by admin on November 13, 2020

Friend and long-time follower of this blog, Mike Splane, sent in excellent comments on my two most recent posts, which I think both warrant some more discussion.

First, commenting on my Transition post, he says: “I totally agree that you got much stronger strategically during the time that I’ve known you. Is there anything you can point to about the chess parties that was particularly helpful?” (I had written that his chess parties helped improve my strategic understanding. If I had had this resource 20 years earlier, perhaps I would have been able to get to FIDE Master or Senior Master.)

I sort of hinted at the answer already, but let me make it more explicit. For the first forty years of my chess career, 90 percent of my analysis was, “If he does this I do that, and then if he does this I do that,” etc. That is, it was highly focused on calculation. But in quiet positions, or positions where there weren’t clear variations to calculate, I would often feel at sea. This was especially true in the opening-to-early-middlegame transition, say moves 15 to 25, when the pieces have been brought out but usually the battle has not yet reached a critical point.

As I wrote in my post, “What do you do, or what do you think about, when you don’t appear to be doing much at all?”

What Mike’s parties showed me was that there are lots of things you can think about. Some of them came up in my last post: What side of the board should I be playing on? Does this trade that I’m considering improve the position of my pieces or the opponent’s pieces? When is the right time for me to relieve the pawn tension — or should I perhaps induce my opponent to relieve it? There are some other good ones we’ve talked about in his parties, such as “What is my worst piece and how can I make it better?” And one that I have mentioned over and over in this blog is the Mike Splane Question: “How am I going to win this game?” Although sometimes that question seems like wishful thinking, it’s amazing how useful it is.

I don’t want to claim originality for any of these questions, except maybe the Mike Splane Question. You can find them in other places. In fact, for readers who are wondering where to get started with strategic thinking, a very good place is Jeremy Silman’s How to Reassess Your Chess, and his theory of imbalances. That book showed me that there can be a lot going on in a position that is supposedly “equal.” A nice example was my game with Belle (Year 12), where in an “equal” position I allowed an unfavorable imbalance in pawn structure to create favorable imbalances in open lines and piece play. That was very Silman-esque thinking, 25 years before I read Silman. Alas, it’s not something I did very consistently.

As I discussed in one of my Chess Life articles, you can remember Silman’s imbalances by the acronym “IMPLODeS + K”:

I = Initiative. (Who is attacking?) M = Material. P = Pawn structure. (Perhaps I should add to this: Pawn breaks, because that is the first thing that GM Robert Hess says that he looks at. Who controls more pawn breaks?) L = Lines and squares. (Who has access to better lines and squares for their pieces? Which of my pieces have good lines and squares, and which don’t?) O = Officers. (This is a synonym for minor pieces. Who has the two bishops? Who has the better minor pieces?) De = Development. S = Space. Finally, Silman unaccountably omitted an important imbalance: K = King safety. (Whose king is more exposed to attack?)

Although Silman’s imbalances are primarily a tool for evaluating a position, they can also be used as a tool for strategic planning, because you can make a plan that tries to “steer the conversation” toward the imbalances that are favorable to you. Again, some of this is well known. If you have the two bishops, you’d like to open the board and play on both wings. But what I like about Silman’s imbalances is that they give you a framework for thinking about all sorts of positions.

Anyway, thanks to Mike’s questions and Silman’s imbalances, maybe the amount of time I spend calculating variations is down to 75 percent instead of 90 percent. Especially in those early middlegame positions where it seems as if there is nothing to calculate.

Now let’s switch gears and talk about something much more concrete: the game against Matt Noble that I analyzed in Year 13. Here is the position on Black’s move 25.

Position after 25. Qe2. Black to move.

FEN: 1r6/4npk1/3p2p1/qppP3r/4P2P/1PN5/P1K1Q3/6RR b – – 0 25

Here Mike writes: “Take a look at 25. … Re5 and tell me how White stops the dual threats of b4 and Nd5. If 26. Rg5 Nd5! 27. Nd5 Qa2+ 28. Kd3 (or Kd1) Qb3+ gives you three pawns for the piece and a raging attack, or an ending with 4 passed pawns for a knight.” By comparison, he gives the move I actually played (25. … b4) two question marks.

This is an extraordinarily difficult position to annotate, because the move that makes the most sense from the strategic point of view is not actually the objectively best move, as determined (almost surely correctly) by the computer.

I love Mike’s suggestion of 25. … Re5. It answers two questions. What is my worst piece, and how can I make it better? The answer: my knight on e7, which is not contributing at all to the attack. By moving my rook to e5, I immediately threaten … Nxd5. Second question: How can I make my own threats stronger? Obviously the move … b4 is in the air, but if I play it immediately White might just play 26. Na4, and it’s not clear that I have gained much. But after 25. … Re5, … b4 is now a killer threat.

The computer does not love Mike’s suggestion. Fritz gives White a whopping 1.6-pawn advantage after 25. … Re5 26. Qf3! This compares unfavorably to the 0.4-pawn advantage for Black after 25. … b4 26. Na4, and to the 3.0-pawn advantage for Black (!!) after 25. … b4 26. Qxh5?, which is what happened in the game.

What’s the deal? Well, I looked at 26. Qf3 before consulting the computer, but I thought that 26. … b4 27. Na4 c4 looked very good for Black. Unfortunately, my analysis (and Mike’s) was not concrete enough, because after 28. h5! White’s attack comes first. The computer analysis (best moves for both sides) goes 28. … Qa7 29. hg fg 30. Qh3. But let’s try to free ourselves from the computer and if-this-then-that thinking. In conceptual terms, the big problem is that after 28. h5 Black cannot prevent the opening of the h-file, because 28. … g5 is met by 29. h6+! Kg6 30. Qh5+ and either Qxf7+ or Rf1+, depending on what Black does. Once the h-file comes open, Black is dead meat. The subsidiary problem, which 26. Qf3 pointed out, is that the f-file and f7 pawn in particular are also weak.

So we have to reluctantly concede that 25. … Re5, while strategically well-motivated from the attacking point of view, unfortunately ignores the important role of the rook in stopping White’s h-pawn advance.

One difficult thing about opposite-side-castled positions is that they are so schizophrenic. White is busted on the queenside. Black is busted on the kingside. Everything comes down to who can throw just enough monkey wrenches into the opponent’s plans so that their attack can come first. It’s hard to evaluate such a position on general principles.

Now that brings us back to the move I played, 25. … b4. Should this get an exclam, as I gave it, or a double question mark, as Mike gave it? I have no definitive answer to this question. I’d like to hear what other readers think. But here are my thoughts.

  1. There is no question that I was excited, probably too much so, by the idea of sacrificing the exchange. I was still under the spell of the anthologies, which make you think that good chess is sacrificial chess. Remember, in the Silman approach, material is just one of eight imbalances: I, M, P, L, O, De, S, and K. There are lots of ways to win that don’t involve sacrificing M (or winning M, for that matter). If I am chronically biased in favor of moves like … b4 to the point where I’m not even considering moves like … Re5, then I’m not playing good chess.
  2. Nevertheless, there are some very good points about the exchange sacrifice. Most important, as I mentioned above, is that it steers the conversation toward the imbalances that favor me. If White takes on h5, the queen is actually in the way of White’s own attack; she has to back out again so that he can play h4-h5. In essence, I am sacrificing the exchange to gain two tempi, and in a position like this, two tempi are huge. Their significance is that now the game will be decided, not by White’s kingside attack, but by Black’s queenside attack and whether it is strong enough to justify an exchange sac. White’s kingside attack is now irrelevant. That’s a heck of an argument in favor of 25. … b4.
  3. By the way, if White doesn’t take the bait but plays 26. Na4, then Black can comfortably play 26. … Rbh8, shoring up the kingside. I couldn’t do this on the previous move because 25. … Rbh8 would have been met by 26. Qxb5 Qxb5 27. Nxb5 Rxh4 28. Rxh4 Rxh4 29. Nxd6, winning a pawn and defending e4 just in the nick of time. So from this point of view, 25. … b4 is an “improving” move just like 25. … Re5; it improves … Rbh8, a not quite satisfactory defense, to the point where it becomes satisfactory.
  4. Finally, we have to talk about a factor that the books will probably say is unimportant, but I consider to be hugely important: chess psychology. Do you know anything about your opponent that makes it possible to predict their response? In this case, I was pretty sure that Matt would answer 25. … b4 with 26. Qxh5. One of my other favorite opponents in North Carolina, Bernie Schmidt, would never trade queens voluntarily, and perhaps I will show you a game where I used that bias against him.
  5. Another chess psychology factor is not specific to the opponent, but general. When my opponent makes a move that he thinks creates a threat that must be defended, I always look for a way to not defend it. This sows doubt and mental anguish in the opponent’s mind. This case is a classic example. White played 25. Qe2 for the specific purpose of threatening my rook so that I would have to move it or defend it. By playing 25. … b4, I tell him that he is wrong. And this puts him in a difficult psychological position. Either he accepts my sacrifice and gives me a dangerous-looking attack, or else he doesn’t accept my sacrifice, and in this case the question is: What the hell did he move his queen to e2 for? If he’s not going to take the rook, the queen is no better on e2 than it was on d2. For many players, it is a bitter pill to find out that their previous move was a waste of time.

My final verdict is that for all these reasons: gaining tempi, steering the conversation, improving my position, psychology of the individual, and psychology of the human species, 25. … b4! was an excellent move. It was only not excellent if I arrived at it through a thought pattern that Mike calls “tunnel vision,” where I make up my mind what I want to do and don’t even give other ideas a chance.

There are still some things about this game that mystify me. It boggles my mind that in the key variation 29. Qa1 (instead of 29. Kxc4?) Qe2 30. bc Qe3+ 31. Kc2 Kg8, even though I’m down an exchange and a pawn and it’s White’s move, the computer says it’s a win for Black.

Position after 31. … Kg8 (analysis). White to move.

FEN: 1r4k1/4np2/3p2p1/3P4/2P1P2P/4q3/2K5/Q5RR w – – 0 32

By contrast, this almost identical position, after 29. Qa1 Qe2 30. bc Kg8 (a move I would be more likely to play; prophylaxis), is only a draw for Black.

Position after 30. … Kg8 (analysis). White to move.

FEN: 1r4k1/4np2/3p2p1/3P4/2P1P2P/2K5/4q3/Q5RR w – – 0 31


Alas, I think that a full exploration of the above two positions might leave us disappointed in the end. It would all depend on innumerable intricacies and subtleties and “if-this-then-that” variations. And in the end, we would not actually be able to learn anything of lasting value from it. The only conclusion would be that there are some positions only a computer can play accurately, which is probably true but depressing.

So I’m not going to say anything more about these two positions, but for readers who want to try challenging their analysis skills against the computer, go for it!

Print Friendly, PDF & Email

{ 6 comments… read them below or add one }

Marshall Polaris November 13, 2020 at 3:02 pm

25… b4 looks very reasonable to me and I don’t think it needs any exclamation points or question marks. It’s clearly challenging to White, and it’s hard to believe that Black won’t at least find a perpetual after Qxh5, so it seems hard for me to criticize it.


admin November 15, 2020 at 8:34 am

I think this is a good practical way to look at it. Black is sacrificing material for an attack that seems to give very good chances of a draw and significant chances of winning. My original assessment of “aggressive, risk-taking” may be the more important point, and the question of whether it’s sound or unsound is not as important.


Mike Splane November 16, 2020 at 11:07 pm

Well, Dana, you’ve changed my argument. My point wasn’t to determine the objective truth about the best possible move in the diagrammed position. My key point was you often lock onto candidate moves and don’t consider other good possibilities. I could easily go through old columns and show you at least twenty examples. In the game with Belle, at move 44, for example, you could also look at 44. Kf4 or 44. Rf4+, both of which slow down, or even stop, Black’s king-side pawn advances. Again, I’m not claiming one move is better than another.

Second, and this is a new point I don’t think we have ever discussed at any of my chess parties. you can NEVER win a game is your opponent plays perfectly. You can’t evaluate the strength of a move or positional idea by using computer analysis. If your opponent uses a computer engine it is called cheating. You can’t justify choosing one move over another by using computer analysis, computer analysis is irrelevant in a competition between two humans.

You win by giving your opponent problems to solve and waiting for the mistakes. They always come! People get tired. They miss candidate moves, they misjudge the strength or weakness of pawn moves, they overvalue material advantages., they get impatient and play impulsive moves, they recapture with the wrong piece, they expose their king, they underestimate the strength or weakness of passed pawns, they defend when they should attack and vice versa. If you are on the defensive, you can be pretty sure your opponent will falter at some point. You just have to keep your pieces active and your chances will come. If they don’t, your opponent is playing master level chess – it happens – so then you just shrug your shoulders, congratulate your opponent, forget about the loss, and move on to the next game.

Going back to the diagram, Re5 sets problems and b4 sets problems. Would you rather be down material in an unclear and possibly lost position or keep even material with all your threats intact. Personally I think the former is a lot of fun so I don’t blame you for your decision. Marshall makes a good point about your move giving practical chances for a perpetual check.

BTW, Now that I’ve seen the analysis of the computer engine, I’ve become convinced that you played the right move.


Marshall Polaris November 22, 2020 at 11:35 pm

> My key point was you often lock onto candidate moves and don’t consider other good possibilities.

Incidentally, I definitely suffer from this problem, and I think if I had encountered this position in a game, I also would have started analyzing b4 and played it after convincing myself it was pretty good, without seriously considering Re5.

One reason I would prioritize b4 here is that I am very belligerent. If my opponent plays Qe2, it’s just my nature to do absolutely anything except move my rook. I think this tendency is generally valuable because it leads me to find my opponent’s mistakes, but it also sometimes causes me to be biased against the best move.


Mike Splane November 18, 2020 at 7:40 pm

Hi Dana,

I’m sorry if it sounds like I am singling you out for criticism. I’m not. I keep mentioning tunnel vision because I see it as a common problem among most players so I thought it was worth discussing and making other players aware of it as a potential problem to be avoided. During game discussions at my chess parties. I feel like I’m looking at more possibilities than the other players, both the masters and non masters. Most of the time those other possibilities are no good. But I love exploring the richness of chess so I continue to seek out all of the possibilities in each position.

I can only remember one exception. One afternoon we were looking at an extremely complicated position, I played a safe move that gave me a long term endgame advantage. Mike Arne wanted to analyze out all of the tactics. I said something like, ” you can’t play chess that way, you’ll lose on time.” Gjon Feinstein immediately contradicted me. “YOU can,” he said. “I’ve seen you do it many times.” That was a good critical remark that helped me realize I was getting lazy in my old age.

I remember you asking me why I played the safe move and I replied, “because I’m winning the endgame.” You criticized me for that remark when you write up the game in your chess blog, claiming it was overly optimistic. As you correctly pointed out, the endgame was a long way off. Yet, it fits in with my philosophy in my previous comment, put pressure on the opponent. My opponent had to attack BECAUSE the ending was lost. All I had to do was wait for the inevitable mistakes. Sure enough, my opponent lost the initiative, had to trade down, and I won the ending because of that safe move. Against a computer opponent it was probably the wrong move.

Here’s what you wrote in your blog post
“Mike’s comment on this move at the chess party was totally characteristic of him. Why did he play this move? “I’m winning the endgame!” he said.

On the one hand, this is totally nuts. Black has lots of chances after this to equalize or even stand better. On the other hand, it’s so Mike. Remember the Mike Splane question: “How am I going to win this game?” He asks it in almost every game, whether his position is better or worse. Here, how he is going to win the game is very simple. He’s going to win Black’s two weak queen-side pawns, at c4 and a6, and then he’s going to run his connected passers to pay-dirt.

The deeper lesson here is that the “objective evaluation” or “computer evaluation” doesn’t always matter in a contest between humans. If you have a plan and your opponent doesn’t, you’re winning.”

I’m mentioning this game in some detail to help your readers get a better idea of why my chess parties have been so valuable for everybody involved. We have established a level of trust and openness so everybody feels free to comment without fear of being ridiculed. Everybody is open to learning new ways of thinking. Criticism is accepted because it is meant to be constructive. It’s a community in the best sense of the word.


admin November 21, 2020 at 7:49 am

I don’t want to leave your two lengthy comments unanswered, so here goes.

First, as a general warning to all chess players, “tunnel vision” is a common problem for players at all levels. As humans we all have biases, and one of the most important ways to improve is to consciously broaden the set of possible moves we look at. The worst sin is to look at only one move. In this case you’re not really analyzing, you’re forecasting. “I’m going to play Qg5, so what is going to happen next?”

A book that helped me understand the ideal thinking process is Kotov’s famous “Think Like a Grandmaster.” For any readers who have not looked at it, you should — at least to familiarize yourself with the concept of “candidate moves.” Kotov says that you should formulate a list of candidate moves, and then analyze each one *once and only once*. This is very different from what many players, including me, often do: (A) Analyze one move, realize it doesn’t work, panic and start looking for other possibilities. Or (B) Identify two possibilities, then keep on vacillating back and forth between them while the clock burns.

Unfortunately, there are certain gaps in Kotov. He doesn’t explain where the candidate moves come from in the first place. And he doesn’t explain how to modify this procedure when you are in time trouble. That’s why I call it an “ideal” thinking process. It’s what you strive for, but it’s just not practical in all positions and at all times. If you are consistently far away from this ideal, your results will suffer.

Now proceeding from Mike’s general advice to his specific advice to me, I feel as if it’s a little bit off base, which is why I have avoided responding to it. One of my key weapons in chess is imagination — considering moves that other players wouldn’t consider — so it’s a little bit strange to hear that I’m not being imaginative enough. I think that Mike is basing his comments in large part on watching me play speed chess and to some extent on our chess parties. The second is an unusual situation because, in order to be heard among several other players, you often have to argue the case for a certain move very forcefully, which (if you’re wrong) looks a little bit as if you have “tunnel vision.” Speed chess is also a different situation from tournament chess. But let’s talk about that.

I’ll say it right now, I am a crappy speed chess player. In 5-minute chess, I’ve always played below my tournament rating. Part of the reason is that I have never figured out a way to implement anything like Kotov’s method so rapidly. To keep up, I have to use a lot of intuition, I have to cut corners, and I often have to do exactly what you shouldn’t do — go with my gut instinct, without giving other moves a chance. In other words, “tunnel vision.” I completely admit to having tunnel vision as a speed chess player. I also have incredible admiration for players like our mutual friend Gjon Feinstein, who is incredible at speed chess because he somehow manages to consider multiple possibilities in just a few seconds.

In one sense, this *is* relevant to tournament chess, because often in tournament chess one gets into time trouble. I find that if I get below the 1-minute-per-move threshold, I start cutting corners and start having tunnel vision. My strategy for dealing with this has mostly been to try very hard to never get below 1 minute per move. However, I would definitely appreciate practical, workable advice on how to avoid tunnel vision on those regrettable occasions when I do have less than a minute per move. This is especially important these days, when many FIDE tournaments have a per-move time increment of 30 seconds, and so GM’s learn how to “play on the increment.” I can’t do that. Playing on the increment for me is an invitation to disaster.

This has been a long reply, I know, but I didn’t want to look as if I was ignoring Mike’s comments!


Leave a Comment

Previous post:

Next post: