Sklansky's puzzle results from a Non TTH simulator

The following simulations assume:


1. Tough tight aggressive game.


2. Ten players.


3. Nine opponents are dealt random hands that they will try to play well.


4. Stakes: Ten Twenty. Sorry, I was half way thru before I realized it was supposed to be 15/30.


5. Rake: Ten dollars per player per hour.


6. Played at the speed of 30 hands per hour.


7. Results expressed in terms of Big Bets Per Hour [BBPH].


Raise with AKo UTG: Win: 13 BBPH Std Dev: 11 BBPH


Call with QTo UTG: Lose: 5.4 BBPH Std Dev: 13.3 BBPH


Now for the next two tests we'll have to modify assumptions one and three

to get four Middle Limpers.


The Blinds will still be dealt random hands.


UTG and the player after him will be dealt garbage.


We'll deal players 5 6 7 8 typical Limping hands.


Players 5 6 7 8 will be forced to just Call on the First round of PreFlop Betting.


Player 9 will get garbage.


Call with ATs on Button: Win: 13.6 BBPH Std Dev: 24.6 BBPH


Raise with ATs on Button: Win: 22.5 BBPH Std Dev: 27.4 BBPH


Raise with JTo on Button: Lose: 8.3 BBPH Std Dev: 24.6 BBPH

I think there is something wrong with the way you wrote things. Your standard deviations are so large there are no statistical differences in any of your results. A a specifice example an estimate of +22.5 with a sd of 27.4 is no significantly different from an estimate of -8.3 with a standard deviation of 24.6. I suspect that you need to divide your standard deviations by the square root of the number of hands in your simulation.

I would expect simulations to overestimate AK under the gun for reasons already stated, and more importantly, underestimate how raising with ATs on the button will result in free cards and greater action when a flush is made.

He doesn't mean that the measured earn rate of that hand has that standard deviation. He means that if you were to play that situation 30 times, this would be the stadard deviation of your results.


There is a difference in stdev of a hand and the stdev of the measured earn rate of the hand.

Is "Erin" is a girl?


I rarely see posts from girls on here. If Erin is a girl, and using a sample of 2 by adding Angelina, can I assume that non-Asian girls all approach the problems of poker - and the use of simulators - in a similar way?


Is it safe to assume that girls play relatively non-reactively, and may play either slightly to the left or right of the dynamic of the game - without changing gears - for hours?


Girls seem to call and lay down by less changeable rules than guys. Anybody else notice that? Can I use the errors typical of simulators - and of people who put a lot of stock in them - to tweak the likelihood a girl is making one type of error versus another in a particular situation?


Sklansky would say I shouldn't worry about girls, because I encounter them so rarely.


I'm not saying girls don't call pre-flop like rocks, take an aggressive bluff shot on busted hands in the presence of weakness, and generally bluff with approximately the correct frequency. But they never bluff-reraise, or something, there must be some common error I can't put my finger on. Maybe it's that what their bluffing frequency or situation is is an easy read, because they call too tight so you more often know what they're holding when you don't see their cards...


Anybody want to rile these girls up by declaring some predictable error that they all share in common?


(And always will:)


- the guy

"Raise with JTo on Button: Lose: 8.3 BBPH"


Would you run it with JTs under those conditions and see what happens please.

Let's say you play the same hand 12,000 times and add up your results,

both good and bad, every 30 hands.


At 30 hands per hour that would give you 400 data points.


Then take the std deviation of those 400 data points to get the std dev per hour.


Is this unsound?