[ QUOTE ]
I'm an engineer, and a data driven guy. So based on the data I have available, it seemed like a good laydown. It's kind of Bayesean thinking, but given that KK hasn't been paying off in the past, and all the action preflop, it seemed like the prudent thing to do.
[/ QUOTE ]
I'm an engineer, too, so I understand your concern for having hard data. I spend a fair bit of time writing little scripts to extract data from my Poker Tracker database to analyze my play.
[ QUOTE ]
I guess what it really boils down to is do you base your decisions on what the "theoretical" results should be or what "actual" results are.
[/ QUOTE ]
Well, you need to take into account two things:
<ul type="square">[*]How much reason you have to believe that the theory is accurate[*]How many samples you think you need for the actual results to converge to the theory[/list]It turns out that a deck of randomly shuffled cards can be modeled
perfectly. We can compute exactly the probabilty that you will win with a given set of hole cards, assuming no one folds (including you). So, we have very strong reason to trust the theoretical results.
As for your empirical data, you have only 14 data points, and you had 4 wins (4/14=28%). Lets say you flipped a coin 14 times, and it landed heads only 4 times. Someone else flips the coin 1000 times and it landed heads 501 times. Do you really think you should conclude that the coin is weighted... only when you're holding it?