Karpov: Super-interesting discussion of luck in chess vs backgammon, attributed to [Bill] Robertie.
Best way I saw it put is [by] Robertie…
'Chess is also a probabilistic game but there is no obvious source of randomness so people don't notice [] it is a probabilistic game. [Where it is, in fact] as probabilistic as backgammon but in a different way.
Let's suppose Gary Kasparov, a 2800 player, goes up against a Grand Master who is a 2600 player. If they play a number of games over [] time, Kasparov is going to win 3 out of 4 of the decisive games pretty reliably. He's not going to win them all even though he's clearly the better player. And if Kasparov plays somebody [] he's 400 points better than, he'll win 90-odd percent of the time, but the other guy will win a few.
Why is this happening? It looks the same as if you put two backgammon players, playing 25-point matches, against each other. You'd see the same distribution. You'd see the better player winning a higher percentage, and it would be a pretty steady percentage over time, as long as their skill remained constant.
If you stuck these players in black boxes, [] all you knew from the outside is over here in Box #1 the results go WLWWWLWL and over here in Box #2 the results go WWLWWWL[. Which] box has the two chess players and which box has the two backgammon players? You can't tell the difference. You just get these strings of results and somebody is winning most of them but not all. It's obvious why this is happening in backgammon because you have the dice going on. But why is it happening in chess?
Chess is like this. Imagine it is dark and you have two people with flashlights. One guy has a big flashlight with a wide beam and the other guy has a little flashlight with a narrow beam. [Out] in front of them is a forest, and in the forest is
some treasure. You say to each player, "Okay, you guys go [] in the forest and first one to find the treasure wins." Well, the guy with the big flashlight is the favorite to find the [booty], but it's no guarantee, because the other guy could luck his way and find it [first].
And a game of chess is like that. You look at this position, and there are all these variations [] string out in front of you. You know, "I could do this, or I could do that. And I've only got time to analyze three of these ...." You're searching this tree of possibilities. Depending on how fast you think and how much you know about the game, your tree search is more efficient but it's not guaranteed to get you to the best answer.
The other guy, who's maybe clumsy[,] doesn't think as fast as you and doesn't know as much about the game, could still walk into the main variation almost by chance. That's where the chance in chess comes from. It's entirely invisible. In fact you might feel a little uncomfortable even referring to it as "chance." But [really that's] what's going on, two people basically searching a tree at different speeds.
They are still trying to find the main line, the line [describing] correct play on both sides. The better player mostly searches it better[,] but not always.
That makes chess a probabilistic game, [] that means it's got a lot in common with these other games, such as backgammon and poker.'
No comments:
Post a Comment
sure, you can comment -- but why?