Exchange sacrifices galore

Last weekend, I played the BayAreaChess Thanksgiving tournament. My play was pretty bad, but my results (2 wins, 1 draw, 2 losses against all players rated a bit higher than I) were okay, and I hit a new rating high of 2050 after the tournament. This game, the most exciting but definitely not the best-played, features a total of three exchange sacrifices, followed by a sharp endgame culminating in a pawn race. All in all, it was an amusing if inaccurate game.

Simon Rubinstein-Salzedo (2037) vs. Julian Lin (2106)

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8. f3
Be7 9. Qd2 O-O 10. O-O-O Qc7 11. Kb1

Unfortunately, I don’t know my lines in the Najdorf as well as I should. The main move is 11. g4, but I was under the impression that white is supposed to wait until black plays Nbd7 to play g4. (At least, that was my rough heuristic.) But my move is okay as well.

11…Nbd7 12. Qf2

But I still didn’t play g4 immediately, because I wanted to stop black from playing Nb6 first.

12…Rfc8 13. g4 Qd8

Black is now planning to follow up with Rxc3. So, at this point I had to decide how to evaluate the position after the exchange sacrifice. I thought that if I could isolate some of black’s pieces on the kingside and out of trouble, then black would fail to get full compensation. So, I was willing to encourage the sacrifice.

14. g5 Nh5 15. Rg1



Exchange sacrifice #1. I was pretty optimistic here, because the knight on h5 isn’t actively helping in the attack for the time being, and the e7 bishop also isn’t currently terribly useful.

16. bxc3 Qc7 17. Qd2

An interesting idea is to clarify matters on the kingside and give up the c-pawn with 17. Qh4. My idea was to give up the dark-squared bishop when the knight goes to b6 (since allowing it to get to a4 is unpleasant) and not worry too much when the other knight goes to f4.

17…Nb6 18. Bxb6

Unfortunately, the e3 bishop was doing double duty: stopping one knight from getting to b6, and the other from getting to f4. But I thought that allowing the knight to get to a4 would lead to very bad things for me.

18…Qxb6 19. Bd3 a5 20. Ka1 a4 21. Nc1 Qc5


This was a crafty move, and it took me a little while to understand what was going on: black threatens to play 22…Qa3.

22. Ne2

The idea is to meet 22…Qa3 with 23. c4 b5 24. Qc3


Another nice move, bringing the bishop into the game.

23. Rb1?

After 23. c4, black is a bit better, but white is in better shape than in the game.



24. Rb4

Exchange sacrifice #2. I was originally planning to play 24. Rb5, but after 24…Qa3, white is simply lost. For example, 25. Rb2 Nf4, and c3 falls. So, this exchange sacrifice was forced.

24…Bxb4 25. cxb4 Qf2 26. Qe1 Qxf3

That was the best I could do: go down a pawn, but my structure is somewhat improved, and at least for the time being, black’s attack has been slowed. Objectively, though, I think my position is lost. So, the goal is to do the best I can setting up a fortress and not making any new weaknesses.

27. h4 Qe3 28. Qc1 Qb6

I was mildly surprised by this decision. Okay, black has some attacking chances still, but I expected that my opponent would prefer an endgame with a clear extra pawn. And, to be honest, I was hoping that he would trade queens, given that I have worked quite hard on endgames recently and thought I might hold, even if I didn’t really deserve to.

29. a3 Rc8 30. Rf1 g6 31. Qd2 Bg4!


Good idea: my bishop is relatively useless, being blocked in by my own pawns. (Capablanca’s rule: when you have a bishop, you should put your pawns on the opposite colored squares.)

32. Rd1 Bxe2 33. Bxe2 Nf4? 

This is what I expected, and it’s the most logical move, but 33…Ng3 simply wins another pawn. After this inaccuracy, I can put up some serious resistance.

34. Bg4! Rd8 35. c4

I had been counting on the strength of this pawn structure for some time. It’s no longer easy for black to break through, since d5 never works, b5 probably doesn’t work, a4 might be weak, and white might be able to play c5 in certain circumstances. At this point, I thought I had reasonable chances to hold.

35…Qc6 36. Qc2 Ra8 37. Rc1 b6

This attempt to get the rook back to the c-file was logical enough, but it’s not clear what the point is: I just keep my queen and rook on the c-file and shuffle my king, and how is black going to do anything?

38. Kb1 Kg7 39. Ka1 Ra7 40. Kb2 Ra8

So, he figured out that that idea wasn’t going to work. But what is happening next?

41. Ka1 Rh8


Yeah, that seems like a better plan. For some time, I had been thinking about playing h5 gxh5 Bf5, but now is my last chance to do it.

42. h5 gxh5 43. Bf5 h6 44. gxh6+ Kxh6? 45. Qh2?

A better move is 45. Rg2, trapping the king on the h-file. The idea is that, while there are several things black wants to do, it’s hard to do all of them: if the queen goes to a8 to prepare Rg8 and a rook trade, then I have Qd2, hitting the d-pawn. If the queen stays on c6, then I can play Qh2 and Qg3 (sacrificing the c-pawn), setting up threats on the g-file.

45…Rg8 46. Qh4 Rg5 47. Qh2 Rg2 48. Qh4 Rg5 49. Qh2


I’m not going to complain about a draw here. But he won’t have any of that.


Exchange sacrifice #3. This is not the best way to continue though: 49…Rg2 50. Qh4 Kg7 51. Rc3 Ne2 is completely winning for black. His move might be winning too, but it gives me more chances. One point to notice is that I’m not seriously threatening anything: he could just take some time to put his king in a safer spot before continuing, and I wouldn’t be able to do anything in the meantime.

50. exf5 Qf3 51. Qa2 Qe3 52. Rd1 Qc3+?

It’s better to win the f-pawn first with 52…Kg5 and then push the h-pawn. At the very least, it buys time for a pawn race, which, as we shall see, is very important later in the game.

53. Kb1 Nd3?


It’s probably still a draw with best play, but now white has most of the winning chances.

54. Qd2+ Qxd2 55. Rxd2 e4 56. Re2 Kg5 57. Rxe4 Kxf5 58. Rd4 Ne5

I didn’t know what was going on here during the game, but I knew what I had to do: activate my queenside pawns before black’s kingside pawns get too dangerous. First plan of action: defend my c-pawn with my king so that I can win the d- and b-pawns, and try to do so quickly enough that I can still stop the black pawns on the kingside. But will it work?

59. Kc2 Kg5 60. Kc3


So far, we have stayed in drawing territory. But now black has a decision to make: which pawn to push? Is it better to have two relatively advanced pawns, or one very advanced pawn?


It’s sometimes hard to figure out what to do when you have two passed pawns, but here’s a pretty good heuristic: bet everything on the one that’s the most advanced. This heuristic would lead to the correct move in the case: 60…h4, which draws. For example 61. Rxd6 h3 62. Rd1 Kf4 63. Kd4 h2 64. Kd5 Ng4 65. Rh1 Ne3+ 66. Kc6 Kg3 67. c5 bxc5 68. b5 Nc4 69. Kxc5 Ne5, and black can stop the b-pawn with the knight, so the game should end in a draw.

61. Rxd6 h4 62. Kd4?

In turn, I’m not sure what I missed, but 62. Rxb6 h3 63. Rd6 h2 64. Rd1 is pretty clearly winning for white. I guess the pressure of a long game was taking a toll on both of us by this point.


Better is 62…Nf3+!, because there are too many forks lurking.

63. c5!?

This is probably winning, and it’s very exciting to go into a pawn race and get some new queens on the board, but 63. Rxb6 h3 64. Re6 is a much easier and less stressful way to win.

63…h3 64. cxb6

Not 64. c6?? h2 65. c7 h1Q 66. c8Q Qd1+ 67. Kc4 Qc1+, winning the queen.

64…h2 65. b7 h1=Q 66. b8=Q


I evaluated this as probably winning, but I’m not sure, and anyway with limited time remaining, I didn’t know what would happen. After the expected 66…Qc1, it’s still a tough game. Fortunately for me, though, he quickly went astray.


Maybe this isn’t a horrible move objectively, but it puts the queen in a vulnerable spot, and that’s not a smart thing to do in a complicated position after 5 hours with little time remaining.

67. Kc5 Ne5??

But this loses instantly.

68. Qd8+ Kg4 69. Rd4 1-0

My opponent deserves a lot of credit for extracting a lot from an exchange sacrifice that I didn’t initially believe in (Rxc3). I knew that in the Dragon, Rxc3 always works, but I was more skeptical about its strength in the Najdorf. I guess now I know better to allow it in the future. I also need to know my lines in the Najdorf better if I’m going to continue to play open Sicilians: it’s simply irresponsible to enter into such sharp lines without being well-prepared. (In fact, all three of my white games in this tournament featured the Najdorf. It was only in the third game that I responded relatively well.) Probably I would see the most benefit in my chess ability if I were to work on my openings and learn some systems properly. But how do I do that in the most effective way possible? I could memorize a bunch of lines blindly à la Moonwalking with Einstein using memory palaces or something similar, but I have fundamental objections to using such techniques, since they cause people to remember things for all the same reasons. Anyway, I’m constantly telling my students not to memorize things. Alternatively, I could play through a bunch of games and hope that the lines stick, but usually they don’t, as there is so much to learn. So, how do you learn openings?

About Simon

Hi. I'm Simon Rubinstein-Salzedo. I'm a mathematics postdoc at Dartmouth College. I'm also a musician; I play piano and cello, and I also sometimes compose music and study musicology. I also like to play chess and write calligraphy. This blog is a catalogue of some of my thoughts. I write them down so that I understand them better. But sometimes other people find them interesting as well, so I happily share them with my small corner of the world.
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s