I really enjoyed watching the first game of the World Championship yesterday, even though I'll be the first to admit I'm a complete patzer who didn't really know what was going on half the time. What really annoyed me though, was the number of people in the chat spouting computer suggestions and eval numbers, claiming a player "blundered" if they played anything other than the first best move, and thinking that a position is "won" if a player has a +1 (or -1) advantage.
I was watching the Sesse analysis and it amazed me that a particular thing kept happening: the evaluation would be a number, let's pick 0, and yet Sesse's evaluation for it's best move would be something like 0.67 if it was white to move, or -0.67 if it was black to move. This number would get closer to 0 as the search depth increased. In other words, even a really powerful computer took it's sweet time figuring things out. Yet people expect humans - they are the best human chess players in the world, but they're still humans - to somehow convert this minute "advantage" into a win, when in fact, that number is the result of a "mighty computer" getting its sums wrong, up until a certain depth is reached at which point it adjusts the evaluation accordingly. By this logic, the +0.2 evaluation given by Stockfish at the start of a game "proves" that chess is a forced win for white.
As I said above, I suck at chess.... but I do know that a computer evaluation is not a final verdict, it's a guess. OK, a very well educated guess, but still..... Magnus Carlsen probably did have a winning position at one point, but the road to winning is long and complex, which is why we saw a +4.5 eval and not a "mate in N" eval. So please, if you're going to give computer lines and evaluations, at least learn what they mean.
To draw parallels with other sports, that would be like criticising Messi for missing an "easy" shot by saying "Why didn't he just kick the ball into the goal?". It's a massive oversimplification of a beautifully complex game.
The irony is that if both sides had played perfectly, matching the computer move for move, it would probably have been a draw, and those same people would moan at the players for being "boring".
I was watching the Sesse analysis and it amazed me that a particular thing kept happening: the evaluation would be a number, let's pick 0, and yet Sesse's evaluation for it's best move would be something like 0.67 if it was white to move, or -0.67 if it was black to move. This number would get closer to 0 as the search depth increased. In other words, even a really powerful computer took it's sweet time figuring things out. Yet people expect humans - they are the best human chess players in the world, but they're still humans - to somehow convert this minute "advantage" into a win, when in fact, that number is the result of a "mighty computer" getting its sums wrong, up until a certain depth is reached at which point it adjusts the evaluation accordingly. By this logic, the +0.2 evaluation given by Stockfish at the start of a game "proves" that chess is a forced win for white.
As I said above, I suck at chess.... but I do know that a computer evaluation is not a final verdict, it's a guess. OK, a very well educated guess, but still..... Magnus Carlsen probably did have a winning position at one point, but the road to winning is long and complex, which is why we saw a +4.5 eval and not a "mate in N" eval. So please, if you're going to give computer lines and evaluations, at least learn what they mean.
To draw parallels with other sports, that would be like criticising Messi for missing an "easy" shot by saying "Why didn't he just kick the ball into the goal?". It's a massive oversimplification of a beautifully complex game.
The irony is that if both sides had played perfectly, matching the computer move for move, it would probably have been a draw, and those same people would moan at the players for being "boring".
