50 Years of Chess: Year 48

by admin on September 13, 2021

The last five years of my retrospective are kind of hard to write because (a) they are so recent that anybody who has been following this blog has seen the games before; and (b) I also haven’t been playing much. One tournament in 2016, none in 2017, two in 2018, two in 2019, and one in 2020. Only six tournaments in five years!

Nevertheless, the two tournaments I played in 2019 were both memorable and fun, in different ways. In January I played in the Bay Area International, which was held for the first time in five years. It was a great chance to test myself against really tough opposition. I did okay but not great, with one win, three losses and four draws.

Ten months later I played in the Kolty Chess Club Championship. This was my first time playing at “the Kolty,” but I already knew a lot of the players there from going to Mike Splane’s chess parties. It was a seven-round tournament, one day a week. I like that kind of tournament; it gives the drama some time to build.

Going into the tournament, I wasn’t expecting to win it. I expected to be rusty, and also I had to take a half-point bye in round one because of a schedule conflict, so I would be starting out half a step behind the competition. I was the fourth seed, behind Mike Splane (#1), a five-time champion; Vlad Getselevich (#2), a somewhat unknown quantity; and Juande Perea (#3), a three-time champion. I expected that I would have to play at least one and perhaps two of those people.

But the tournament worked out in a completely unexpected way. Eric Steger, a class-A player who started out as the #13 seed, decided to play the tournament of his life. In three consecutive rounds he beat the three top seeds, and both Getselevich and Perea withdrew after their losses. Mike Splane kept playing after his loss to Steger but was not in his top form, perhaps because of the medical issues that had already started and that eventually took his life this year.

So with the highest-rated players out of contention, the big question was whether anybody could catch up to Steger. After five rounds he was a perfect 5-0. I gave up a draw in round four in addition to my half-point bye in round one, so I had a record of 4-1. In round six we were paired against each other, and it was basically win or go home for me. With a win or a draw Eric would eliminate me from contention for first place.

I’m showing you the game, so you can probably guess who won. It’s a game that I’m very proud of, because I won a must-win game against a player who had just beaten three masters in a row, and also because it was not easy. It’s a great example of how to stir up trouble in a position that has not gone very well for you.

The game is also interesting (and I mentioned this in my blog last time I wrote about it) because I was able to successfully use a “scouting report” that Mike Splane gave me before the game. He told me that he had never seen Steger play a sacrifice. That was really interesting because if he never plays sacrifices himself, he is unlikely to anticipate sacrifices from his opponent. On move 20 an opportunity presented itself to play a complicated, speculative pawn sacrifice. I really couldn’t determine whether it was sound or not, but I played it anyway, in part because I thought he would be more uncomfortable in the resulting position than I was. Sure enough, he accepted the sac but then he almost immediately went wrong. The win after that was pretty easy.

Even after looking at it on the computer, I still don’t know if my sacrifice was “objectively correct.” But from the practical point of view — and especially in view of my opponent’s psychology — it was exactly the right thing to do.

Dana Mackenzie — Eric Steger

1. e4 d6 2. f4 Nf6 3. d3 g6 4. Be2 c5

Transposing from a Pirc into a Grand Prix Sicilian. I play this system a lot, and it’s convenient that I can learn one system for two different openings.

5. c3 Nc6 6. Na3 Rb8 7. Bd2 …

I have to say something about this move, too. For a long time after I started playing the Grand Prix Sicilian, I would develop my bishop on e3, thinking that I was protecting the weak dark squares in my position. Eventually my chess friend and fellow master, Gjon Feinstein, asked me why I kept putting my bishop there. That’s the first time it occurred to me that putting a piece on a weak square (e3, in this case) does not make it a strong square! It only makes that piece a target.

By not putting the bishop on e3, I also leave that square free for my knight later on.

7. … b5 8. Nc2 Bg7 9. Ne3 e6 10. Bf3?! …

This is taking eccentricity to an extreme. More natural is 10. Nf3. By playing all those moves with my knight (b1-a3-c2-e3) and now with my bishop (f1-e2-f3) I’m getting behind in development.

10. … Qb6 11. Ne2 a5 12. O-O O-O 13. Qe1 Ba6!

I like this move, getting ready to probe the weakness created by my 10th move.

14. Rd1 …

This position is so comical that it needs a diagram.

Position after 14. Rd1. Black to move.

FEN: 1r3rk1/5pbp/bqnppnp1/ppp5/4PP2/2PPNB2/PP1BN1PP/3RQRK1 b – – 0 14

All of my baroque maneuvers have gotten me into a position where all of my pieces are piled up like football players. As Mike Splane would say, my pieces are all looking at the rear ends of my other pieces. Compare this to Black’s development, where the pieces are well spaced and the bishops are on nice diagonals where they attack (or will attack) White pawns. There’s no question who has played the opening better, and it’s not me.

The one saving grace is that there is no direct contact yet between the two armies, so there is still time for White to start untangling his position. Also Black does have one problem: the “Swiss cheese pawn formation” on the kingside, h7-g6-f7-e6, which leaves the dark squares weak. I had won a similar game in round two where Black gave himself that formation and I brought my queen to h4, and I was eager to try again.

14. … Rfd8 15. Qh4 d5 16. e5 Ne8 17. Ng4 …

This hyperactive knight continues its journey across the board. For the first time, though, I have a concrete plan. Mike Splane would ask, “How am I going to win this game?” For me, the answer was, “Get my knight to f6.”

17. … Rd7!

My opponent is playing very well. He isn’t the least bit afraid of 18. Nf6+ because he would win a pawn after 18. … Nxf6 19. ef Qd8.

18. Be3 …

Possibly 18. Be1 was better. The reason 18. Nf6+ didn’t work was that my queen was undefended. 18. Be1 would take care of that problem, so that Nf6+ would become a threat.

18. … b4

Possibly 18. … h5 was better, for the reason just mentioned. If 19. Nf6+ Nxf6 20. ef Qd8 21. Bxc5 Bxf6 Black’s position is clearly better, as he controls all the attacking lines.

19. d4?! …

This is the point where I started writing about the game in my previous post, after skipping the first 18 moves. I took 25 minutes on this move, looking at more patient options like 19. Bf2 (probably best) or 19. Rd2 or 19. b3, but these moves had little appeal to me.

I finally decided that I wanted to play a move with more fighting spirit. Although objectively I think Black could have gotten the advantage, I can’t call this move an out-and-out blunder. I had the sense after 18. … b4 that my opponent was starting to take over the narrative of the position and forcing me to react to him. This move, along with my 20th and 21st moves, was a very deliberate attempt to “take back the microphone.”

19. … c4?

This move, on the other hand, is in my opinion very misguided. First, it releases the pressure on White’s center and thus allows me to pursue my kingside attacking ambitions with fewer worries. And second, even if you think that … c4 is a good move for Black, you should at least play 19. … Bxe2 first, trading off your bishop so that it doesn’t get locked behind the pawn chain. Third, he played this move much too quickly. (Perhaps, because of my 25-minute think on the previous move, he had already decided what to do?) From the strategic point of view, this was a very important moment. He should at least have considered the pawn trade on c3 and the pawn trade on d4, opening lines for his pieces.

Position after 19. … c4. White to move.

FEN: 1r2n1k1/3r1pbp/bqn1p1p1/p2pP3/1ppP1PNQ/2P1BB2/PP2N1PP/3R1RK1 w – – 0 20

The most important move of the game. I was sure that Black had made a mistake by not taking my knight, so I asked, what is a good way to punish him for that mistake?

20. f5! …

A bombshell! I’m sure that my opponent wasn’t expecting this. It’s a beautiful example of a sweeper-sealer. It sweeps open the c1-h6 diagonal so that my bishop can participate in the attack. It also sweeps open the f4 square, which my knight can use effectively. Mike Splane would say, “You lost a pawn but you gained two pieces.” Long-term, it can also sweep open the f-file for my rook. Furthermore, it seals off the f5 square, which would have been a great defensive and offensive square for Black’s knight, but now will not be available. And it also weakens the defense of the d5 pawn, which Black will now have to treat with velvet gloves. All this for the price of a mere pawn! From the strategic point of view, it’s a wonderful move.

Nevertheless, it was hard to be sure whether the move worked tactically, so it took a bit of courage for me to play it. I could just end up a pawn down.

20. … ef 21. Nf6+!? …

The computer — both Rybka in 2019, and Fritz today — thinks that White should play 21. Nh6+ Kh8 22. Qf2 (to prevent any captures on e5). Then g2-g4 is in the cards, and the knight on h6 helps put pressure on the f5 pawn.

Nevertheless, I already had decided on my answer to the Mike Splane Question: I was going to win by putting my knight on f6.

21. … Nxf6 22. ef Qd8 23. Bg5 h6

Of course, Black can’t allow the “bone-in-the-throat” pawn on f6 to remain. Steger must have been optimistic at this point; trading the h6 pawn for the f6 pawn looks like a pretty good deal for him.

24. Bxh6 …

Position after 24. Bxh6. Black to move.

FEN: 1r1q2k1/3r1pb1/b1n2PpB/p2p1p2/1ppP3Q/2P2B2/PP2N1PP/3R1RK1 b – – 0 24

The final turning point of the game. Black played the very bad move:

24. … Qxf6??

You’d have to ask Eric why he played this move. Maybe he thought I would have to trade queens, after which he would have a very comfortable pawn-up ending. Or maybe he overlooked my next move, or missed something after that move.

From the pure chess point of view, it’s too bad that he didn’t play the correct move, 24. … Bxf6, because I would be very curious to see what would have happened next. My plan was to play 25. Qh3 (diagram).

Position after 25. Qh3 (analysis). Black to move.

FEN: 1r1q2k1/3r1p2/b1n2bpB/p2p1p2/1ppP4/2P2B1Q/PP2N1PP/3R1RK1 b – – 0 25

My claim is that White has full practical compensation for the pawn. This may seem unbelievable, because the computer evaluation of the position starts out at around +0.8 to +1.0 pawns for Black, in other words the computer believes that White has hardly any compensation. But when you go deeper into those computer lines, the alleged advantage for Black tends to melt away.

I’ll skip the computer analysis, though, and try to explain this position in human terms. Why is the position so hard for Black? First, he has to choose between many possible plans. Should he ignore my attack and storm the queenside with … Rb2, possibly taking on a2 and doubling his rooks on the seventh rank? Or should he place his bets on all-out defense, putting his bishop on c8 and knight on e7? Or maybe should he try to blockade the kingside with … g5 and … f4?

Second, no matter what plan Black chooses, it is hard for him to keep all of his weak points covered. The paramount target for White is the pawn on f5, which White will attack right away with g2-g4. Another target is d5. Also, Black has several loose pieces on a6, c6, and f6. And of course there is the huge problem of the Black king, which may come under fire on either the g-file (after White trades off his g-pawn) or the h-file. That king just has no safe place to go. And finally, off in the middle distance, White has dangerous sacrificial threats like Ne2-f4xg6 or Ne2-g3xf5, which could pulverize Black’s kingside.

While Black has a lot of things to worry about, White by comparison has very few things to think about. He will play g2-g4, and try to pile pieces on the f-pawn. At some point White will have to decide whether to play his knight to g3 or to f4; that is the only real subtlety in the position.

Just for giggles, and for people who like “What might have been?” questions, here is a sample computer line where the computer initially likes Black, and then shifts strongly towards White: 25. … bc 26. bc Rb2 27. g4 Ne7. So far this seems quite reasonable for Black; the knight on e7 defends a whole lot of weak points (d5, f5, g6). The problem, though, is that Black’s pieces are poorly coordinated, with the knight in the way of the queen and the rook in the way of the bishop (which wants to go to c8). Now the computer suggests 28. Ng3 f4 (trying to keep lines closed) 29. Bxf4 Rc2 (diagram).

Position after 29. … Rc2 (analysis). White to move.

FEN: 3q2k1/3rnp2/b4bp1/p2p4/2pP1BP1/2P2BNQ/P1r4P/3R1RK1 w – – 0 30

Initially, when Fritz got to this position, it thought White had a slight advantage with 30. g5. But then, after about a minute of thought, it comes up with the unbelievable move 30. Nf5!! Holy cow! First, you have to see that if 30. … gf 31. gf White will have an irresistible attack on the g- file and h-file, and if Black’s king tries to run away with 31. … Kf8 then one of those loose bishops drops off after 32. Qh6+ Bg7 33. Qxa6. You saw that, right? Alternatively, if Black goes about his business with 30. … Rxc3, you have to see that White will blast the kingside apart with 31. Nh6+ Kf8 32. Nxf7!

Only in my dreams could I ever play this way, but it’s really inspiring to see what possibilities there were in the position. No wonder I couldn’t figure out whether I had compensation for the pawn.

Anyway, let’s now go back to reality. Back to the third diagram above. As noted earlier, Black blundered with 24. … Qxf6??, and I played

25. Bg5 Qe6

Black is kind of unlucky that the perfectly reasonable move 25. … Qd6 loses the exchange to 26. Bf4. Probably he should play that anyway, and try to draw the endgame after 26. … Qf6 27. Qxf6 Bxf6 28. Bxb8 Nxb8. The move he chose leads to a quick debacle.

26. Nf4 Qe3+ 27. Kh1 Re8?

Here, as I mentioned in my earlier post, Black has an amazing possibility of sacrificing the queen for a bishop and two extremely dangerous passed pawns: 27. … bc 28. Nxd5 Qxg5! 29. Qxg5 cb. This is the sort of thing that Black should try even if he can’t work it out completely, because every other move loses very quickly. However, remember that Mike Splane’s “scouting report” said that Steger never plays sacrifices. So it’s very unlikely that he will come up with a queen sacrifice in this situation, and indeed he didn’t.

28. Nxd5 Qe6 29. Nf6+ Black resigns

Final position.

FEN: 4r1k1/3r1pb1/b1n1qNp1/p4pB1/1ppP3Q/2P2B2/PP4PP/3R1R1K b – – 0 29

If 29. … Bxf6 30. Bxf6 White’s strategy of dominating the dark squares is complete, and there is no stopping Qh8 mate. If 29. … Kf8 White doesn’t settle for winning the exchange with 30. Nxd7+? because there is a much better move: 30. Rde1!, and Black has to give up his queen to avoid checkmate.

I like this final position because it shows you the value of the Mike Splane Question (“How will I win this game?”). Very early on, as early as move 17, I had the idea that I was going to win by getting a knight to f6. In fact, both of my knights eventually went to that square! Of course there was nothing forced about it. The game could have gone differently and who knows, maybe I would have won with Nf5 as in that crazy computer variation that I showed you. But the Mike Splane Question gives you a North Star to sail your boat towards, and it’s really amazing how often it takes you all the way to your destination.

Oh, by the way, I should mention that I also won in the last round against Paulo Santanna, the #5 seed. Eric Steger kept pace by winning his last game, so we both ended up with 6-1 scores. However, I won the title of club champion on tiebreak. I feel a little bit bad about spoiling what was probably the best tournament of Eric’s life, but I can’t feel too bad! For me it was the first victory in a multi-day, open tournament in many years. I have only won four such tournaments: the 1988 Georgia Congress (Atlanta), the 1989 Roosevelt Open (Dayton, OH), the 1993 Roosevelt Open (Dayton), and the 2019 Kolty Club Championship (San Jose). Twenty-six years is a pretty long wait!


  • Try to control the “narrative” of the game. If your opponent is starting to dictate the action, look for enterprising ways to take back the microphone, even if it costs a little material. Sometimes it’s not the smarter player who wins, it’s the more determined player.
  • Learn about sweeper-sealers. These “impossible” pawn sacs open lines for your own pieces while sealing off good squares from your opponent’s pieces.
  • To play a sacrifice, you don’t always need to analyze it all the way to a win. In fact, it may be a waste of time to attempt to. You should analyze enough to make sure your opponent doesn’t have an easy defense. It’s a big plus if your sacrifice forces him to make difficult decisions, and if you have so many different threats that it’s hard for him to stop them all.
  • Ask the Mike Splane Question: How will I win this game? But remember that it’s only a guide; be prepared to change your plan if circumstances require it.
Print Friendly, PDF & Email


50 Years of Chess: Years 45-47

by admin on September 1, 2021

As the finish line of my retrospective approaches, we’re going to put on a “finishing kick” and cover three years in one post. The reason is that I have almost no games to show you from the years 2016 to 2018. I played in only one tournament, with 5 games, in 2016; no tournaments in 2017; and two tournaments, with 13 total games, in 2018. I didn’t play a single tournament game for more than two years, from March 2016 to June 2018.

There was a reason, of course. The only other time I ever went two years without playing was when I was in graduate school (1980-1982), when I had to temporarily give up chess because I didn’t have time for it. Likewise, in 2016 and 2017 chess took a back seat to the book I was writing with Judea Pearl: The Book of Why. I have zero regrets about this, because that book was the most important thing I have done in my career as a writer.

By 2018, the book was finished, and it was published that May. It was the end of a long journey, but every step was worth it. Our book was a critical success and managed to hit Amazon’s Top 100 list, for one day. (It never managed to dent any “real” bestseller lists; that would be the equivalent of being in the Amazon Top 100 for a week.) I was especially happy when our book was reviewed in the New York Times and selected as one of the top books of 2018 by NPR’s Science Friday.

Another piece of good news came along that spring: I was offered a chance to be “journalist in residence” at the Simons Institute for the Theory of Computing in Berkeley for eight-week summer session. I learned a lot about quantum computing that summer, probably just enough to realize how much I don’t know.

The eight weeks of my internship coincided exactly with the eight weeks of the summer Tuesday Night Marathon (chess tournament) at the Mechanics Institute. I had always wanted to play in a Tuesday Night Marathon, but it had never been practical because it’s such a long drive from Santa Cruz to San Francisco. But if you’re in Berkeley, it’s an easy subway ride! So this was my perfect opportunity to kill two birds with one stone: to return to tournament chess and play in one of the country’s most historic chess venues.

Alas, the things we look forward to don’t always work out as well as we had hoped. I struggled in that tournament, going 5-3 against mostly class-A level opponents. In retrospect, it’s scarcely surprising. I had not played in two years and had my mind on other things, like book reviews and book readings and quantum computing and, oh by the way, did I mention that my nephew got married and I took a quick trip to Newport, Rhode Island, for his wedding? Yeah, that happened too.

I did have one outstanding win in the seventh round, where I sacrificed two pawns and the exchange for a winning kingside attack. However, I blogged about that game in great detail and I really have nothing to add to what I wrote before. Feel free to check that game out, especially if you’re interested in sacrificial attacks against the French Defense.

But since I don’t want to just repeat myself, I’ll have go to plan B and annotate a different game, which I played in round five. It is another wild and exciting game. Although it ends in a draw, I’m sure you will enjoy it. I did blog about it at the time, but only about one crucial position; I’ve never showed the entire game. One reason I’m selecting this one is that it also gave rise to one of the best online post-mortems I’ve ever had on this blog. Larry Smith, Mike Splane, and Juande Perea all sent really great comments, and this will give me the opportunity to add their discoveries to my own analysis.

Greg Sarafian — Dana Mackenzie, 6/27/2018

My opponent was an expert who hit his lifetime peak rating (2014) in this tournament.

1. d4 Nf6 2. Nf3 d5 3. Bg5 Ne4 4. Bf4 e6 5. c3 Nd7 6. Nbd2 Ndf6 7. h3 Be7 8. e3 O-O 9. Bd3 b6 10. O-O Bb7 11. Ne5 c5

Both sides have completed their development normally and now the action is starting to heat up, so perhaps it’s time for a diagram.

Position after 11. … c5. White to move.

FEN: r2q1rk1/pb2bppp/1p2pn2/2ppN3/3PnB2/2PBP2P/PP1N1PP1/R2Q1RK1 w – – 0 12

The two big questions in this position are: 1) If White drives my knight away with f2-f3, will it create weaknesses in his position that I can take advantage of? And 2) Will my bishop on b7 ever find any kind of purpose in life? I did not find satisfactory answers to either of these questions in the game.

12. f3 Nxd2 13. Qxd2 Nh5

Trying to create some air for my kingside pieces before White has the chance to play g2-g4.

14. Bh2 f6?!

A little bit too impatient. Another option might be 14. … Bh4, trying to put pressure on White’s weakened dark squares. After 15. g4 Nf6 the bishop looks a little bit exposed, but it’s not clear if White can actually take advantage of it. A more flexible idea is 14. … g6 15. g4 Ng7, with the threat of … f6 trapping White’s knight. Both of these variations are evaluated by the computer as completely equal. Basically, I should not be scared of White’s space advantage because my pieces are still mobile enough.

15. Ng4 …

Either overlooking or rejecting the idea of 15. Ng6! hg 16. Bxg6 winning a pawn for White. I think that Black would still have pretty good chances to hold a draw. After 16. … f5 17. Bxh5 Kh7 18. f4 (forced, to keep the bishop from being trapped) I already have the makings of a pretty good white-square blockade. Of course this means I will have to play 50 moves of patient defense, not exactly my idea of a fun game.

15. … Bd6 16. Qf2 Bxh2+ 17. Kxh2 Qc7+ 18. f4 …

Position after 18. f4. Black to move.

r4rk1/pbq3pp/1p2pp2/2pp3n/3P1PN1/2PBP2P/PP3QPK/R4R2 b – – 0 18

Now I lashed out with a very impatient move.

18. … g5?

This move creates multiple weaknesses, on h6, f6, f5, and the g-pawn itself. I was thinking more with my heart than my head here. I wanted to shake White’s position up and do anything to create counterplay. I’m sure that the ratings and the clock situation also entered my mind; I want to speed the game up and put pressure on a lower-rated player.

But is this the type of position where we should be looking for a tactical solution? In general, you should go for tactics when there are good strategic reasons for thinking they will work in your favor. Here, that is not the case. Black’s pieces are no better posted, and in fact worse posted, than White’s. The QB and QR are out of action, and the knight on h5 is kind of in no man’s land. The only thing working in Black’s favor is the b8-h2 diagonal, where Black’s queen eyes White’s king, but that is a very small foundation to build an attack on. I’m really playing hope chess here — I’ll just try to put pressure on f4 and hope something good happens.

To approach the position more strategically, Black should ask, “What is my worst piece and how can I improve it?” That’s really easy. Both sides have very strangely posted knights that have no moves at the moment. Where would I like to see my knight go to? Well, the outpost on e4 beckons. And it’s not that hard to get to. It takes only four moves: g7, e8, d6, e4. But of course, to even get started I need to play 18. … g6. I think I was wary of letting White’s knight go to h6, but the move 18. … g5, which I actually played, has the same “defect” and more.

19. Qf3 Qg7 20. g3 Kh8 21. Rg1 e5!?

Burning all the bridges behind me. This move isn’t sound, but at least there is something to be said for the fact that it’s consistent. Damn the torpedoes, and full speed ahead. Now the game gets really, really crazy.

Position after 21. … e5. White to move.

FEN: r4r1k/pb4qp/1p3p2/2ppp1pn/3P1PN1/2PBPQPP/PP5K/R5R1 w – – 0 22

22. Nxe5! Nxf4!

Both knights sacrifice themselves in desperado fashion.

23. gf fe 24. Qh5! …

A very calm and strong move by White. I was hoping for 24. de, when my bishop can finally get into the game with 24. … d4. Actually, White is probably doing well in that line too, but I totally agree with Sarafian’s decision to keep the genie (the Black bishop) in the bottle.

24. … ef 25. Rxg5 Qc7 26. Re5? …

For the first time White’s resolve starts to weaken. The problem with this move is that it turns White’s best attacker into a defender, and you have to ask whether this is really necessary. The computer, with its nerves of steel (or copper, or semiconductor), says that 26. Rag1! is completely winning. The point is that after either 26. … fe+ or 26. … f3+ White can simply play 27. Kh1. There is no need to use any of White’s attacking pieces for defense, because Black does not have enough attackers to create real threats! Plus, White has a really sneaky threat that Sarafian perhaps didn’t see. After, say, 26. Rag1 fe+ 27. Kh1 White threatens 28. Bxh7!! Qxh7 29. Qxh7+ Kxh7 30. Rh5 mate. The two-rook checkmate is not a mating pattern you see very often this early in the game. Black’s only real defense is to play 27. … Rf7, but now 28. Qg4 sets up Alekhine’s gun and threatens mate on g8, and Black is forced into full-fledged defense with 28. … Qc8 etc. The lesson here is that White’s threats are simply stronger than Black’s, so White has no need to play any defensive moves.

This is all really easy to say because I have a computer. It’s much harder to make this kind of assessment over the board, with your clock ticking. White’s move plays it safe, but it gives Black a little more time and a little more hope.

26. … fe 27. Re1? …

I have to give this move a question mark because technically this is where White loses his advantage. The computer says that 27. Rg1 is still absolutely winning for White, +3.70 pawns. The main reason is that it once again sets up the “sneaky threat” of 28. Bxh7!

But in the real world, this move follows completely logically from the previous one. If White had seen the idea of Rg1 followed by Bxh7, he would have played it last move. White did not see that threat, and instead he does the normal human thing, which is to defend and try to shut down Black’s counterplay.

I really have trouble criticizing any of Sarafian’s moves. The only thing he did wrong was that he didn’t play like a computer.

27. … cd 28. cd Rf2+ 29. Kh1 …

Position after 29. Kh1. Black to move.

FEN: r6k/pbq4p/1p6/3pR2Q/3P4/3Bp2P/PP3r2/4R2K b – – 0 29

The last couple moves for Black were more or less automatic, but now we have arrived at a crucial moment. Black’s prospects have definitely improved, but the position is exceedingly dangerous for both sides. This is the position where I started my previous blog post about this game. A key thing to realize, by the way, is that I have 17 minutes left for the rest of the game. No time control, no extra infusion of time to look forward to. I used 12 of my 17 minutes on this move — very poor clock management — and still didn’t play the best move.

In a complicated position, with little time left, I think it’s important to start with the things you know about the position. As I wrote in my earlier post, one thing that I thought I knew was that White’s rook on e5 is pinned. He moves it, he gets checkmated. But that’s not quite correct! What I overlooked is that if he moves it with check (i.e., Re8+) then he doesn’t get checkmated and in fact he takes over the initiative.

Once you realize this, it makes Black’s decision very easy. The most important thing for Black is to keep that e5 rook pinned. Therefore there are only two moves that are even worth considering: 29. … R8f8 and 29. … Rg8. Both of those moves prevent Re8+. Of the two, 29. … Rg8 is much more promising because it closes the mating net around White’s king so that he can’t escape. After this, 29. … Rfg2 becomes a real possibility, and both White rooks are stuck! One can’t leave the first rank, the other can’t leave the fifth rank.

So in thirty seconds or less, with no calculation of variations, we have already arrived at Black’s best move, 29. … Rg8. Now is where the fun stuff begins. Several readers of my previous post came up with some great tactics for both sides. It’s likely that White will take the opportunity to grab the e-pawn, 30. R1xe3. First Larry Smith suggested that I could continue 30. … Qg7, which looks extremely strong because it threatens two checkmates at once, but then he noted that 31. Rg5! is a strong response. I saw this during the game, and it’s the reason that I didn’t play 29. … Rg8. I can’t take the rook because I would get mated on h7. Larry did point out, though, that I could play 31. … Rf1+!, a neat deflection idea. After 32. Bxf1 Qxg5 33. Qxg5 Rxg5 34. Bg2 I am alive, but there is no question that White has all the winning chances in the endgame because of his active rook and his better bishop.

Then Mike Splane chimed in and suggested a different and even more amazing way to deflect White’s bishop: 30. R1xe3 Ba6! His original idea after 31. Bxa6 was to play 31. … Qc2, but then Juande Perea pointed out that an even more crushing variation for Black is 31. … Qc1+ 32. Re1 Qf4!, and mate is unstoppable! Wow!

Then Larry Smith joined in again and said that after 30. R1xe3 Ba6! White has to play 31. Qh4! “hitting the rook on f2 and eyeing f6.” That last fact becomes crucial after 31. … Rfg2, when White miraculously saves the game with 32. Qf6+ Qg7 33. Re7!! This sweet queen sac was discovered by Juande Perea. Of course, if 33. … Qxf6 then 34. Rxh7 is mate, so Black has to acquiesce to a draw with 33. … Rg1+ 34. Kh2 Rg2+. Sadly, … Qg2 mate is impossible because the queen is pinned.

What an incredible variation! And I love the old-fashioned way we found it. Nobody turned on their computers, they just looked at their boards at home and found one key idea after another in collaborative fashion.

The bottom line is that 29. … Rg8 30. R1xe3 Ba6 is “only a draw,” but White would have had to earn that draw with some sensational moves. There is still a possibility that Black could play for a win with 29. … Rg8 30. R1xe3 Rfg2 immediately. Fritz evaluates this at +1 pawn for Black.

I didn’t see any of this, and after letting my clock tick down to 5 minutes I played more “hope chess.”

29. … e2?? 30. Re8+ …

You can imagine my reaction: Oh, s***! That rook isn’t pinned?

30. … Rxe8 31. Qxe8+ Kg7 32. Rg1+ Kf6 33. Qf8+ Qf7

Position after 33. … Qf7. White to move.

FEN: 5Q2/pb3q1p/1p3k2/3p4/3P4/3B3P/PP2pr2/6RK w – – 0 34

There is one more twist in this amazing game! Three years ago I wrote, “Amazingly Black does seem to be surviving” — but I was wrong! White wins by force in this position. Do you see how?

First of all, let me say that my opponent did not find the win. In fact, I think he was hugely relieved to bail out to a draw with 34. Qd6+ Qe6 35. Qf8+ 1/2 – 1/2. Let’s not forget that a draw was a good result for him, as the lower-rated player.

Nevertheless, White is completely winning after 34. Qh8+! This little two-step, Qf8+ followed by Qh8+, is a nice example of what I call “moving your opponent’s pieces.” First he lures my queen to f7, where it is more exposed. Then after 34. Qh8+! Ke6 35. Qe5+ Kd7 my queen and king are lined up awkwardly. Of course, 36. Rg7? would be a mistake, as 36. … e1Q+ draws. But if White plays 36. Bxe2!, Black’s position collapses faster than a house of cards. White’s main threat is still to win the queen with Rg7, and there is just no good defense. For example, if 36. … Qe7 37. Rg7 Rf7 38. Bg4+ (or 38. Bb5+), Black can resign.

All in all, I think that both players were probably glad to escape with a draw!


  1. Before plunging into tactical complications, ask whether there is any strategic reason for thinking the complications will work out in your favor. Is this a position for tactics or strategy?
  2. On the other hand, once you have committed yourself to a tactical solution, don’t go halfway. Against human opponents, relentless tactical pressure will often save or win positions that are objectively “losing” (according to the computer).
  3. In a complicated position, try to simplify it by identifying the things you know about the position.
  4. In highly tactical positions, mate threats rule. In fact, having a bona fide mate threat is sometimes almost like having an extra attacker. There are so many examples in this game. In the notes to moves 26 and 27, White’s “sneaky threat” of Bxh7 Qxh7 Qxh7+ Kxh7 Rh5 mate completely changes the evaluation of the position. Black has no attack because he has to stop this threat immediately. And in the note to move 29, we saw mate threats galore: White’s mate threat of Qxh7 mate. Black’s threats of … Qh2 mate or … Qg2 mate in various variations. Finally, White’s threat of Rxh7 mate after Juande’s sensational queen sacrifice 33. Re7!!
  5. Sometimes you can win by moving your opponent’s pieces to squares where they don’t want to be.
Print Friendly, PDF & Email


50 Years of Chess — Year 44

August 22, 2021

I want to thank the people who wrote to me here or by e-mail or on Facebook after Mike Splane’s death. In this blog, I’ve been privileged to have the opportunity to steal — oops, um, er, I mean “write about” — many of his ideas. Although I’ve said it before, I’ll say it again: […]

Print Friendly, PDF & Email
Read the full article →

Mike Splane, Farewell and Legacy

August 12, 2021

Last night I received a very unexpected e-mail from Ken Case, one of the regulars at Mike Splane’s chess parties. Ken said that Mike has passed away. Mike has been in poor health for at least the last two or three years. The source of some but not all of his problems was diabetes. Mike […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Year 43

July 31, 2021

I’d like to interrupt my retrospective to mention an event that is going on in the chess world here and now: the 2021 World Cup. Principally, I’d like to congratulate Sam Shankland for advancing farther than any other American contestant, all the way to the “elite eight,” as they would call it in basketball. As […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Almost Another One for the Ages

July 27, 2021

In Year 35 of this retrospective, I wrote a post called One for the Ages, in which I showed my lifetime masterpiece, Mackenzie-Pruess. In that game I debuted a new opening variation, the Bryntse Gambit (which had been played before in correspondence chess but never, to my knowledge, in OTB chess). The Bryntse Gambit is […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Year 42

July 18, 2021

Wow! We’re moving on up to 2013 in this retrospective, and in this year I have a rich selection of interesting games to show you. But — here’s the catch — I’ve already written about almost all of them on this blog. Here are some of the “greatest hits” from that year: Zildzic-Mackenzie: I sacrificed […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Year 41

July 11, 2021

Recently I wrote a post in which I listed my best and worst tournaments by rating improvement or rating loss. But one of the highlights of my chess career was a tournament where I gained only 2 rating points! I also didn’t win any prize money or beat any USCF masters. (I did, by a […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Year 40

July 4, 2021

We’re on the home stretch now! For so long the games in this rear-view mirror have been so old that it seems almost as if they were played by a different person. But now we’re up to the year 2011, which doesn’t seem like that long ago. As I mentioned in my last post, 2011 […]

Print Friendly, PDF & Email
Read the full article →

50 Years of Chess: Year 39

June 26, 2021

In 2010, according to the USCF online database, my rating hit its lowest point ever. At the 2010 CalChess Labor Day Classic (the northern California state championship) I withdrew after four rounds with a score of 0.5-3.5, culminating in a loss against a player rated more than 200 points below me. This came after a […]

Print Friendly, PDF & Email
Read the full article →