Jerome Gambit: Another Day, Another Refutation

I just viewed the "Refuting the Jerome Gambit #shorts #chess" short video by chesspages1512 on YouTube.

The creator has about 55 short videos on his site with the goal of "Helping with chess skills through exposure to an analysis of games and positions". Check them out and enjoy.

The notion of refuting the Jerome Gambit is not new. I addressed this early on in the blog - 15 years ago - in "But is this *stuff* playable?" and answered (among other places) 5 years ago with "Yes, the Jerome Gambit is Playable (Part 1 & 2)".

I waxed somewhat philosophical in "Jerome Gambit: This 'Refuted' Thing (Part 1, A Basic Truth) & (Part 2, What to Do?)" but got down to basics with "What Makes A Gambit Playable?"

How I remember, long ago, when the Jerome Gambit was not even a twinkle in my eye, the regular complaint of a chess buddy, "I was winning right up to the point where you checkmated me!"

So, let's just enjoy the effort of chesspages1512, and leave it at that, shall we? The Jerome Gambit video features chess moves on a 2D board, without written or spoken commentary. It is a cause for some thinking, and we can be thankful for that as well. (I have added diagrams and subtracted background music.)

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 

4...Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Ng6 

7.Qd5+ Kf8 8.Qxc5+ d6 9.Qe3 Nf6 

This position shows up in over 400 games in The Database.

10.f4 

Checking with the computers:

Komodo 13.02 64 bit liked this move at 20 ply, but at 30 ply preferred 10.0-0 by 1/100th of a pawn;

Stockfish 15 chose 10.f4 early in its analysis but then raced by it, preferring 10.h4 (a novelty according to The Database) at 30 ply. 

The video also gives the alternative 10.Nc3, which appears in about 100 games in The Database. It continued 10...Kf7 11.O-O Re8 12.d3 Kg8 13.f4

Analysis Diagram

Here there is a look at the sideline 13.Qg3 which chesspages1512 indicated appeared in the game Aman Hambleton - Thalia Cervantes Landerio, 3 1 blitz, Titled Tuesday, Chess.com, 2022, (see this blog's "Jerome Gambit: GM vs WGM") with then the novelty 13...Nh5 followed by 14.Qf3 Qh4. Black has the extra piece for two pawns, better development, and an equally safe King. 

The video then continues (after 13.f4) 13...d5 14.f5 (this is new) Bxf5 15.Rxf5 d4 16.Qe1 dxc3 17.Qxc3 Qd6) 

Back to the main line in the video.

10...d5

Some day someone will play the creative line suggested by the Stockfish 15 (not its first choice, mind you): 10...Nxe4!? 11.Qxe4 Qh4+ 12.Kf1 Bd7 13.Qe1 (sad) Qxe1+ 14.Kxe1 Nxf4 15.Kf2 Nxg2 16.Kg3 Bc6 17.Rf1+ Kg8 18.Nc3 Re8.  

11.e5 

At 30 ply Stockfish 15 instead prefers 11.d3 slightly over 11.f5 Ne5 12.f4 Nc6 13.e5. (That second line has a Hypermodern feel to it.) 

11...Nh4 12.Qc5+ Kf7 13.exf6 Nxg2+ 14.Kf2 Nxf4 

15.fxg7 

It is disappointing that the "best" that Stockfish 15 can come up with here is 15.Qe7+ Qxe7 16.fxe7 Kxe7 when Black is a pawn ahead with better development.

The video also show 15.d3 Nh3+ 16.Kg2 gxf6 which the computer evaluates as more than a Rook better for Black. 

15...Qh4+ 16.Kf3 Qg4+ 17.Ke3 Rae8+ 

Black has a checkmate in 3 moves.

So: Truth in labelling.

I suspect, though, that if you play the Jerome Gambit, if you go over this blog post - and over the video - you will come up with more ideas for White, to add to those that Stockfish 15, Komodo 13.02 and I have come up with.

Let me know what you discover.

Jerome Gambit: A Slip on Defense

We see this again and again: Make a slip on attack and you lose the attack; make a slip on defense and  you lose the game.

This time we see mbokhari, with 877 games in The Database, deliver the bad news.

mbokhari - ayaz44jamali

10 0 blitz, lichess.org, 2022.

1.e4 e5 2.Bc4 Nc6 3.Nf3 Nd4 

The Blackburne Shilling Gambit.

4.Bxf7+ 

The Blackburne Shilling Jerome Gambit.

4...Kxf7 5.Nxe5+ Ke8 6.c3 Nc6 

This retreat shows up in 147 games in The Database. White scores 60%. This can be compared with the more solid 6...Ne6, which appears in 272 games, with White scoring 51%.

I was shocked by Stockfish 15's 3rd-best suggestion of 6...Nc2?!. A check with The Database shows that it actually has been played 3 times by Fazmeister, about a decade ago on FICS. He won 1 of those games.

7.Qh5+ 

This straight agression is a bit better than 7.Nxc6 e.g. 7...dxc6 8.d4 c5?! 9.d5 (9.Qh5+!?) 9...c6 10.Qh5+ g6 11.Qe5+ Ne7? (11...Kf7 12.Qxh8 h6 13.Qe5 = ) 12.Qxh8 cxd5 13.Bh6 dxe4 14.Qxf8+ Kd7 15.Qxd8+ Kxd8 16.Bg5 Bf5 17.Nd2 Kd7 18.Bxe7 Kxe7 19.O-O-O Rd8 20.Nc4 Rxd1+ 21.Rxd1 b6 22.Ne3 h6 23.Nxf5+ gxf5 24.Rd5 h5 25.Rxf5 a6 26.Rxh5 b5 27.Rxc5 b4 28.cxb4 a5 29.bxa5 e3 30.fxe3 Ke6 31.e4 Kf6 32.Rd5 Kg6 33.a6 Kf7 34.a7 Ke6 35.a8=Q Kf7 36.Qb7+ Kg6 37.Qc6+ Kf7 38.Rd7+ Ke8 39.Qc8 checkmate, mbokhari - Manikanth20, lichess.org, 2021. 

7...Ke7  

Blitz. It happens.

mbokhari has also seen 7...g6 in several games 

8.Nxg6 Nf6 (8...hxg6 9.Qxh8 Qg5 (9...Nf6 10.d3 d6 11.Bg5 Qd7 Black resigned, mbokhari - AlisheR7292, lichess.org,  2021; 9...Nge7 10.d4 d6 11.Bg5 b6 12.Bh6 Ba6 13.Qxf8+ Kd7 14.Qxd8+ Rxd8 15.Nd2 Kc8 16.Be3 Kb8 17.c4 Nb4 18.O-O Nc2 19.Rac1 Nxe3 20.fxe3 c5 21.d5 g5 22. Rf7 Ng6 23.g3 Rf8 24.Rxf8+ Nxf8 25.Rf1 Ng6 26.b3 b5 27.Rf6 Ne5 28.Rxd6 Bb7 29.cxb5 g4 30.a4 Nf3+ 31.Nxf3 gxf3 32.Kf2 Kc7 33.Rf6 a5 34.Kxf3 Kd7 35.h4 Bc8 36.g4 Ke7 37.Rc6 Bd7 38.Rxc5 Kf6 39.Kf4 Kg6 40.e5 Kh6 41.e6 Be8 42.d6 Bg6 43.d7 Bc2 44.d8=Q Bxb3 45.h5 Black resigned, mbokhari - Atuobi_89, lichess.org, 2020; 9... Qe7 10.Qxg8 Qxe4+ 11.Kd1 Qxg2 12.Re1+ Ne7 13.d3 Qf3+ 14.Kc2 Qxf2+ 15.Bd2 d6 16.Qxg6+ Qf7 17.Qxf7+ Kxf7 18.c4 d5 19.Nc3 dxc4 20.dxc4 Bf5+ 21.Kc1 a6 22.b3 b5 23.Rf1 bxc4 24.bxc4 Rb8 25.Nd5 Nxd5 26.cxd5 Kf6 27.Bc3+ Kg5 28.Rh1 Be4 29.Rg1+ Kh4 30.Bf6+ Kh3 31.Be5 Bh6+ 32.Kd1 Bf3+ 33.Kc2 Be3 34.Rg3+ Kh4 35.Rxf3 Bg5 36.Bxc7 Rc8 37.d6 Kg4 38.Rc3 Bf6 39.Rc4+ Kf5 40.Rf1+ Ke6 41.d7 Kxd7 42.Rxf6 Rxc7 43.Rf7+ Black resigned, mbokhari - Mosa123zakhor, lichess.org, 2020) 10.O-O Nce7 11.d4 Qf6 12.Qxf6 Nxf6 13.e5 Nh7 14.Be3 d6 15.exd6 cxd6 16.Nd2 Bd7 17.Ne4 Nc6 18.a4 Kd8 19.Ng5 Kc7 20.Nxh7 Black resigned, mbokhari - Dr_OmarDawod, lichess.org, 2022) 9.Qh4 Rg8 10.Nxf8 Rxf8 (10...Qe7!?) 11.d3 Ne5 (11...Qe7?! 12.Bg5 Qf7 13.O-O d6 14.Nd2 Ne5 15.d4? Ng6 16.Qg3 Nh5 17.Qe3 Nhf4 18.Bxf4 Qxf4 19.Qxf4 Nxf4 20.Rfe1 Bd7 21.a4 Ke7 22.e5 d5 23.c4 c6 24.cxd5 cxd5 25.Nf3 Rg8 26.g3 Nd3 27.Re2 Bg4 28.Kg2 Raf8 29.Re3 Bxf3+ 30.Rxf3 Rxf3 31.Kxf3 Nxb2? (31...Rc8) 32.Rb1 Rf8+ 33.Kg2 Nd3 34.Rxb7+ Ke6 35.f4 a5 36.Rxh7 Rb8? (36...Nxf4+ 37.gxf4 Rxf4 = ) 37.Rh6+ Kf5 38.Kf3 Ne1+ 39.Ke2 Rb1? (39...Nc2) 40.Rf6+? (40.e6) 40...Ke4 41.e6 Nf3 42.e7 Rb2+ 43.Kf1 Rb1+? (43...Nxh2+) 44.Ke2 Rb2+ 45.Kd1? (45.Kf1) 45...Rd2+? (45...Kd3 46.Rc6 Rf2 47.Kc1 Re2 48.Rc7 Nxh2) 46.Kc1 Kd3 47.e8=Q Rc2+ 48.Kb1 Nd2+ 49.Ka1 Nb3+ 50.Kb1 Nd2+ 51.Ka1 Nb3+ 52.Kb1 Nd2+ Drawn, mbokhari - Ibrahim71, lichess.org, 2022) 12.O-O Nxd3?! 13.Bg5 Nxb2? (13...Qe7) 14.e5 h6 15.Bxf6 Rxf6 16.Qxf6 Qxf6 17.exf6 Kf7 18.Nd2 d6 19.Ne4 Bf5 20.f3 d5 21.g4?! (21.Rab1) 21...dxe4 22.gxf5 exf3? (22...e3) 23.Rxf3 Kxf6 24.Rb1 Nc4 25.Rxb7 c6 26.Rh7 Rg8+ 27.Kf2 Ne5 28.Rxh6+ Kg7 29.Rg3+ Kf7 30.Rxg8 Kxg8 31.Re6 Ng4+ 32.Kg3 Nxh2 Black resigned, mbokhari - Haykonarek123, lichess.org, 2022

8.Qf7+ Kd6 9.Nc4+ Kc5 10.Qd5 checkmate

Jerome Gambit: Attack Quickly

Everything happens quickly in a bullet game, especially an attack.

In the following game, angelcamina - who specializes in 1 0 bullet games with the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+) shows how it is done.

angelcamina - ethylotete_123

1 0 bullet, lichess.org, 2023

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 

4...Kxf7 5.Nxe5+ Ke8 

Declining the second piece has been examined in "Jerome Gambit for Dummies 2.0 (Part 4)" and "The Jerome Gambit Does That To You" (and elsewhere)

I was amused - but not surprised - that the lichess.org computer labelled this move a blunder.

If you have no idea what kind of devastating attack your opponent has just unleashed upon you, it is quite normal to decide not to choke on sacrificed pieces, but to put your King closer to safety. 

Verdict: psychologically sound, chessically suspect.

Why suspect?

While The Database shows little difference between retreating the King to e8 (421 games, White scores 54%) and retreating the King to f8 (614 games, White scores 56%); Stockfish 15 (at 34 ply) shows a definite preference for 5...Kf8, rating it as about 2 pawns better for Black, while rating 5...Ke8 as about a half pawn better for White. 

6.Qh5+ 

As mentioned in "Jerome Gambit: Over Quickly" and still true today

White takes advantage of the fact that Black's King has moved back to e8 - an aggresive check pushes forward his attack.

The computers recommend 6.Nxc6, which would be met by 6...Qh4!?, a counter-attack that remains unplayed.

6...g6  

Or 6...Ke7 7.Qf7+ Kd6 8.Nc4 checkmate, as in angelcamina - SammmDBest, lichess.org, 2021. 

7.Nxg6 

7...hxg6 

Natural, and likely the first thing to come to mind in a bullet game.

Still, it it interesting to point out that the computer recommends 7...Bxf2+, although the Database shows White scores 16 - 7 - 1 (69%) against the move.

angelcamina has also seen

7...Qf6 8.Nxh8+ Kd8 9.Qxc5 Qxh8 10.Qf8 checkmate, angelcamina - DenBlood, lichess.org, 2022; and

7...Nf6 8.Qh4 Nxe4 9.Qxe4+ Ne7 10.Nxh8 d6 11.Qxh7 Be6 12.O-O Qd7 13.Ng6 Nxg6 14.Qxg6+ Bf7 15.Re1+ Kf8 16.Qh6+ Kg8 17.c3 Qf5 18.d4 Bb6 19.Bg5 d5 20.Bf6 Qh7 21.Qg5+ Qg6 22.Qh4 Qh7 23.Qg3+ Qg6 24.Qxg6+ Bxg6 25.Nd2 c6 26.Re7 Bc7 27.Rae1 Bd6 28.Rd7 Bf8 29.Nf3 Rb8 30.Be7 b5 31.Bxf8 Kxf8 32.Ree7 Be8 33.Rxa7 White won on time, angelcamina -  stockfishdanirealno1, lichess.org, 2021) 

8.Qxh8 

White has a Rook and 3 pawns for two pieces, but, more importantly, he has strong play against the enemy King.

8...Kf7 

Better than 8...Kf8 9.O-O d6 10.c3 Be6 11.d4 Bb6 12.d5 Bf7 13.dxc6 bxc6 14.Bh6+ Ke7 15.Bg5+ Nf6 16.Bxf6+ Ke6 17.Bxd8 Rxd8 18.Qxd8 Bxf2+ 19.Rxf2 d5 20.exd5+ cxd5 21.Rf6+ Ke5 22.Nd2 c5 23.Qe7+ Be6 24.Qxe6 checkmate, angelcamina - MegaBane, 1 0 bullet, lichess.org 2019. 

9.Qh7+ Kf6 10.c3 

Okay, this move isn't "faster than a speeding bullet", but it is powerful, nonetheless.

10...d5 11.d4 Bd6 

With a plan to return some of the sacrificed material and stifle the attack - although it has a slight flaw.

12.e5+ Bxe5 13.dxe5+ Nxe5 14.Qh4+ g5 15.Qxg5+ Black resigned

X-ray attack on the unprotected Queen.

Jerome Gambit: Slipping Away

Sometimes a game goes quickly, sometimes it seems to just slip away...

Wall, Bill - Jergens

internet, 2023

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 

4...Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Ng6 

A quick look at The Database shows 6...Ng6 to be the most popular of the three main responses for Black. There are 3,690 games. White scores 57%.

Next in popularity is 6...Ke6, with 3,129 games. White scores 55%.

Then comes 6...g6, with 1,700 games. White scores 52%.

(Of course, if you look only at Bill Wall's games, he scores 99%, 85% and 97%, respectively. Might I recommend 1...e6 ?)

7.Qxc5 d6 8.Qd5+ Be6 

This pawn offer was most recently discussed in "Jerome Gambit: He Who Takes the Queen's Knight's Pawn". 

9.Qxb7 

Stockfish 15, at 30 ply, prefers 8.Qb5 by less than a half pawn. That's not much in human vs human play.

9...Nf6 

Bill has also faced 

9...Nh4 in Wall,B - CheckMe, Chess.com, 2010 (1-0, 23); 

9...N8e7 in Wall,B - Guest1352598, PlayChess.com, 2021 (1-0, 21) and Wall,B - Gilex, 2022 (1-0, 21); and

9...Ne5 in Wall,B - Guest249301, PlayChess.com, 2013, (1-0, 30) 

10.Nc3 Re8 11.O-O Kg8 

Black is okay.

White needs to do something with his extra pawns.

12.f4 Bc4 

This provocation does not gain much (temporarily preventing d2-d4) and does not lose much (only time).

13.d3 Bf7 14.Qa6 d5 

15.e5 Nd7

Black would have gotten more out of the move 15...d4. The game is about even now.

16.d4 Re6 

It is tempting to chase the enemy Queen away, but hitting the center with 16...c5 was more to the point. The pieces on the Kingside seem to call out to White's pawns.

17.Qd3 Ne7 

A retreat is in order, but this move hobbles the Rook and does nothing about the coming e5-e6 fork. Probably 17...Re8 with the idea of meeting 18.f5 with 18...Ngxe5 19.dxe5 Nxe5, returning some of the sacrificed material, was the way to go. 

18.f5 Rb6 19.g4 

Or advancing the e-pawn forthwith.

19...Nc6 20.e6 Bxe6 21.fxe6 Ndb8 

Black's pieces have shifted toward the Queenside, leaving his King in great danger.

22.Rf7 Qh4 23.Qf5 Black resigned

The Jerome Gambit is A Mess

 

In club play, the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+) is a lot like bumping the chess board, watching pieces fall over, then deciding to play on, anyway.

In blitz games the knowledgeable player can take advantage of his opponent's unfortunate need to take thinking time to figure things out in a messy position. This is the value of learning some specifics about the opening beforehand.

When both players are discovering the finer points of an opening as they play it, things can become a bit of a mess - as in the following game.  

Selrah - tc2303

5 3 blitz, lichess.org, 2023

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 

4...Kxf7 5. Nxe5+ Nxe5 6. Qh5+ g6 7.Qxe5 Qe7 

This is Whistler's defense, which Alonzo Wheeler Jerome faced in the games of his correspondence match with Lt. G. N. Whistler, secretary of the Lexington, Kentucky Chess Club, in 1876.

If White now accepts the offer of a Rook with 8.Qxh8, he enters into difficult complications.

"Objectively" - if we can trust Stockfish 15 to be objective - the position is then rated as only a pawn shy of Black being a Queen better.

However, The Database has 649 games with 8.Qxh8, and White scores 62%. This is despite the fact that the average rating for White and for Black in those games is almost identical.

The conclusion has to be that White often navigates the confusion of the Jerome Gambit better.

8.Qxh8 

It is far safer for White (even if it still leads to Black's advantage) for his Queen to back away with 8.Qf4+. 

8...Kf8 

Black defends (his Knight) first before attacking (with 8...Qxe4+) - in fact, this suggests that he was defending his Bishop on the previous move, rather than knowingly playing one of the strongest anti-Jerome lines.

White's King can now escape danger with 9.0-0, 9.d3 or 9.Nc3.

9.Na3 

This move is a bit of a puzzler. I am not sure what is going on. Perhaps it is a mouse slip for 9.Nc3.

9...Qxe4+ 

This time Black sees the move.

The irony is that Stockfish 15 now sees White as a piece better after 10.Kf1. What has happened to dilute the power of Black's Queen's capture (with check) on e4?

The answer is that in some of White's defensive lines he now can produce a useful check because Black's King has moved to f8 - for example, 10.Kf1 Qh4 11.d4 Bxd4 12.Bh6+ Qxh6 13.Qxd4 and White is up the exchange.

Complicated? Like I said, the Jerome Gambit is a mess. Plus, remember, this is a blitz game.

10.Kd1 

Thinking about the earlier suggested line starting with 8.Qxe4+, it did not much matter if White responded by moving his King to f1 or d1 - he was going to be in grave danger either way (although d1 was a little bit less weak).

Here, however, White needs to protect his g-pawn with 10.Kf1. The text leads to great difficulties. 

 10...Qxg2 

The alternative 10...Qg4+ is more than sufficient, e.g. 11.Ke1 Bd4 12.Qxh7 Qxg2 13.Qh4 (13.Rf1 leads to being checkmated in 30 moves) Qxh1+ 14.Ke2 Bf6 15.Qg3 and Black's extra piece (and the two Bishops) gives him a clear advantage.

11.Re1 

11...Qxf2 

Following the blitz rule of thumb: when in doubt, grab material.

Black is on unfamiliar ground and so does not yet focus on how the light squares affect the fate of White's King - otherwise, he would have liberated his Bishop immediately with 11...d5 with threats of ...Bg4+. But second chances do come.

Now Stockfish 15 suggests that White play 12.d4, threatening Bh6+ and strongly encouraging Black to seek a draw with 12...Qf3+ 13.Re2 Qf1+ 14.Re1 Qf3+, etc.

12.Qe5 

White is happy to get his Queen back in play and team up with the Rook. Unfortunately, it is not enough.

12...d6 13.Qe8+ Kg7 14.d4 Bxd4 15.Re7+Nxe7 16.Qxe7+ Kg8 

Alas, it does not work both ways: White does not have a way to a draw by multiple Queen checks.

17.c3 Bg4+ 18.Qe2 Qxe2 checkmate

Someone who glanced at the final position might have erroniously concluded "White should have castled to safety, as Black's King did".

Jerome Gambit: The Return of Cliff Hardy (Addendum)

 

I was able to find on TwitchTV a video including the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+) game between "Cliff Hardy" and Grandmaster Martin Kraemer (playing under the handle DrawDenied_Twitch) covered in the last 3 posts (1, 2, 3).

Although I speak French like a Spanish cow, my German is even weaker - having picked up what little I know mostly from Randspringer magazines - so Grandmaster Kraemer's German language commentary and the chat in the background is best appreciated by those who sprechen deutsche.

I was able to note someone in the chat during the early moves of the game labelling Cliff Hardy as ein Trickspeiler - a trickster - which seems appropriate for someone who plays the Jerome Gambit.

Eventually there was the chat comment, Now I have it: it is the Jerome-Gambit. (Thank you, Google Translate.)

As Cliff commented to me

I even made out a tiny bit of English which explained a couple of his feelings on the moves. He said, "Easy peasy" when I allowed 26...Bg4+, which I guess was justified and "idiot" when I had the chance to win a piece with 36.Rf5+.

There was also a chat reference to Grandmaster Aman Hambleton, whose YouTube video on the Jerome  has  had almost 3/4 million views - pretty impressive. 

In the Twitch TV video, I watched with excitment as the grandmaster's clock ran down almost to zero. Would he escape a time forfeit? I watched a second time, again expecting him to flag. It was like the old joke 

Two friends decide to go to the movies together. The picture had a scene with a horse race in it. Before the race starts, one friend turns to the other and says "I'll bet you fifty dollars the black horse wins." The second friend says "OK, you're on!" The scene ends with the black horse barely winning, so the loser pays up. The second friend decides to confess; "I have to admit that I saw this movie last week." The first replies, "So did I, but I didn't think that black horse could possibly win a second time!"

Jerome Gambit: The Return of Cliff Hardy (Part 3)

 

[continued from the previous post]

Cliff Hardy (2478) - GM Martin Kraemer (2895)

10 0 rapid, Lichess, 2023

This may not look like the typical Jerome Gambit blitz game, but it has a significant, familiar feature - the defender has to use time to figure out how to respond, and that puts him behind on the clock. - Rick

36.Rf6??

Missing 37.Rf5+! Kb6 (not 37...Ka4?? 38.Ra5 mate!) 38.Rxh5 = and with a massive time advantage and suddenly an equal position on the board, I would have had an excellent chance of winning the game.

36...Bg4 37.Rxh6 Rxh2 38.a3 Rg2 39.Rxd6 h2 40.Rh6 Rg3+ 41.Kd4 Rh3 42.c4+ Kb6 

43.Rg6 h1=Q 44.c5+ Kb5 45.Rxg4?? Qd1?

The grand disaster misses mate in one with 45...Qa1! but since he only had 8 seconds left on his clock, I guess we can be charitable to him.

The rest of the game is a sprint, but the grandmaster is up to it, clockwise. - Rick

46.Ke5 Qxd3??

Oops! 46...Qxg4 is better.

47.Rg5 Qc3+ 48.Kf5 Rf3+ 49.Ke6 Qf6+ 50.Kd7 Qf7+ 51.Kd6 Qf6+ 52.Kd7 Qxg5 53.e5 Qg7+ 54.Kd6 Qf8+ 55.Kd7 Rf7+ 56.Ke6 Qe7 mate

Checkmating me with 4.2 seconds left on his clock.

As Maxwell Smart used to say, "Missed by that much."

The clock is the 33rd piece. - Rick

I got beaten nicely but he only just made it in time before his clock ran out and if I'd taken more time (I finished the game with 1 minute and 12 seconds left on my clock) and possibly found 37.Rf5+!, I could have completely turned the tables!