[Buzz]

[ QUOTE ]
I admittedly am not familiar with your simulations.

[/ QUOTE ]

Niss - The Wilson turbo-Omaha-8 simulator has a screen with blank spaces to put cards and with all the cards in a full deck shown below from which to click, hold, and pull any card to the blank space where you want that card. If you don’t specify any card for a space, the simulator presumably deals a random card to that space.

There are five blank spaces for the board, and then four blank spaces for every player in the game. If I stipulate nine players for the simulation, seat #1 disappears, leaving seats #s 2,3,4,5,6,7,8,9,10. I generally pull Hero’s cards to seat #10, and then sometimes (depending on the simulation) drag some cards to other spaces.

[ QUOTE ]
My question is, does your data derive from thousands of runs of 8 random hands that always play through to the end?

[/ QUOTE ]

I specified 10,000 runs for each of these simulations. (10,000 separate deals apparently with different random cards distributed to the blank spaces each time).

The opponents I chose play all the hands through to the end. Wilson has about fifty or so other characters who play according to certain specified parameters, some tighter than others, some more aggressive than others, some more tenacious than others. You get different results using different line-ups. For the results in his book, Bill Boston used a ten-handed tough, tight, aggressive line-up. I could have done that too, or I could have chosen a line-up of other various opponents with other various characteristics, some loose, some tight, etc.

Instead I want to know how each hand would fare if the opponents stayed to the end, rather than folding for unfavorable situations. I don’t generally want simulation opponents (who actually would end up beating Hero if they had stayed) folding prematurely. And I wanted to see how each test hand would fare if Hero stuck it out to the end. And I used a nine player line-up rather than a ten player line-up, because nine players is standard for a full Omaha-8 table in the Los Angeles area.

[ QUOTE ]
Becuase I wonder if that is not a realistic way of considering 3-4-5-6. In reality, if someone plays at you, chances are you are either behind or looking at 50% at best. If you are not behind, you're probably looking at a small pot. So I wonder if valuing 3-4-5-6 based upon simulations of thousands of random hands that don't fold is appropriate.

[/ QUOTE ]

Good point. Obviously in a real game neither you nor your opponents will see every flop and stick it out to the end. But each table at which I play has different individual opponents who play somewhat differently - and at any particular table there’s a gradual exchange of opponents, so that the dynamics of the game continually change. It’s impossible to know what to expect from an unknown group of opponents.

At any rate, these particular simulations were for each of the specified hands (specific cards dragged to seat #10) playing eight non-folding opponents with unspecified cards. The Wilson simulator presumably dealt random cards to all the opponents, also dealt random cards to the board, and then tabulated the results for a number of data screens. It did this 10,000 times in an amazingly very short time.

However, as you suggest, the simulations are clearly not realistic. They only indicate how often a hand would win (and with what) if nobody folded. In real games people sometimes fold and for various reasons.

Everybody knows that ace-deuce is the best two-card combination for low, followed by ace-trey, deuce-trey, and ace-four. Almost all players in my games will see the flop with almost any holding that includes any of these, especially the first two.

So starting out with three-four as the best two-card low combination in your hand, you’re probably up against at least one opponent, and probably more than one, who has a better starting low combination than you. If you insist on not continuing with three-four for your low unless there’s an ace plus a deuce on the flop, I don’t see how you can play a hand with a three-four combination unless the hand also has an ace or a deuce (or a pair of kings or queens plus some suitedness). Thus if that’s the way you play, then 3456 is (logically) unplayable as a starting hand, even if double suited (except from the unraised big blind).

As knoll wrote in his response to Jeff’s post about the 3456 hand: [ QUOTE ]
Not much to be said here. You know you made a questionable preflop call, got trapped, then got sniped. Your play was fine on every street, just got an awful river. But as we all know that's omaha.

[/ QUOTE ]

In other words, Jeff played a skimpy starting hand, and then got out-drawn on the river by someone who didn’t have proper odds to see the river. However, that could happen with any hand, even a premium starting hand.

I’ll agree it’s less likely to happen with a premium starting hand, but that’s reflected in the higher simulation results for a premium starting hand.

I thought your comparison of a 3456 hand with a 9TJK hand suggested an interesting way to perhaps put non-paired, aceless high only hands in perspective. We all know that 3456 is not a very good starting hand, but there seems some lingering doubt about non-paired, aceless high only hands.

(I’m not suggesting anyone should play or not play either the 3456 hand or the 9TJK hand. Not that it matters, but I’ll play or not play both of them sometimes, depending. When and if I do play them, I like them to at least be single suited.)

Lastly, a few other comparisons:

hand.....high.....low.....scoop.....total
9TJKs-.....478.....0.....710.....1188
3456s-.....304.....593.....345.....1242
6789s-.....458.....9.....243.....710
9TJQs-.....514.....0.....704.....1218
AA23s.....314.....1599.....1016.....2929

Except for the premium hand, the 3456 hand seems the best of the lot, winning slightly more than it’s fair share (which would seem to be 1111.11 in a 9 handed game) while the hand with all middle cards is clearly the worst.

Buzz

just a note regarding the high and low columns:
The high and low totals shown evidently represent half pots divided by two plus quarter pots divided by four, etc. Thus 9TJKs- could win a half pot either (1) by tying for high with no opponent qualifying for low, or (2) by not sharing high but splitting with a low hand. We could probably come up with a pretty good estimation of how often 9TJKs was tied for high and how often it split with low, but I don’t know as it matters much. A half pot is a half pot, whether you share with high or low.