Quote:
Originally Posted by Aaron W.
There's probably a way to measure that, but I don't know what it is. But this is also the type of game where knowledge of a sub-game is only sometimes useful. For example, a full knowledge of getting 3 in a row in a 3 by 3 Tic-Tac-Toe board doesn't inform you about the trivial nature of getting 3 in a row in a 4 by 4 Tic-Tac-Toe board.
Right, I think that suggests we could study this evolution of an expanding game.
Quote:
The computers kill us these days.
Bah, I can be terrible with language. By "can" I meant "possible", I am surprised its even possible, for example for a human to win at a game of go. And I found an example but its from a year ago, and I would believe that possibility has changed already.
Quote:
The deep learning algorithms are in the same condition.
Yes they simply have a computational advantage. But it's interesting then to consider the limits of that advantage.
Quote:
Modern artificial intelligence learns from playing lots of games and that stands as the empirical evidence. Computers struggle with this because two identical situations (say an open-raise from UTG) can lead to two different decisions based on historical knowledge of the player's tendencies as well as other player dynamics at the table. But computers don't really have a means (yet) of picking up those sorts of details, and so still get crushed in multiway situations.
That's back to the unbelievable part for me how can a human pick up on these tendencies and not a computer. Do you mean to imply they are social, or cultural? Hex is another game that just seems solvable to me by some sort of reduction/recursion. But I have new questions now to go after.
So games then are that type of problem that you must brute force in a sense. Well this might not be true but the games that are easy or "broken" ie solvable I think evolve. Or we can ask what are the factors of a good game. I think most games evolved through play and rule change and this is how we have these incredibly difficult to solve games that seemingly still have some finiteness to them (I probably don't use the term finiteness in the correct way).
It's all incredibly interesting to me. I need more direction to explore it though. I appreciate the points (from everyone).