WTB SNG simulator

[stupidsucker]

I credit my biggest knowledge gain in limit poker to turbo texas holdem. A simulator for limit. I know that a lot of people do not like the program, but if used properly then it can be an incredible tool.

I made all of my own profiles, and used the automatic test feauture to mold my game.

If I had a SnG simulator that had simalar features I could blow the game wide open.(under my own self serving biased oppinion)

It would need to have the ability to:

Create profiles to play as programed
Log results for each individual
High speed simulation


I could easily use a tool like this and really study the finer points of a sng game.

I could test accuarcy of the ICM
I could test accuarcy of curent variance equations
I could find fine points of dos and donts


Anyone seriously up to the challenge then PM me. I would like to run some ideas by you.

[eastbay]

I have been thinking about this for some time, and have a primitive capability to do it now (that requires programming knowledge to use.)

The most difficult aspect of this is how to define strategies for play. A small domain-specific language would have to be created to do this reasonably. I have thought about how to design such a language but it is a huge task to do it well enough that the results could be useable by most players.

If you think about all the variables that go into every NL SnG decision, it is a very complex process, and specifying a realistic strategy for a full SnG is a very difficult task.

I'm kind of curious how the Wilson software works in this regard for limit play. Can you outline how the user-defined strategies work?

Say, for example, I want to define how to play QQ in the Wilson software. Preflop is easy enough, but how is flop, turn, and river play defined, with all of the huge number of possibilities for boards and action?

eastbay

[eastbay]

[ QUOTE ]

I'm kind of curious how the Wilson software works in this regard for limit play. Can you outline how the user-defined strategies work?

Say, for example, I want to define how to play QQ in the Wilson software. Preflop is easy enough, but how is flop, turn, and river play defined, with all of the huge number of possibilities for boards and action?

eastbay

[/ QUOTE ]

I just dl'ed the demo and took a look at the strategy specification.

It looks like there are several thousand parameters to a full strategy specification for a limit ring game. This is pretty intricate and I'm impressed with the level of detail.

Now think about generalizing one of these strategies to no-limit tournament play. Now instead of specifying what to do when a player bets, you have to specify what to do when he bets half pot, or pot, or all-in, or some other scheme of classifying bets. And of course it matters if that half-pot amount is a small fraction of his stack, or a quarter of his stack, or all he's got. Even for a simple scheme like that that's 9 parameters where there was only 1 before.

There's a combinatorial explosion of possibilities here, and reducing it to a tractable set of parameters is not at all straigtforward.

I think you could easily spend 6 months full time trying to devise a parameterization for full NL SnG strategy that 1) allowed for sufficient flexibility to define realistic strategy and 2) didn't require enumerating tens of thousands of parameters to complete a single strategy.

eastbay

[microbet]

I like the Wilson software a lot, but it could certainly be better.

This project would be HUGE. I wouldn't consider it unless it were part of starting a company to develop and market the product. That would definitely include getting business/marketing people in on it from the beginning. Of course there are a lot of people who do huge non-commercial programming projects.

Anyway, as for me, I don't have time now. Maybe in 6 months.

[eastbay]

[ QUOTE ]
I like the Wilson software a lot, but it could certainly be better.

This project would be HUGE. I wouldn't consider it unless it were part of starting a company to develop and market the product. That would definitely include getting business/marketing people in on it from the beginning. Of course there are a lot of people who do huge non-commercial programming projects.

Anyway, as for me, I don't have time now. Maybe in 6 months.

[/ QUOTE ]

Unless the customer is Bill Gates, doing this on a contract basis for one individual would be insane.

I agree that this is a project of company startup proportions only.

eastbay