Would You Make these Positional Sacrifices?

Forgive me for beginning with an obvious statement (sometimes I feel it necessary to be obvious). Sacrifices, especially long-term sacrifices, are one of the most exciting aspects of playing and watching chess. One must appreciate Bobby Fischer's famous queen sacrifice in the "Game of the Century" against Donald Byrne, no matter how many times Ben Finegold calls it overrated.

Like in the Fischer game, the best response is to ignore the gift of the queen and capture the knight, but of course my opponent had to capture the queen. He attacked it, didn't he? That reduces his advantage to (a measly) +2. It was the subsequent pawn capture that was a bridge too far. But it looks playable, and the top computer move: Q-g5, is not obvious. (And does it win a pawn? No, it does not.) Now my pieces are great and I quickly take over.

When the game ended, I felt that I had paralleled Fischer. It wasn't some "queen sac" that leads to mate in 1 or 2 moves (yawn); it was a queen sac to make my other pieces better. I felt like Fischer, that is, until I analyzed my game with the computer. Computers are such wet blankets (and I mean that figuratively). 

The computer cannot, however (and I know it wants to) tell me that my sacrifice didn't work. I won, didn't I?... No, don't worry, I'm not so foolish as to conflate results with efficacy. What I'm suggesting is that my move might have given me (specifically me) a better chance of winning (against my specific opponent/against any opponent) than the alternative computer move in the situation. 

My opponent was rated 1900. The Game was 2 minutes, with 10 second increments.  I am only so good at chess. Was it more likely for me, were I to encounter this position 100 more times against equally strong players (or even stronger players), capable of winning more frequently with my faulty move than the computer's boring one?

I'm going to be a little self indulgent, if that's okay with me (this is a blog, after all). In a college political science class, I was introduced to the "Guess 2/3 the average" game theory game. The way it works is everyone in the room privately writes down a number between 1 and 100, with the goal being to write the number that would ultimately be 2/3 of the average of all the numbers written down. In this instance, the winner received five dollars. (I wanted to win those five dollars.) You can decide what number you would write down now, if you want to play along. I eventually settled on 19.

I reasoned that if we all wrote random numbers, 2/3 would be 33, but we all would know this, so therefore many people will write 22. Of course if most people did that, the right answer would be 14.6, and so on, but I wasn't willing to give my class that much credit, so I settled between people being able to think two steps ahead, and people being able to think three steps ahead. 

The class average was 41. (I suppose I must now admit that I did not attend an Ivy.) Before this was revealed, the teacher explained (to the surprise of many) how you could keep reducing a number by 2/3 until you got to the "Nash equilibrium" of 1. 1 is what a computer would have guessed, and it would be even wronger than I had been. Sometimes the "smartest" choice is the wrong one.

The winner had written down 29 - I have no idea why - and I profited not at all (except for this thrilling anecdote I can tell both in chess blogs and at parties).

What inspired me to write this article was the game below, which I played earlier today. (If you are a chess computer, you're probably better off not watching my opening (or the rest of the game) - you'll hate it.)

I really thought I was controlling the board after 13. Nxd5, so upon review, I'm amazed by the utter wrongness my moves. However, I can't help but think my opponent shared my belief that I was winning the whole time I wasn't. Black is scrambling to meet my threats every turn, and they're scary threats. It looks good for white, it feels good for white, and if it quacks like a duck... does that expression apply here?

In this instance my opponent was rated 2000 and the game was 6 minutes with an 8 second increment. The question I continue to ask is: did I improve my chances of winning despite the computer's demonstrative objections?

But I don't want to limit this analysis by suggesting that faulty tactics are simply a club player way of beating other club players (and although my example below is a rapid game I don't want to relegate this concept to solely blitz and rapid). One of the games I enjoyed at the ongoing Saint Louis Rapid & Blitz tournament was Aronian's game against Maxime Vachier-Lagrave in which he played the novelty, N-g8.

This move did not, according to the computers, or the analysts, improve Aronian's position. But there's more to it than that, even at the highest level. When encountering an unusual or surprising move, one must consider that your opponent has seen (or prepared) something that you have not. And simply because it wasn't considered, one must spend possibly valuable time considering it. Additionally, take the opponent's fallibility out of the equation for a moment, and focus on one's own, or my own. Even if I find the best move in one position, that does not mean the following moves will be easy for me to find, while an inferior line might be much easier for me to play. There is a hard-to-quantify advantage to be had in all of this.  

I did not write this as a longwinded way of justifying my mistakes. I wrote this to point out - yes, longwindedly - that the Nash equilibrium does not always win five dollars, and that the most accurate move might not always lead to the best chance of victory. If the best move has a natural-looking rebuttal from your opponent, nullifying it, or declawing it, and an inferior move requires counterintuitive and/or precise responses to overcome, which is the correct choice? Perhaps our goal should not be to emulate a computer at all times, however superior they may be, when facing a person, and while being a person.

One must also (reluctantly) consider the larger picture (even if it is contrary to my thesis). We want to practice good habits in our games, and shed bad ones. Suspect moves that you get away with because your opponent misplays his or her response is not the best habit to form. But hey, that's why I check the computer after my games, so I don't get away with anything (more than I already have). But I feel I'm getting dangerously general. Let me be more specific. The whole thing boils down to two simple questions: 1. Do you find the two sacrifices I made effective? and 2. Would you yourself making them?  

I'll conclude with a dubious analogy (which I think is appropriate). Steve Nash averaged about three turnovers per game in his NBA career, but he still won two MVPs. That's my kind of Nash equilibrium.