bad moves can be good

Sometimes the best move may not be the one that yields the best probability of a win. An example of this is Tal, who often played unsound moves that raised more problems than his opponent could solve. I am not Tal, but here is a Blitz game I played on the Chessbase site yesterday.  The interesting move is the sacrifice of the b2 pawn. Although the sacrifice was unsound, the opponent couls not solve the threats.

---

It depends from what point of view. A poor move may only be poor if the opponent can find the refutation. So a move that yields the highest probability of a win against a human may not yield the highest probability of a win against an 8-core computer calculating 25 moves deep.

To be concrete,consider a move that is slightly inferior but for which there is only one hard-to-find move that does not lose, vs a pedestrian "best" move for which there are 4 or more moves that maintain the equiliblium. In my view, the "inferior" move is better, especially in Blitz.For example, in the game above,  9. Ra1 may not be the best move, but after 10. Ne2 to counter the check, the only Black move (winning) is 10...Kd8, far from obvious and losing the ability to castle, which of course is refuted by the Rook sacrifice.

In sum, what I mean is that one could play every move in such a way that the opponent had only ONE good reply, he would almost always win (provided that that this one good reply did not win the game instantly...), because there are few players if any that can find the best move every time. Post-analyzing some of my games with  a computer has shown cases where there were series of 10-moves where each side had 4 or more moves each time that did not change the balance of the game; clearly one is not going to get a won game in situations like this.

The rook sac was nice, but you had very little for the two pawns except for hoping he wouldn't make that one blunder.

This is different than playing a dubious move but one with many tactical chances. This was more like hoping for an outright blunder, because you had very minimal compensation. Compare this to Tal: he may have played unsound moves, but he would reject many of these ones, and he would not beat his opponents playing ones as shallow as in the game, but instead with more tactically challenging moves. He would at least have decent compensation.

---

I agree, but isn't it all a question of play level? - the philosophy is the same: Tal played a (Grandmaster) inferior move because he figured that his opponent (in an OTB normal game) would not find the Grandmaster refutation. I played an "inferior" move (the second pawn sacrifice)in a blitz game figuring that my 1200-rated opponent might not find the refutation that meant moving the King and losing the right to castle. (In fact it was more like trying to apply Fischer's "...sac, sac...mate!".)

In fact had he played the King move, I still had pretty good compensation for the pawns with a huge lead in development and his King unable to castle. Fire evaluates the position as -/+-80, which is not enough to win for Black. In a Blitz game between low-rated players, such an advantage is meaningless.

---

Yeah... it is. I guess it depends on how often you can expect stuff like that to work at a certain level, but I would still not want to get into a habit of playing moves for the sake of creating some kind of tactical threat even if it's very unsound. There are moves that are borderline unsound but create a lot of problems so they are ok, if you're that kind of player, but there isn't always a good way to sacrifice, sometimes not even ones that really create problems!

I think we're all told that when we are worse we should look for counterchances, and many times I have made the mistake where when I'm worse I try to create some tactics to at least make him calculate a bit... the problem is the tactics favored him, and my opponent was no joke, he certainly could handle some tactics, and so though I was a pawn down I ended up losing a ton of material and lost very quickly just by getting too desperate. I think that if I instead put up a tough defense (it takes lots of technique to win pawn up, probably like 30-50 moves, not 10!), and only try for some swindle when it was a little more realisitc, I would have done much much better.

So I'm basically saying you should take the principle into moderation. Sometimes you keep your opponent off balance playing those "unsound" moves, but other times the move is so unsound as long as your opponent doesn't completely hallucinate your position will fall apart. So although tricks like in the game may work, I wouldn't get into a habit of it.

Some part of a mistake is always correct - Tartakower

So, I have had two games recently where i put the oppentent in check where they had a few options, if they chose the right option(s) I would be in serious trouble, but if they chose the wrong option I would win material and save myself from a bad postion.  In both these instances they chose incorrectly and I won both games. 

For one game I was about to be mated, so it was a last ditch effort to survive against a much stronger player.  I feel like it is a victory when I am able to draw against him. 

The second instance was with a player on, or about, my level.  Not judging by rating, but just based on the game.  I was in a position where I could make a move where if he/she answered correctly I would be in trouble, but if he/she answered incorrectly I would be in a great position.  I had a feeling he/she would not make the correct move, but I passed it up because I was otherwise in better shape.  After getting into trouble I decided to make the move anyway, and they chose wrong and resigned when I took both rooks. 

So I am trying to figure out, whether it would have been good to risk my advantage on a move that could have led to a disadvantage or great gains, depending.  How does one decide when to take these risks? 

Oh, and if you look at my last game (the one in question) I made a bunch of mistakes.  For sure when I could have mated instead of taking his queen.  So try to keep laughter at a minimum, please. 

I have two competing philosophies here. The more important one is "If I can see it, so can he". This means that no matter how desperate I am, I won't play a move that I can refute personally. Well... unless I'm drunk, and that doesn't count because I'm a terrible player when I'm bombed. The less important philosophy is that you can only lose once per game. This obviously doesn't effect positions where I'm winning or equal. But from a losing position I'm happy to make a move that I think is dubious (or at very least is not the most precise defense) provided that A: I can't personally see the refutation and B: I feel that the complications raised by the "inferior" move are enough that it's more challenging to beat than the more precise (but still losing) continuation. *shrugs* I'm not saying that I strive to play dry chess (I do occasionally, but more often I look to play open-creative games). Only that I try and play the BEST move possible every time I get to push a piece.

---

You make some very good points. I agree that in a losing position, there's no reason not to play for complications or to go for a swindle.

---

I agree, but  in the above case, I had not seen his "refutation" with the King move either, I was just playing attacking moves because if I stopped attacking I would probably lose. But I HAD seen the threat that materialized when he didn't find the refutation. So the point is that one has a better chance of winning by making threats than by refuting them, especially in blitz games. And to turn your statement around, if I can't find the refutation, there is a pretty good chance that an opponent of equal strength will not find it either....

Thanks Travis,

Like Josh says in Chessmaster "never make a move that depends on your oppenent making a mistake" or something like that. 

I think what you are saying is that sometimes it is better to make an uncertain move than one which is the best predictable outcome but which is still losing.

---

Like Dan Heisman (I think) writes in his Chesscafe column, one should avoid playing "hope chess", which is what Josh is saying, and I don't disagree. But both of them encourage playing moves that create threats versus moves that increase security. Finding the right balance is the trick.

Frank Marshall made a career of winning games by means of "swindles"; one should not make moves that depend on the opponent making a mistake, yes, but one should not work too hard on solving the opponent's problems either...

nice game nice attack. Maybe i should try an opening like that

This is a very interesting topic. I played a game recently in which a gave up a pawn that allowed him to fork my rook and king. Had he not taken the rook at the end of the combination, however, I would have given my opponent a reasonble advantage (.8 pawns according to shredder) but because he took the "poison rook" he allowed me an interesting checkmate. The only reason I played this move is that the time control of the game was three minutes and i figured that even though I was risking an unfavorable position, i might be able to last it out anyway or attempt to retake the advantage.

Here is another example from a game I finished today.

 

 

 

 

 

 

now here according to  computer analysis, my best move is Qg4 and wins, with plenty of chances to make mistakes. Instead I went for the throat by capturing the pawn and going for the black king right away. Bxa7!?

After this move, Black has only one move that can hold the game ...Qb6!, but this move is terribly hard to find when your King looks like he is about to be mated! So what else is Black to play? There is no other move, and he lost quickly by playing ...Nxd4.

I hold that the aggressive move Bxa7!? is a "better" move than Qg4 for players of our level because it is aggressive, leaves no room at all for Black to make a mistake, makes it unlikely that Black will find the move Qb6 (I had not seen it either), and leaves less room for White to make mistakes in the rest of the game.

Here is another example of winning with a bad move against a player rated 2000. On move 16, I made an unsound sacrifice (Nxp) based on the fact that the opponent would likely interpose the Bishop instead of moving his King to the corner when I checked with the Bishop. I figured that he would prefer blocking with a developing move rather than moving his King to the corner and facing a dangerous attack with the knight (that would have led to nothing and leaving me a piece down. He did interpose the Bishop and I won against a player rated 200 points higher than me.