Quote:
Originally Posted by zizek
...And to preempt the obvious response that "well chess must not be a game of skill under that definition", the difference is that in the perfect game of chess, a certain player (probably white) is ALWAYS going to win, whereas at a table with 6 people playing perfect poker, the winner would be completely random.
Isn't the assignment of white/black determined by a coin flip? Between players with identical strategies, the winner of a chess game is completely random (assuming the unknown solution to chess is "white wins" or "black wins", rather than "draw"). Chess can be replaced with any symmetric strategy game. Your argument suggests that all such games are games of luck.
I see that a few others have beaten me to this.
As far as approaches to predominance go, David's seems pretty reasonable as far as convincing laymen. One issue is that the classification of the game changes dynamically as the player population changes. Various forms of poker might be a game of skill today, but, in a hypothetical future where everyone who plays poker has become much better at it, the average player might play well enough such that the expert only has a 70% probability.
This again brings us to the issues raised by the above paradox. Any empirical approach to measuring skill/chance based on results of actual real-life gameplay will fail for any game where players' strategies are sufficiently close to each other. Poker games which were skill under this measure would cease to be skill
if all of its players started playing much more skillfully.
Whatever the predominance test really "is", I'd say that an ideal property would be that it returns the same result regardless of the particular population that chooses to play the game at any given time. Skill-ness should be an inherent property of a game. So perhaps we need some framework for establishing a dummy "average player" for any given game. Maybe take a random person who has never played before and give them X hours of training?