Quote:
Originally Posted by venice10
My problem with general theory threads is that nobody defines what they are talking about. It took about 100+ posts before people started realizing they were talking about multiple different things.
While I'm sympathic to those trying to use statistical definitions for analysis, the problem they have is that in real life poker statisitics are not useable. One major assumption in statistics is that the pool from which the samples are pulled is homogeneous. As Mike Caro pointed out years ago, the skill level at your table is going to have a huge effect on your daily results. Longer term, the more you play the same players, the more they learn about you and adjust their play. Your pool changes over time. For example, your results are going to vary considerably if you are playing at a table full of bad whales compared to a table of the top 9 players in your room.
In terms of swinginess, the variable to consider isn't so much tight or loose, but rather how aggressive one is playing. The classic LLSNL player, the loose passive, isn't going to swing very much. They'll play a lot of hands, but mostly will fold on the flop because they missed. Their stack doesn't move much. When they win a hand, it is mostly calling down and having the best hand. A maniac is going to be pushing all hands to the limit. Most of the time, people will just fold from fear of losing their stack. That allows the maniac to build his stack. However, when he loses, he loses big.
Just some thoughts on the matter.
Wanted to get back to this because it's important.
Let's pretend we build a regression model and regress winrate on talent, stakes, opponent skill, age, amount of sleep last night, etc, etc, etc. The assumption of homoskedasticity will be violated because our dependent variable will vary at different amounts over different values of X. Mike Caro is correct about this. But it doesn't invalidate all analysis of poke statistics.
The consequences of this violation are that the standard errors for our coefficient estimates will be biased, and so our inferences will be suspect. We won't be able to say why winrate varies with X reliably.
However, the coefficient estimates themselves remain unbiased. And, thus, can still be used for prediction. We would still be able to effectively predict winrate, but we couldn't reliably test hypothoses.