| Sklansky gets it wrong ... |
[Oct. 16th, 2008|10:33 pm] |
| [ | mood |
| | full | ] |
| [ | music |
| | Secret Agent on SomaFM | ] | ... though it is a piddling detail a bit outside of poker. He says:
While no one plays perfect chess, it's likely one could play perfect chess. And if you did, you'd never lose."
Leave asides the question of whether one could play perfect chess, i.e. that there's an algorithm for doing so. In a footnote he continues:
"Well, you probably would never lose. In chess, the player with the white pieces moves first, and that privilege confers an unbalanced advantage throughout the game. So it is possible that if two "prefect" players played against each other, white would win every time. But more likely, as in tic-tac-toe, they'd draw game after game."
The problem is with the unbalanced advantage bit. Humans think white has an edge, because it seems to us that it is easier to win with white than black. But it could in fact be the case the white starts out in zugzwang, and it is simply the case that we're not built to see it.
There are several examples of games where the second player wins if the starting conditions are right - Nim comes to mind. It can even be the case that humans think the first player has an advantage, but the reality can be somewhat different - thrusting vs. parry-riposte in fencing, for example.
Anyway, I suspect what sent me off to posting is that I've seen this example touted by people who should know better before, but seeing it from David was a bit of a surprise. |
|
|
| Comments: |
Fitchneal usually confers the advantage to the player that moves first. A second player win is generally only possible if the first player doesn't have a strategy/clueless/new.
This is possibly a good example of what I'm talking about. It _appears_ the first player has the advantage, at least to humans. But it could turn out that there's an overpowering but difficult deep strategy that really gives the second player an advantage, if only he had the wit to discover it.
This sort of stuff happens in wargaming all the time. Folks will play a wargame, and come to the conclusion "Game X is imbalanced in favor of side Y", and that will become the standard belief until someone comes around and shows a strategy the destroys side Y using the standard strategy. Then folks change to "Game X is imbalanced in favor of side Z" and we iterate.
Now, many games like chess have stabilized to the point where we're pretty confident in saying one side has the advantage - but this is a different sort of confidence than we would have about making claims for games like tic-tac-toe, Nim, or checkers, where we can prove what the results would be regardless of how the opponent plays.
If I understand you, I think Fitchneal is probably a pretty good example of that. The object being to get the king off the board, if the first player moves appropriately, it is very very difficult for the second player to meet the winning conditions. Generally what winds up happening is a stalemate.
Aren't the morris games heavily weighted toward first player? I seem to remember sitting in on a discussion about strategy in morris games and that being discussed at some length.
Fox and Geese have varying weight, depending on how many foxes and how many geese. That isn't the same thing, tho.
There's a whole field of mathematics that studies two-player, zero-sum, perfect-information, turn-based games (like chess, checkers, tic-tac-toe, etc. where both sides can see the whole board and there is no randomness). There's even an algebra of games: if X and Y are games, each with its own board, then X+Y is defined as the game with BOTH boards in which the first player can decide whether to make a move on the X board or the Y board. Based on this, one can define subtraction, multiplication, division, square roots, etc. as operations on games (which is the sort of abstract nonsense that mathematicians like me love).
Anyway, there are plenty of games in which the second player wins, if both play perfectly. I don't think anybody has completely analyzed chess, but the borderline between first-player-wins and second-player-wins is so delicate that it's quite possible that Black wins at chess.
The parry-riposte is a huge part of modern fencing, but in historical European martial arts, it doesn't come up very often; instead, historical authors describe and advocate counterattacks that defend and offend with a single action (what Italians call "stessi tempo" and Germans call "meisterhauen")
Both Italian and German authors find these single-time counterattacks so effective, that they substantially build their martial systems around them, so that the focus as a fighter is on creating a situation in which any action the opponent makes will fall into one of several prepared counters. Italians spend a lot of time describing the effectiveness of such positions (the 16th and 17th century authors, at least, often use the term "constraining" to describe it), though the Germans conceptualize it rather differently ("crowding in on his every opening")...
Anyway, pretty well zugzwang-y. | |