Doing something is always better than doing nothing, right? Don’t our chess teachers tell us always to have a plan?
At Mike Splane’s latest chess party, we looked at a game that will have you seriously questioning that wisdom. The game was Miles-Huebner, Wijk aan Zee 1984. Mike thought that this would be a good game for us to analyze at the party because there are essentially no tactics: the game is completely played at a strategic level. What can we learn from such a game?
Here’s a brief synopsis of the game. Miles got a slight advantage out of the opening due to the two bishops and his flexible pawn duo at e4 and d4. He proceeds to do nothing for 20 moves. He shuffles his king from g1 to g2 to g1 to g2. He dances in a circle with his light-squared bishop, going from e4 to f3 to g2 to f1 to c4.
Meanwhile, Huebner rearranges his pieces and puts every one of them on its ideal square. At the end of this, Miles makes one pawn move and Huebner’s position collapses. Presumably, after the game, Huebner asks, “What the **** just happened here?”
Well, let’s see.
Anthony Miles — Robert Huebner, Wijk aan Zee 1984
1. d4 Nf6 2. c4 c6 3. Nf3 d5 4. Qc2 dc 5. Qxc4 Bf5 6. g3 Nbd7 7. Bg2 e6 8. O-O Be7 9. Nc3 O-O
FEN: r2q1rk1/pp1nbppp/2p1pn2/5b2/2QP4/2N2NP1/PP2PPBP/R1B2RK1 w – – 0 10
Position after 9. … O-O. White to move.
Surprisingly, even in this apparently harmless position White has to be a little bit alert. Mike wanted to play 10. Rd1 here, but it loses material. After 10. … Bc2! White’s queen is trapped! If 11. Rd2 Nb6. So White would have to give up the exchange.
Amusingly, according to ChessBase more people have played 10. Rd1? in this position (played 48 times) than Tony Miles’ move (played 44 times):
10. Bf4.
The most popular move, if you’re wondering, is 10. Re1, threatening e4. Black usually prevents this with 10. … Ne4. After Miles’ slightly unusual move, which does not threaten e4, Huebner sticks to the tried and true:
10. … Ne4.
Because this move was not required yet, Rybka suggests 10. … Qb6 first. White’s next move shows that he really does want to play Rd1, not Re1.
11. a4.
The point is that 11. Rd1?! would still be weak because of 11. … Nxc3! Then 12. bc Bc2! would once again cost White some material. And 12. Qxc3 Be4 would just be barren equality. So Miles plays a4 to give his queen a flight square. However, there is an objection to this move: it weakens b4, which becomes a perfect outpost for a Black minor piece. But which one?
11. … a5 12. Rfd1.
Mission accomplished for White. Now what should Black do?
FEN: r2q1rk1/1p1nbppp/2p1p3/p4b2/P1QPnB2/2N2NP1/1P2PPBP/R2R2K1 b – – 0 12
Position after 12. Rfd1. Black to move.
This was a position that Mike really wanted to dissect, because he wanted to know: “What is Black’s plan in such positions?” Unfortunately, I think we completely failed to answer his question at the chess party.
What happened was that about four candidate moves got suggested, each with its own strong advocates. We proceeded to talk about those four moves for more than half an hour, without moving a piece or looking at any lines. There was a total lack of communication. The people who wanted to play … Nb6 didn’t listen to the people who wanted to play … Nd6. The people who wanted to play … Qb6 didn’t listen to the people who wanted to play … Nb6. The people who wanted to play … g5 (that was me) didn’t listen to the people who wanted to play … Qb6. And so it went, in a circle, getting nowhere. Finally, out of complete desperation, Gjon Feinstein (who wanted to play … Nb6) went to the board and managed to drag us through some analysis of his move, but without really convincing anybody that it was best. I’m afraid I’ve never seen a more wasted half hour at one of Mike’s parties.
Just to give the computer a chance, I put this position on Rybka and let it go for 25 minutes, almost as long as we spent at the party. After that time, its top choice is 12. … Nd6 (+0.10 pawns for White). Its second choice is 12. … g5 (My move! Yay!) (+0.19 pawns). The other moves we talked about finished sixth (12. … Nb6, +0.26 pawns) and seventh (12. … Qb6, +0.27 pawns).
I think the bottom line is that none of these moves is a mistake. They are just choices for how you want to play the game. The move Huebner chose was 12. … Qb6, which was a mistake according to the book that Mike was reading, but in my opinion no more of a mistake than anything else.
If you’re curious, the reason I wanted to play 12. … g5 was that otherwise I saw Black’s position potentially drifting into passivity, and I wanted to take advantage of the space-gaining move … Ne4 by gaining some more space. Ultimately, if I were playing Black, I would have dreams of playing … f5 and maybe working up a kingside attack. Also, 12. … g5 would have preserved the two bishops for Black, which the move in the game does not.
Anyway, after 12. … Qb6 the game continued 13. Nh4! Bxh4 14. Nxe4 Bxe4 15. Bxe4 Be7 16. Bf3 Nf6 17. e4 (diagram).
FEN: r4rk1/1p2bppp/1qp1pn2/p7/P1QPPB2/5BP1/1P3P1P/R2R2K1 b – – 0 17
Position after 17. e4. Black to move.
So what should we make of this position? Well, White definitely has a slight advantage, because he has the two bishops, which are eyeing both sides of the board, and he has an advantage in space because of the pawn duo at e4 and d4. The duo is not easy to challenge with either … c5 or … e5. This is exactly the sort of “drifting into passivity” position that I was afraid of five moves earlier.
I even made a somewhat facetious remark to Mike: “If you wanted to learn how to play such positions for Black, the answer is that you should avoid them.”
However, Craig Mar was also at the party and did not share my dim view of Black’s position. He said that, knowing of my impatience for playing a waiting game, if he were playing against me he would try to get me into a position like Black’s. But in reality the situation is not so bad for Black. What Huebner undoubtedly did here was ask, “Where should I place my pieces?” Well, obviously the rooks should be doubled on the d-file, where they can focus on the most sensitive target in White’s position, the d4-pawn. It would be nice also to attack d4 with the queen and bishop, say the queen on f6 and the bishop on b6. Finally, where does the knight go? Well, White has done Black a great favor by playing a4, because this gives the knight a beautiful square on b4. The knight can never be driven away from there, and it eyes the very important square d5. That does two things: It discourages White from advancing d4-d5, because of massive exchanges on d5 leading to a draw. It also discourages White from advancing e4-e5, because Black’s knight would then gain another ideal outpost on d5.
So over the next 20 moves we’ll see Huebner head toward that arrangement of pieces. Notice that he does not worry about playing a pawn break with … c5 or … e5, reasoning that the correct pawn break will present itself after the pieces have been put in the best places.
That, my friend, is how you play this position for Black.
What about White? Well, his ideal setup is not so clear, and Miles seemingly stumbles towards it by accident. But the fact that Miles is still in control of the position, even after “wasting” several tempi, is actually part of the point. One reason that he has the advantage after move 17 is that he has the luxury of time. He does not have to commit himself yet. The two bishops will always be an advantage. The pawns at e4 and d4 will always give him more elbow room. In fact, it is absolutely imperative for him to “do nothing” in the center, because “doing something” would mean playing d5 or e5, and both of those moves are bad.
The only place where White can “do something” is on the kingside. It takes a while, but he does eventually do something there, as we will see.
The game continued: 17. … Rad8 18. Be3 Qc7 19. Rac1 Qd7 20. Rc3 Bb4 21. Rcd3 Qe7 22. Kg2 Rd7 23. b3 Rfd8 24. Bg5 h6 25. Bc1 Ba3 26. Be3 Bb4 27. h4 Ne8 28. Qc2 Nc7 29. Qe2 Na6 30. Kg1 Bd6 31. Bg2 Nb4 32. R3d2 Bc7 33. Qg4 Kf8 34. Bf1 Bb6 35. Bc4 Qf6 36. Kg2 (diagram).
FEN: 3r1k2/1p1r1pp1/1bp1pq1p/p7/PnBPP1QP/1P2B1P1/3R1PK1/3R4 b – – 0 36
Position after 36. Kg2. Black to move.
This position is a turning point, and I wish we had spent more time on this position in the party than we did on move 12.
First, let me point out that White’s “doing nothing” moves were not quite as pointless as they may have seemed. 24. Bg5 provoked … h6, which as we’ll see was exceedingly important. The first 22. Kg2 served no purpose that I can see. However, 30. Kg1 was a good move, freeing g2 for the maneuver Bg2-f1. That maneuver, which seems so unthreatening, had a purpose too: to discourage … c5 (which can be met in many cases by Bb5) and to bring the bishop to c4, where it eyes the e6 pawn. In fact, we can imagine Miles going through the same sort of thought process on move 30 that Huebner did on move 17: Where do my pieces belong? The rooks and dark-squared bishop are clearly in good places, but the queen and light-squared bishop were doing nothing. 33. Qg4 was a really nice move, the sort of move that grates on your opponent like sandpaper. And the second 36. Kg2 had a point, too. If Miles was already thinking about Qh5 and g4 as a possible plan, he needed to play Kg2 first, to cover the f3 square. Notice how GM’s keep all the weak points in their position guarded.
However, I do not want to over-praise Miles. The computer actually thinks he has frittered away his advantage (which was never great to begin with). Here, Rybka says it is time for Huebner to “do nothing” and just mark time with … Bb6-c7-b6 and/or … Qf6-e7-f6. That puts the onus on White to “do something,” and in Rybka’s estimation there is nothing he can do. So Rybka gives the position a big fat 0.00, dead equal.
Alas for Huebner, he was too determined to “do something.” After focusing all of his pressure on d4, he was undoubtedly dismayed to see that he still can’t take the pawn because of 36. … Bxd4? 37. e5! — a neat tactical stroke that takes advantage of the pin on the d-file. Huebner naturally thinks that if he defends the d7 rook one more time, White will finally be forced to do something about the attack on d4. So Huebner played
36. … Ke7?
The losing move! The king is now in no man’s land: too close to the center for comfort, too far from the kingside to defend the weaknesses there. As I mentioned earlier, the pawn on h6 turns out to be the Achilles heel. Miles now plays the impressive move:
37. Qh5! …
Once again the pawn is taboo: 37. … Bxd4? 38. Bxd4 Rxd4 39. Qc5+! (even better than taking on a5) d6 40. e5, winning the rook.
Huebner realizes his king has wandered too far, and races back to safety with 37. … Kf8. But it’s too late! Now Miles playes 38. g4! One little pawn move and it’s all over. The threat is g4-g5, and if Black takes on g5 then Bxg5 skewers the queen and rook. Huebner therefore removed his rook from the booby-trapped square d8, with 38. … Rc8, but 39. g5 hg 40. Bxg5 was still decisive. White threatens mate on h8 and the queen on f6. Perhaps Huebner was under the impression that the move 40. … g6 would save him, but Miles then played the coup de grace, 41. Qh7! and Huebner resigned, because if 41. … Qg7 42. Bh6 pins and wins the queen.
Possibly the suddenness of Black’s collapse had something to do with time pressure, but in the end, “do-nothing” chess prevailed. That’s partly because of the nature of the position, which gave White time to play that way, and it’s also because Miles’s “do-nothing” moves actually did some subtle things, such as inducing … h6, freeing the queen to go to g4 and h5, and defending the f3 square so that the g-pawn could advance.
A very fine game by both Miles and Huebner, but especially by Miles. To me, such a game is like a painting, where the individual brush strokes and the order they are placed are not as important as the overall picture. As a chess player I tend to focus on individual moves, but it is clear that both players in this game thought in terms of the “big picture.” It is difficult to reach that level of mastery, but something we should all strive for.
{ 6 comments… read them below or add one }
Interesting topic. I always have trouble in positions where the best plan is to sit tight. I have, however, picked up a thing or two from watching my top ex-students draw Grandmasters, sometimes in their sleep.
Anyways, I was not impressed with Black’s position at move 37. He simply has too many weaknesses and too few reasonable moves. You suggested Qf6-e7-f6 or Bb6-c7-b6 as solid moves to hold the position. I thought White must have something. So I cheated and opened my pet engine Komodo.
Very quickly, my computer turns up its nose on 37… Bc7. The plan is remarkably similar to the game. 38.Qh5 Kg8 39.g4 Re8 40.g5 hxg5 41.hxg5 Qg6 (forced) 42.Qxg6 fxg6 and Black’s pawn structure is a mess. Since Black can’t stop g4-g5, I don’t see many alternatives. For example, 39… Qe7 to avoid a queen trade only begs for an eventual Qh8 checkmate!
More interesting is 37… Qe7. Here Komodo sees something shocking. It suggests Kh1, Kh2 or Kg1, all with identical evaluations. Alright, the computer wants to pass. And then my jaw dropped. After 38.Kh1, the computer thinks Black is in zugzwang! OK, he’s not losing material right away, but 38… Qf6 is at +0.34 and the next best moves Ra8 and Rc8 are +0.54. White continues logically with 39.f4. If Black retreats 39… Qe7, then 40.f5 looks scary even to the naked eye. Alternatively, 39… Kg8 runs into 40.Rf1 and (you guessed it) 41.f5. Likewise 39… Re8 begs for the same 40.Rf1 and 41.f5.
After filling up my hash table with tons of deep variations (occasionally leading the engine with my intuition), I returned to move 37 (after 36… Qe7 37.Kh1). Eventually the computer settles on this completely ridiculous line: 37… Qe8 38.Kg1 Qe7 39.Kg2 Qe8 (if Qf6 then f4 again) 40.Qf3 Qe7 41.g4. Once again, Black is on the ropes (+0.68 if you’re counting at home).
Obviously, I was not expecting a gaggle of chess masters to discover these insane variations over crackers, cheese and wine. However, computers find all kinds of possibilities (for both sides) that are hidden to the naked eye. Therefore, I must find it highly unlikely that, in a middlegame with all the pieces on the board, that moving back and forth is a viable plan against a silicon opponent.
And no, I do not claim to have discovered the final word on the position at move 37.
Well, Komodo is obviously way beyond my poor copy of Rybka 3. It’s somewhat reassuring to hear that White does have a computer-approved advantage on move 37. That was my initial feeling back on move 17 — that the best way for Black to play such a position was to avoid it.
But if you can’t avoid it, I think that the three options are: break with … c5, break with … e5, or just sit. The third one is what Huebner did, which I would never have the patience for, and it didn’t work after all. Probably he should have played one of the pawn breaks — but I don’t know when.
“This was a position that Mike really wanted to dissect, because he wanted to know: ‘What is Black’s plan in such positions?'”
This was not the position I wanted to discuss. I wanted to go about six moves deeper and then break into small groups, with each group charged with coming up with three active plans for Black. I wanted to hear how each group came up with their plans, what principles they used.
For example, your group may have decided that the only danger to Black is a kingside pawn storm by White, so Black’s plan should be to freeze White’s pawn structure on the kingside with Kh8, Rg8 and g5 and g4
Another group might have decided that breaking up Whites d4 e4 pawn duo was the right idea, then decided on a plan to eliminate the e4 pawn by leaving the rook on f8 and playing h6, Nh7 Ng5 and f5.
Another group might have come up with a plan to take advantage of White’s pawn weaknesses on b2, a4 and d4 by playing Ra8-a6-b6-b4. Another group might have suggested trying to relieve Blacks somewhat cramped position by exchanging a set of minor pieces with h6 Nh7 and either bishop or knight to g5. It was the plan, the concept, that comes first, then the moves follow.
“I’m afraid I’ve never seen a more wasted half hour at one of Mike’s parties.”
I fully agree that this period was not fruitful. This position is not one that can be solved by analysis of variations. It needs to be solved by planning and assessment. But nobody was talking about plans. I let the partygoers take over the discussion instead of cutting it off and moving on to the position and thinking processes that I wanted to focus on.
I learned what I already suspected, it is a complete waste of time to try to solve some positions by calculation of variations. And I learned not to let the partygoers take over the discussion when I need to be cutting it off.
Thanks for your feedback Dana. It was quite useful.
Mike
1. Is 36…Rxd4 beyond the pale here? That would have been fun to chew on at the chess party, from what I can see looking at it for a few minutes. (I’ll follow up with more of my ideas here, if discussion ensues.)
2. What’s with this worship of high-class abacus evaluations of positions in which their evaluations are in hundredths of a pawn, because there is nothing to ‘win’ within their ever-lengthening but still limited heavily-pruned full-width search capabilities?
Hi Hal, Mike and Gjon and I looked at your idea yesterday, and unfortunately it has a hole in it. After 36. … Rxd4 White can play exactly the same move: 37. e5! The key thing to realize is that 37. … Rxg4 38. ef Rxd2 39. Rxd2 is much worse for Black than it looks, because of the back rank mate threat. (If 39. … Bxe3? 40. Rd8 mate.) You’ll see very quickly that Black in fact has to lose the bishop; for example, 39. … Bc7 40. Rd7. There are other variations, but Black is losing material in all of them. The “high-class abacus” evaluates the position as +3 pawns for White, and even an abacus is unlikely to be wrong in such a case. 😎
Point taken – thanks! Another demonstration of the wisdom of Kg2, so in this case 37…Rxd2 doesn’t work.