 51.Kg2 Ra1 52.Kf2 a3 53.Ra4 Ke5 54.Kf3 Kd5 55.Kg2 Kc5 0-1** Venus,U-Schaefer,U www.desc-online.de 2006.

Now we can look forward to

*Play the Petroff* by Ronen Har-Zvi and Ankit Gupta in 2011, which, hopefully will give King's Gambiteers all around the world more chances to push the f-pawn at move two! :)

And if they are as generous with interesting ideas in their book as in the chesspub forum, I may well be tempted to get a copy myself!

edit: just opened a new

chesspub thread on 13.Qe1!?