Week 3

Chapter 3: Using All the Information: Estimating Parameters and Bayes' Theorem

Ok, here it is the last chapter on the basics. Bayes' Theorem is quite advanced actually, but it is what we will need later for exploitative play. After just a few trials, samples regress to the mean so quickly that you can use this against your opponent quickly. Gathering tendencies of your opponents and being certain of them to a degree will make you play optimally against them.

Anyway, here are a few topics to discuss:
1) I hope you have read or heard about the Monty Hall problem. That is the holy grail to understanding Bayes' theorem. If you don't know about it, read about it, understand it!
2) I found a kind of interesting article on Bolt's supposed steroid use on 2+2 invoking Bayes' theorem. I thought it was interesting application (Scientific American article).

Anyway, if something is way off the mean, first start doubting the test that assigns that value, then the value itself. Do you think that could be a nice summary of Bayes' applicable to poker?

Oh, and week 3 assignments are here.

Week 2

Chapter 2: Predicting the Future: Variance and Sample Outcome

Hey peeps, very late, but finally here, a disussion on Chapter 2 of the book. The assignment for this week can be found here
Some discussion obviously I want to hear your opinion on it for my 1k post :)

Do you think the distribution of poker winrates follows the central limit theorem? After all, they are not really random.

How do you deal with having periods where you play on tilt affect your WR/variance, or do you think that is part of your poker distribution and that it all "evens out"?

Comment on Week 1

Hypergeometric Distribution:

Wiki page

Just a little comment on probability distributions in poker. This is the most important one when working out stuff in poker, because you can't just put cards in the deck, once they are gone or used, they're gone and all probabilities are worked out considering that.

A couple of weeks ago we had a post dealing with "flushdrawitis" and people being very scared that when there was a "flush draw" on the board. Is that MUBS. Well, we can work out the probability of getting a flush draw, basically getting a 4 suited in a 5 card draw.

f(4;52,13,5) = (13 choose 4) (39 choose 1) / (52 choose 5) = ~10.73%

Don't get distracted by which suit, it's not important here. So if you are up against one opponent (assume you have some random cards), he will have a flush draw about 11% of the time -> not scary, but agains 4 opponents... well, you be the judge.

Solutions for week 1 have been sent out. I'm working on week 2 assignments right now. Should be up tonight. You can still, at any time send in assignments for previous weeks. There is no "late" here.