[vhawk01]

[ QUOTE ]
[ QUOTE ]

Exactly. Trying to PREDICT the past. Not merely describe it. Thus, EV. Using the limited data that we have from the past, predict what would happen in, say, 100 iterations of the past. IOW, what is the EV of each decision/player. Exactly the opposite of all the things you just said.


[/ QUOTE ]

This phrase has no meaning to me, "100 iterations of the past". It is not well-defined. The past happened once. If we iterated it 100 times we would get the same outcome 100 times.

Anyway, I think you kind of get my point, you just disagree. My only goal here is to point out that it's a weird thing we do, pretend we can go back in time and re-roll some dice of our choosing, but not others.

[/ QUOTE ]

Talking about the past is only useful insofar as it helps us in the future. Simply listing a bunch of facts that happened in the past serves no purpose. Its what we draw from those facts, in order to make predictions about the future, that has value. It isn't weird at all. When we play these hypothetical games, what we are trying to do is come up with theories or rules or inferences or strategies, things that will allow us to better solve problems in the future. Looking at Eckstein winning a WS, what do we learn? Well, that depends. There are many inferences we can draw from that. So we need more information. Taking all of the information we have, we can create a synthesis that has predictive power. Something like "fielding a team full of 5'7" grit-monsters = World Series baybeee" or maybe "there is tons of static and noise but gimme the guys who dont make outs and who hit the ball really hard and we'll win more games."

You play poker right? Do you ever review hand histories? Why? Is it to memorize what hand won in hand 146,795 and who held it at the time? Or is it in an attempt to synthesize all the data you have and come up with general strategies and theories to improve your game?