For those you of more recent vintage, who have no idea who Keith Hernandez is, I will say briefly that he was a great first baseman for the NY Mets, and one of the heroes of their 1986 World Championship – why he is in this chess article you will see a bit later.

But first, dear readers, you have seen me tortured, drawn if not quartered, and finally emerge victorious with the Dragon – but one prize still eluded me: I had not beaten a Grandmaster since Arnold Denker in 1984, and it was getting close to twenty years! And, I was on my fourth tournament in Hungary!

It was time to win.

I sat down across the board from the Romanian GM Grigore. I understood my own character; I played according to my style, and I listened to Keith …

Timothy Taylor vs. George Grigore
Paks International, Hungary, November 2003
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Be2 0-0 6.f4!

There is nothing like playing the Four Pawns Attack to announce to your opponent that you are playing for a win!!

6…c5 7.Nf3 cxd4 8.Nd4 Nc6 9.Be3

Alert readers will recall my note to move ten of Kosanovic-Taylor ( TRUE COMBAT 2 )

that “I might have played 10.f4 here.” This is an almost identical position – in fact, if you look this position up in ECO, you will find it under E79 (King's Indian) but if the game continues 9… Bd7 10.0-0 you will have to look it up under B38, which is to say, our old friend, Accelerated Dragon!

So you see I have simply reversed colors! White has a Maroczy bind structure with more space and attacking chances – and I'm happy as can be … all I needed to do was switch sides!

9…Qb6

Probably not best, because of the following trick. Fischer made a draw by playing 9… e5 against Uhlmann, but the latter was happy to take a half point – Benko later said “White should not avoid the fight” and I agree with him – there is no clear path to Black's equality.

10.Nc2!

Offering a poisoned pawn, as 10… Qb2?? loses to 11.Na4, snagging the Queen!

10…Qa5 11.0-0 Nd7

Also 11… Bd7 12.a3, preparing to drive the Queen back, is good for White, Taylor – Regan, Los Angeles 2003.

12.Bd2 Qb6+ 13.Kh1

Again offering the b-pawn, as this time the Queen will be snared by Rb1 if she gets too greedy!

13…Nc5 14.b4!

The b-pawn may not have nine lives, but at least three! Now if 14… Nxb4? 15.Rb1 will eventually pick up a piece. It goes without saying that I was cheery as can be, striking with one tactical blow after another!

But what if I had played 6.Nf3 e5 7.dxe5? No such adventures would have occurred! The simple point is that by this time I had found openings I liked (Sicilian Dragon, Four Pawns Attack) that gave me the attacking opportunities I craved.

Quite the opposite of the Schneider game ( TRUE COMBAT 3 ), here the flow of the game was in accordance with my temperament – I never even thought about a draw!

14…Bxc3 15.bxc5 Qa5 16.Bxc3 Qxc3 17.cxd6 exd6 18.Rf3 Qf6?!

Black should not defend the weakling; rather he should sacrifice it with 18… Qa5, to get play on the d-file.

19.Ne3! Qd4

I saw Black's last move coming – he wants to trade Queens.

Jeremy Silman asked me to put in some maxims for up and coming players, but I don't find too many maxims I believe in – but right now, there is one I invented that applies: “NEVER THINK ON AN INTUITIVE SACRIFICE!”

Everyone who has read Tal's great story about rescuing the hippopotamus from the marsh knows about this: an intuitive sacrifice by definition cannot be calculated, so – you either believe or you don't!

I had absolutely NO INTENTION of exchanging Queens, and so –

20.Qf1!

Instantly!!

20…Qxe4 21.f5

Instantly!

And now Grigore thought for a looooonngg time. What can he do about White's attack? The f-pawn may go to f6, the dark squares are cracking, the f-file is dangerous, the white Knight is going to d5 …

He finally decides to eliminate the Knight and f-pawn, even at the cost of kingside weaknesses, and try to build a dark square blockade.

21…Bxf5 22.Nxf5 gxf5 23.Bd3 Qe7 24.Bxf5 Ne5 25.Re1

My first goal is to keep the black Knight away from g6, where it would be a strong defensive barrier.

25…h6 26.Re4 Kh8 27.Rh3 Kg7 28.Rg3+ Kh8

Enter Keith Hernandez from the visitor’s dugout! In his great book, “If At First” he talks about the little voice he would hear in his head, that would tell him to swing at a pitch. He said, “I always listened to that voice.”

I know just what he means, and I think most chessplayers do: that little voice of intuition, of “feel”, that tells you more than you consciously see.

Right now I got a whisper: there is a forced win in the position.

So I looked.

And looked.

And my clock ticked.

And I looked some more.

Now it would have been possible to “override” the voice, say, “there’s no damn win!” and just make a practical move with some winning chances, but nothing clear – and sometimes I have done that in the past, usually to my regret.

This time I stayed firm.

I don’t know how long I looked, I think thirty or forty minutes, I know I got down to less than five minutes for the rest of the game. But I finally found what I knew was there. (You might wish to give yourself half an hour and see if you can find this!)

And you may also want to check out “Fritz Speaks” at the end of this game!

29.Qc1!! Qf6 30.Rf4!!

My gentlemanly opponent was most gracious after the game, and full of praise for this – if I do say myself – amazing idea! The white rook blocks his own queen, but at the same time puts Black's entire position in zugzwang!

There is no defense, as can be seen: for example, any quiet move, let's say 30… Rab8 then 31.Be4 Qe7 (31… Qe6 32.Bd5) 32.Rf5 Qh4 33.Rh3! and White wins. Other variations are similar. Essentially White has set up a concealed battery (Qc1/Rf4/focal point h6) though my thought process was nowhere near that rational.

I can't tell you how I saw this idea, even now – all I know is I listened to the voice!

30…Rg8

After prolonged thought, Grigore finds the only way to continue the game, banking on my time trouble: he avoids the mating lines by giving up the Exchange.

31.Bh7 Qe6 32.Bxg8 Rxg8 33.Rh4 Rg6 34.Rxg6 Nxg6 35.Rxh6+ Kg7 36.Rh3 Qe2 37.Rb3!!

The last difficult move of the game: now if 37… Qxa2 38.Qb2+ Qxb2 39.Rxb2 b6 40.Ra2 a5 41.Rb2 White wins the endgame.

Grigore keeps Queens on, for swindling chances, but now White's play is relatively easy: all Queen exchanges win, so Black can be steadily driven back.

37…Qe4 38.h3 b6 39.Rg3 Kg8 40.Qf1 Kg7 41.Qa1+ Kg8 42.Rg4 Qe6 43.Qd4 Kf8 44.Re4 Ne5 45.Qd5 Qf6 46.Re1 Kg7 47.Rd1 Qf5 48.Qxd6 f6 49.c5 bxc5 50.Qxc5 Kg6 51.Qxa7 Nd3 52.Kg1 Qe4 53.Rf1 Nf4 54.Qf2 Ne2+ 55.Kh2 Qe5+ 56.g3 Nc3 57.Qf4!

A nice finishing touch: if 57… Qxf4 58.Rxf4 Nxa2 59.Rc4! traps the Knight.

57…Qe6 58.a3 f5 59.g4 , 1-0.

The exchange of Queens is inevitable, therefore, Black Resigned.

And so a GM finally fell. How did I win? A razor sharp opening; an intuitive sacrifice; a magical attacking maneuver; solid technique.

But what I will always remember from this game is that search for the win on move 29, where my brain went past my own rational limits, and I created something, I say again, magical .

Fritz speaks!

When I had my brainstorm at move 29, my “Keith Hernandez moment”, I correctly realized that there was a win in the position, and I found it, namely 29.Qc1!, and I went on to score the full point.

However, what I did not realize until some months after the game (and after I had written the first draft of this article) was that on move 29 there was not ONE win, but actually TWO!

While Fritz is quite happy with my 29.Qc1, and agrees with me that White then wins by force, he points out that I had an alternate, completely different, and indeed shorter and less complicated win: 29.Qf4!

Why did I not find this during my long search at the board? The answer is that I looked at it, in fact 29.Qf4 was one of the first moves I considered, but I rejected it just as quickly, after examining (in my head, which is not error-proof!) the following variation: I saw 29.Qf4 Qf6 30.Qh4 which looked good until I realized Black had the defense 30… Nd7, defending the Q while threatening to mate me on a1! So I abandoned the variation, and eventually came up with 29.Qc1.

However, Fritz points out that after 30…. Nd7, I have the simple 31.h3 (which I admit I completely missed in my calculations) which saves White’s K, while winning a whole piece! Black can’t both defend h6 and his N on d7, so something big goes! And if instead Black plays 30… Qh4 (his only other choice) White wins simply with 31.Rh4 Ng6 (forced) 32.Bg6 fg6 33.Rh6+ Kg7 34.Rgg6+ Kf7 35.Rd6 and White easily wins the ending with two extra pawns.