The Machines.

"If chess is a vast jungle, computers are the chainsaws in a giant environmentally insensitive logging company."  -  Nigel Short


To trust or not trust the silicon beast, that is the question. As we all know, computer chess engines and hardware have become faster and stronger than most of us could have imagined even 20 years ago.  As it is today, many of us take laptops with us to tournaments as little know-it-all companions to show us the way. 
        "Beep Boop, what are you crazy? 2..F5? Have you lost your mind? Who told you the Latvian Gambit was okay?" says Houdini. A Red circle begins to swell. You respond pleadingly, "But I wanted a sharp game, I wanted to take my opponent out of his preparation, he knows the Ruy Lopez better than I do.." Houdini responds in a condescending tone-"Oh please, you see, 10 moves from now, provided you see you what I see and do as I do, you'll be stuck with a position even my old friend Fritz would scoff at." "Oh yeah? Well you see I have this little X in the top right-hand corner of my screen, piss off." 


Trusting a computer's evaluation and move-by-move analysis is a tricky thing. There are times when the machines are god-like. They are never wrong in calculation, they never make tactical errors, they are cold-blooded mathematicians with no concern for your sense of battle and psychological trickery. So how should we use them?


First of all, we should remember that engines like Houdini, Fritz and Rybka were of course designed by humans. But what exactly does that mean for us? It means that someone like you or me designed a machine to play chess. They calculate far better and more accurately than any human ever will by virtue of their binary model, but how do they choose moves? What is the filter by which they come to recognize that 1.h3 sucks and that 1.e4 and d4 are better? Programmers in partnership with strong, IM+ strength players programmed a hierarchy of move-selection processes that tell the engine to focus more closely on certain moves and not others, exactly the same way we learn to play chess with "candidate" moves. We learn to filter out certain moves from our thinking because they look "ugly" or "weak", and because they don't conform to our sense of how "correct" and "good" chess is played based on principles that were first laid out systematically by Wilhelm Steinitz.
           But if the chips do everything we do but better, should we not trust them at all times? No, not at ALL times.


As every tech-savvy tournament player is probably aware, chess engines are great tools to show us where we went wrong in our games; The exact moment of a tactical oversight, a slight inaccuracy in the execution of a plan and so on. But when we begin to take their word over our own, that's perhaps when you should re-think why you enjoy playing chess and what it means to you.  Here's a little story to try and illustrate my point..


I recently came within 100 rating points of my friend and nemesis Micah Hughey. I proposed we play a 4-game match. He agreed. How should I approach this match, I thought. Hmm, well I know he has a pretty restricted opening repertoire, and tends not to try to switch it up based on who he plays, I can expect to see these openings on the board. Great I thought, this'll make preparing for him a breeze. So I open up my Rybka and take a look at what the machine says about what he plays. "Hmm, no real edge there, slight pull there, ah this looks dull, god damn it Hughey, play something interesting!". Then I sit and think for a second about what Micah likes and doesn't like. AH! The light bulb pings. I know, I'll play the Pirc, surely that'll go against his grain a bit. So I start flicking through some opening lines, checking some databases to see who he's played in this line and what setups he tends to use, and I come across a game he played back at the 2005 Canadian Open. Excellent I thought, he won't expect this from me, his prep will go to waste, and perhaps I might gain a slight advantage on the clock. As it turns out, he had in fact played this line a few times as he told me after the game. Unfortunately, the line I had a chosen was one recommended by who else, Houdini. The material balance became odd-- My queen and two pawns vs. his 3 minor pieces. Now according to the non-human magician, I should have a small advantage right? I have more points of material than he does. But how to proceed once the tactics have stopped? "uh, crap." If this situation has ever cropped up on you, I highly recommend you take a step back and realize how a computer evaluates positions and plays chess in general. In my case, I wanted to trust a strong engines evaluation because it's advertised as being 3300+.  But why? In this game the computer led me astray because it's the type of position that maybe only a computer could play properly based on 12+ move calculations, something I'm not capable of. It was tough for me to try and come up with a reasonable plan because I was only prepared to play the position through the eyes of the machine instead of relying on more human concerns. Maybe Houdini thinks I have a slight advantage in a purely theoretical sense, but at the board with my clock ticking, I had a hell of a time trying to do anything constructive and bemoaned the rationale of the number cruncher. I lost. 


As evidenced by this game, humans and machines do not play or "think" about chess in the same way. Silicon monsters make moves based on concrete calculations at a depth that most humans could only dream of. They rationalize courses of action based on "I go here, he goes there". Human thinking is far more abstract and self-preserving and chess games between humans should probably stay that way. Trust the machine so long as it already confirms, at a greater depth,  what your intuition already tells you. 
          We have little men on our shoulders who say things like "Oh don't do that, that's too risky" or "Yeah let him have that pawn, that king looks tasty". Our human-like way of thinking is streamlined for efficient intuitive decisions set on digestible principles . Don't try to be a machine, it leads to the dump.