+EV Starting Hands

[scott2130]

I searched to see if this was covered before and couldn't find it, so if it was, I am doing it again.

I came across stats for starting hands and there +EV. It starts with AA at +2.32 and goes to 32s at -.16. The statistics are based on 122,031,244 pair of pocket cards dealth at the real money tables. The unit for EV is average profit in big bets. Not a bad sample size.

I started to compare this list to recommended starting hands found in the books. The top hands are pretty much the same but as you get down the lists they start to differ. I expected some difference but some hands like 98s is ranked middle in the book and listed at 0.00 in the stats. Others that are middle in the books J9s, 87s, and 76s are either break even or below. Others like 88 are higher in the stats then the books. I have not looked at all of them yet, but was suprised to see as much of a difference that I see at first glance.

My questions are these.

Are the lists in the books based on simulations or real stats like I found? If not, what are they based on?

I got my list from http://www.pokerroom.com/games/evstats/totalStats.php?order=value
are there other such lists or simulations that I can compare? If so, where are they?

[gaming_mouse]

scott,

The discrepancy you see results from the fact that the pokerroom stats are taken over ALL players -- which includes primarily losing players.

Recommendations that you find in books like Sklansky's assume that the hand will be played by a good player, with good postflop skills. Thus many starting cards that are break even or losing when played poorly become profitable in the hands of a good player.

You should interpret the pokerroom EV stats as the minimum EV of the starting cards. Unless, of course, you are worse than the average pokerroom player, which I doubt you are.

HTH,
gm

[StellarWind]

[ QUOTE ]
The discrepancy you see results from the fact that the pokerroom stats are taken over ALL players -- which includes primarily losing players.

[/ QUOTE ]
This also leads to biases that distort the relative EV of different starting hands.

Weak and unschooled players overestimate the value of certain hands compared to other hands. Two categories that come to mind are offsuit blackjack hands and offsuit connectors. Fish play these hands more often and in more dangerous ways than other starting hands of similar value. The result is probably greater losses and lower measured EV for the more popular hands.

I'm not sure these giant databases serve any useful purpose for the practical winning player. I've never thought of one.

[tilto]

Another thing to consider (aside from post flop skill of the player) is the situation in pre-flop. For example, a good player will take position, raising, and number of people in the pot into account when they decide whether or not to play a hand. As I recall, the PokerRoom stats do separate based on position, but not based on other "hand situations." For example, playing 89s against an UTG raise with few other players in the pot is probably not profitable, and something like this might be making the EV stat go south.

[OrianasDaad]

Here's a link I found useful before I got HEPFAP:

A New Guide to the Starting Hands in Texas Hold'em Poker
http://www-2.cs.cmu.edu/People/mummert/poker/

[scott2130]

This is an excellent link and it seems to support what I found. Now I have two more questions after reading the site.

1. Do you recommend making the adjustments to the rakings
as they did?
2. If the theory they lay out is correct, why didn't S&M
make the adjustments in their book?

I guess the second one can only be truly answered by the authors, but lets theorize. If the authors read this post, please comment.

[uuDevil]

This comes up periodically. Mason Malmuth has given a detailed response which you can find in this thread, among others. (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=1370460&page=&view=&s b=5&o=&vc=1)