Bayes or not Bayes, that's the question....

[Shandrax]

I found an article by David Sklansky in cardplayer online and it made me wonder.

First of all, here is the link:
http://www.cardplayer.com/poker_maga...77&m_id=68

Now there is an interesting part about A-K and pair combinations. A-K can be made from 16 different combinations while a pair can only be made from 6 different combinations.

Looking into PoT (p. 229) I find a reference to Bayes theorem and that the odds of getting A-K instead of A-A are 4:3!

This is true, because of how the cards are dealt. You can make A-A in 6 different ways, starting with 4 different aces = 24. You can make A-K in 16 different ways, starting with either ace or king = 32. 32/24 = 4/3

What I am wondering about is why in the article David only used the pure hand combinations and not the possible combinations how these cards can be dealt, because if you do that you get different results. Quite obvious because Davids math is based on 16/6 while mine is based on 32/24.

HELP!!!

Sorry for my stupidity, but I am 38 years old and I knew I had this all in school, but since it was 20 years ago, I simply forgot it all.