1984-04-05 - Humphreys, Jerry 1-0 Elliott, Sean
A tough battle in a Sicilian Najdorf. Jerry sacs an Exchange for a lasting initiative that eventually erodes the Pawn structure around Black's King. The end comes in a frantic time-scramble and Deep Rybka has plenty to say.
[Event "BRISTOL LEAGUE DIV 1"] [Date "1984.04.05"] [White "HUMPHREYS, J."] [Black "ELLIOTT, SJ."] [Result "1-0"] {A notable game between a D&F stalwart who came to us from Bristol University, and a noted Bristol Junior who came to us via the then Hanham & Kingswood club, now Hanham Folk Centre, for whom he was playing in this match. This is an analytical feast for the tacticians amongst you, but beware - Deep Rybka has a lot to say.} 1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. f4 {"I have several times seen Sean Elliott play the Sicilian Najdorf against D&F players. Each time he has run into a prepared line and lost, except on one occasion where he avoided a prepared line and blundered into a mate."} e5 7. Nf3 Nbd7 8. Bd3 {"8 a4 to prepare Bc4 is usual. Black then plays 8...Qc7 and White can play Bd3, Qe2 and eventually Bc4." It is recommended to play 8... b5 here to take advantage of the omission of 8 a4, but Sean chooses something different.} g6 $6 {This is very rare indeed. I could only find two games with this move, one of which was Prasad-Britton, Lloyds Bank Masters, London 1987. Britton later chose to develop his Queen's Bishop with ...b7-b6 and ...Bc8-b7 so there is no resemblance to the current game. I don't like this move and neither does Deep Rybka. The basic problem is the backward Pawn on d6 which now loses its primary defender. Of course White has blocked the d-file with the d3 Bishop but that can always move again. From hereon Jerry takes the view that the position is similar to a Pirc Defence, Austrian Attack where Black has added the weakening move ...e7-e5. Hence Jerry starts a traditional build-up on the Kingside which involves the Queen manoeuvre Qd1-e1-h4.} 9. a4 { Jerry corrects the omission of his last move. To be fair some people play this on move 6!} Bg7 10. O-O Qb6+ {"Black has played this without losing a tempo with ...Qc7 first [as was quite common in this line]. However, he could have tried 10...exf4 11 Bxf4 Qb6+ 12 Kh1 Qxb2." Deep Rybka and I spent a while on this. First there is 13.Qd2 Qb4 14.Nb5!? Qxd2 15.Nxd6+ Ke7 16.Nxc8+ Raxc8 17. Bxd2 Nc5 (not so sure about this) 18.e5 Nxd3!? 19.exf6+ Bxf6 20.cxd3 Bxa1 21. Rxa1 Rhd8 22.Re1+ Kf8 23.d4 which the engines judges to be 0.24 in White's favour. Some people would like the Rooks and some would like the minor pieces, so 'you pays your money and you takes your choice'. Secondly there is 13.Qe1 Nh5 14.Bd2 0-0 15.Rb1 Qa3 16.Rb3 Qc5 17.Nd5 which it gave as roughly +0.7. Personally I might have left Rb3 out of this but even then the evaluation seems a touch generous to White. Nevertheless snatching the Queen's Knight Pawn is supposed to be dangerous so Sean declined and he kept on declining. Having said that, then what was the point of playing ...Qb6?} 11. Kh1 O-O { "Jerry indicates the line 11...exf4 12.Qe1 Nh5?!" The ?! was my original annotation. It doesn't take long to spot 13.Nd5 Qd8 14.Nxf4 Nxf4 15.Bxf4 and White stands better. If Black gets greedy and goes after the b2 Pawn with 15... Bxb2? hoping for 15 Rb1 Be5 maybe, White sacrifices the Exchange with 16 Bxd6! Bxa1 17 Qxa1. The squares f7 and g7 are prime targets here and Black's Rook can't guard both. Deep Rybka has this as well over +3 for White. No surprise, just look at the difference in development and Black's dark squares.} 12. Qe1 exf4 13. Bxf4 Re8 {"If 13...Qxb2 14 Bxd6 menaces Bxf8 and Rb1." Deep Rybka confirms this with the continuation 14...Ng4 15.Bxf8 Bxc3 16.Rb1 Bxe1 17.Rxb2 Bc3 18.Rb3 and White emerges an Exchange up this time!} 14. Be3 Nc5 {This time if 14...Qxb2 then 15 Rb1 Qa4 16 Rb3 traps Her Majesty.} 15. a5 Qc7 $2 {"If 15.. .Qxb2 16.Bxc5 dxc5 17.Rb1 Qa3 18.Rb3 traps the Queen." It would if the a-Pawn was still on a4, but it's now on a5 hence 18...Qxa5. Instead 15...Qxb2 16.Bd4 - threat 17 Nd5 - 16...Qb4 (or 16...Be6 17.Nd5 Qxd4 18.Nxd4 Bxd5 19.Qb4 with a plus for White according to Deep Rybka but no Black Queen!) 17.Bxf6 Bxf6 18.Nd5 Qxe1 19.Nxf6+ wins a piece 19...Kg7?? 20.Nxe8+ adds a Rook. I wrote that "15...Nxd3 16.cxd3 Qxb2 appears to be playable but risky..." It's not playable because of 17.Ra2 Qb3 18.Nd4 Qb4 19.Ra4 Qc5 (If 19...Qb2 20 Rf2 traps the Queen again.) 20.Nf5 Qc7 21.Nxg7 Kxg7 22.Bd4 winning a piece once more. "...but a better Queen retreat seems to be 15...Qd8 to keep an eye on White's intended Qh4." I'm glad I got something right in this note! Deep Rybka says Black should exchange on d3 and play ...Qd8 either way around. The problem with ...Qc7 is that White later picks up a tempo with Nd5.} 16. Qh4 { All part of the planned Kingside attack. Deep Rybka has White at +0.71 whereas it was only +0.3 after 15...Qd8.} Ne6 $2 {"Jerry comments 'This defensive move, over which Black spent nearly twenty minutes, allows him to play ...h7-h5 and restrict White's options'." The ? is mine. I much prefer 16...Be6 to control d5 and get the Rooks connected. In the game the Bishop doesn't get out until move 30. Deep Rybka also doesn't like the text but it offers instead 16... Ng4 17 Ng5 h6 18 Nd5 Qd8 19 Rxf7 hxg5 (or 19...Re5 20 Bc4 with +1 according to the engine.) 20 Bxg5 with a big attack. No human is going to play a move such as 16 ...Ng4 allowing 17 Ng5 attacking h7 and f7, so 16...Ne6 is understandable from that point of view. If we look back at Black's development scheme then although the d6 Pawn isn't a weakness the d5 square is and Black's fianchetto is coming under pressure.} 17. Ng5 {This move is okay but I prefer 17 Nd5 Nxd5 18 exd5 Nf8 and then White has options of 19 Bb6, 19 Rae1 or 19 Ng5. The isolated e-Pawn is gone and structurally White has a very good 4-3 majority on the Queenside. Deep Rybka says that 17 Nd4 is stronger than either. That's very understandable when you see lines like 17...Nxd4 18. Bxd4 Nd7 (Not 18...Nh5 19.Bxg7 Kxg7 or 19...Nxg7 20.Nd5) 20.g4) 19.Bxg7 Kxg7 20.Rxf7+! Kxf7 21.Qxh7+ winning. Instead it offers 17...Ng4 18.Qxg4 Nxd4 19. Nd5 Qd8 (if 19...Bxg4 20 Nxc7 forks the Black Rooks) 20.Qg3 Ne6 but still has White at +1. After the text White does threaten 18 Rxf6 Bxf6 19 Qxh7+ Kf8 20 Nd5 and the Queen is overloaded defending f7 and f6. See how many lines in which Black's 15...Qc7? comes back to haunt him.} h5 {After 17...Nxg5 18 Bxg5 Ng4 (18...Nd7 was given in the original notes) 19 Nd5 White's plus goes up to over 2. The weak squares and lack of development in Black's position are very obvious. The text is the only other sensible defence to the threat of 18 Rxf6 and 19 Qxh7+ as 17...h6 is met by the simple 18 Nxe6 and 19 Bxh6.} 18. e5 $6 { Given an !! originally with the comment "A brilliant move which, in conjunction with the following combination, punches a hole through Black's Kingside and gives him tremendous problems while short of time." Sadly the combination has a hole in it, and also the position that White gets at the end of the combination isn't as good as the one he starts with. We'll get to the analysis in a moment. The first thing to note is that there is a simple and clear advantage for White after 18.Nxe6 Rxe6 19.Bg5 Ng4 20.Nd5. No-one will dispute this, not even Deep Rybka. Secondly I have always wondered exactly what difference 18 e5 made to the forthcoming combination. What if White actually starts with 18 Rxf6? We get these lines: 18.Rxf6 Bxf6 19.Nd5 Qd8 20. Nxf6+ Qxf6 21.Rf1 Qd8 - to maintain the pin on the Knight - 22.g4! - to pry open the h-file. White has a big local superiority as Black's Queenside pieces do little for the defence - and now: a) 22...f6 23.Nxe6 Bxe6 (if 23...Rxe6 24. Bc4 regains the Exchange while keeping the attack) 24.Rxf6 Bf7 25.gxh5 Re6 26. Bg5 Rxf6 27.Bxf6 Qxa5 28.hxg6 wins. The back rank checks are to no avail. b) 22...Nxg5 23.Bxg5 Qd7 24.gxh5 Qh3 (24...d5 is a bit better) 25.hxg6!! fxg6 (if 25...Qxh4 26 gxf7+ Kf8 27 fxe8=Q+wins or 26...Kg7 27 fxe8=N+! wins. Not quite an Excelsior Pawn because it didn't make five consecutive moves.) 26.Bc4+ (24.. .d5 would have prevented this.) 26...Kg7 and now, not unsurprisingly, there's a mate in six moves with 27.Bf6+ Kf8 28.Be7+ Kg7 29.Rf7+ Kg8 30.Rf8+ Kg7 31. Qf6+ Kh6 32.Qh8#. The hole in the game continuation does not appear in these lines so it appears that 18 Rxf6 wins, but you need to see that 22 g4! opens the h-file and that the attack succeeds. Not easy. So what is the point of Jerry's 18 e5? Well, it is to clear the e4 square for the Knight in some lines - this doesn't actually happen -, the fourth rank for the Queen in others - occurs in a couple of lines - but mainly to open up the b1-h7 for the Bishop. Jerry intends to take on f7 and, as comes to pass, he'll then take on g6 and h5.} dxe5 19. Rxf6 Bxf6 20. Nd5 Qd8 21. Nxf6+ Qxf6 22. Rf1 {All as planned by Jerry but this is where the hole in his combination appears. Black should now play 22...Nf4! In the line beginning with 18 Rxf6 the Black Pawn now on e5 was actually on d6 so ...Nf4 was not a candidate move because it was simply en prise. Now it's defended by the e5 Pawn. Deep Rybka continues with 23.g3 which allows 23...Qc6+ 24.Be4 Qb5 25.Kg1 Ne2+ 26.Kg2 f5 when Black is the Exchange up and better. This is not to say White may still have practical chances but a great deal of the activity he gets in the game would have been muffled.} Qd8 $2 {After all the above analysis this is clearly not the right move but Black is by no means lost.} 23. Rxf7 Nxg5 {"If 23...Nf8 then 24 Qc4! is very strong." Indeed this switch across the fourth rank would completely justify Jerry's 18 e5. Lines go: a) 24...Be6 25.Rxf8+ Rxf8 26.Qxe6+ Kh8 27.Qxg6 wins or b) 24.. .Ne6 25.Qe4 Nf8 26.Rxf8+ Kxf8 27.Bc5+ Re7 28.Qxg6 mating or c) 24...Kh8 25. Rh7+! Nxh7 26.Nf7+ Kg7 27.Bh6+ Kf6 28.Nxd8 wins the Queen.} 24. Bxg5 { Astoundingly Deep Rybka says that this natural recapture is not the right move! Instead it suggests 24 Rf1 with the amazing line 24... Nf7 25.Qg3 Re6 26. Bxg6 Rf6 27.Bxh5+ Kf8 28.Bc5+ Nd6 29.Rd1 Qe7 30.Qa3 winning. Unfortunately Black does much better with the simple 24...Rf8 25 Rxf8+ Qxf8 26 Qxg5 Bf5.} Qxa5 $2 {Black was in deep time trouble now but even so he should have seen 24. ..Qd4! simply forcing the swap of Queens. White has no sensible way of continuing while his Rook hangs. The centralising move is always better than snatching a Pawn on the wing. Once again this is a direct result of including the moves 18 e5 dxe5. So the conclusion must be that 18 e5 was a mistake, but it's taken 35 years to prove it!} 25. Rf1 {A practical decision by Jerry who was also in time trouble. However, Deep Rybka has a field day with 25.Rf6! a) 25...Qa1+ 26.Rf1 Qa5 (if 26...Qa2 27.Bc4+ wins the Queen) 27. Bxg6 and the Kingside collapses just as Jerry planned on move 18! b) 25... Re6 26.Bc4 Qa1+ 27.Rf1 or c) 25...e4 26.Rxg6+ Kf7 27.Rf6+ Kg7 28.Qxh5 mates, or the difficult one d) 25...Bf5 26.Qc4+ Kh8 27.Rxf5! gxf5 28.Bf6+ Kh7 29. Qf7+ Kh6 30.Qg7 mate. A lot of these lines have two White units working together to control f1, but it's understandable that Jerry cuts out all the tactics based on the back rank. Currently Black has two units completely out of play and Deep Rybka still evaluates White at nearly +1.} Be6 $2 {"If 25... e4 26 Bxe4 Rxe4 27 Qxe4 Qxg5 28 Qe8+ mates or here 27...Bf5 28 Qxb7." Deep Rybka says that in the latter line White is about +1. The opposite coloured-Bishops are great for the attacking side but White needs to careful too. However, this could well be Black's best line. Alternatively 25...Bf5 can be met by 26 Bxf5 gxf5 27 Bf6! and Black's King will feel the draught, while 25...Kh7?? goes down to 26 Qxh5+, and 25...Re6 could be met by 26 Bc4 but 26 Qf2 is even better. If 25...Kg7 26 Bf6+ followed by 27 Qg5 and Deep Rybka says it's mate in six. Finally 25...Qb6 26 Rf6 Re6 27 Bc4 pins and wins. All these lines other than the first one mean that g6 is ultimately indefensible. Hence Sean lets the Pawn go but at least he develops his Bishop and connects his Rooks. But he could have done this at move 16!} 26. Bxg6 { Suddenly the evaluations go through the roof for White. As Jerry foresaw a long time ago with g6 gone h5 collapses and the black king is at the mercy of four White pieces. It's amazing that Sean makes the time control!} Rf8 27. Re1 $6 {My first preference would have been 27 Bf6 so as to play 28 Qg5 or 28 Qxh5. However, Deep Rybka says 27 Bh7+!! is mate in sixteen moves. Once you appreciate that the h5 Pawn, Black's Bishop and his King's rook all go with check then this line is easier to calculate, but chess engines are great at mating with Queen and Rook in wide open spaces. Humans do take some time to calculate and actually move the pieces. We did actually find 27 Bh7+ in the post-mortem but we probably didn't calculate sixteen moves deep!} Qb5 $2 {The problem with 27 Re1 is that it tied the White Queen down to the protection of the Rook, although Qg3 has certain threats. The text releases that tie and although it does threaten mate on f1 this is easy to defend against. I prefer 27...Rac8 getting the laggard Rook in play.} 28. Bd3 {White still has tricks with 28 Bh7+! but the back rank will have to be sorted out sooner or later. Jerry still has a clear advantage due to Black's exposed King and the a8 Rook doesn't yet feature. However, Deep Rybka's assessment is going up and down like a yo-yo!} Qe8 {"Some spectators thought Black had touched his Bishop here, but clearly he would have resigned had he done so." White's Queen covers c4 as well.} 29. h3 {So no more back rank tricks.} Qf7 30. Bd2 {"30 Rxe5 is also good." White threatens 31 Rxe6 Qe6 32 Bc4 and the natural defence of 30... Rae8 loses a lot of material after 31 Bh6. The text looks like Jerry intends 31 Bc3 and 32 Bxe5. Three moves nearer the time control and Black has nothing constructive in reply.} Bf5 31. Rf1 Qg6 32. Rf3 {Not 32.Rxf5? Rxf5 33.g4 hxg4 34.hxg4 Qf6! is embarrassing.} Kh7 $2 {"Jerry gives 32...Bxd3 33 Rg3 [33 Rxd3 Kh7 34 Rg3 Rf1+ 35 Kh2 Qf7 is evaluated as equal by Deep Rybka. TM] 33...Bxc2 34 Rxg6+ Bxg6 35 Qg5 Kf7 36 Qxe5 Rae8 37 Qd5+ Kg7 38 Qxb7+ Rf7 39 Bc3+ Kh7 40 Qxa6 winning." I can't improve on this and it probably is winning. White will have to combine the advance of the b-Pawn with Kingside threats while watching out for counterplay against g2. That might have been up to the adjudicator though.} 33. Rg3 {Now "33 Rxf5 Rxf5 34 g4 also wins."} Qf7 34. Rg5 {Deep Rybka says 34 Qg5 is mate in seven. It could be right. The text isn't bad either.} Bxd3 35. Rxh5+ Kg8 36. cxd3 $2 {"White missed 36 Rh8+ mating, but, as his position is untenable, sportingly Black resigned instead of letting White chase his King all over the board." White's Rg5 will catch a Queen shortly although there is also 36...Qf6 37 Qg4+ Kf7 38 Rh7+ Ke8 39 Qd7 mate.} 1-0
The games viewer used here was developed by
http://chesstempo.com
and is licensed under
CC BY 2.0.
