yesterday , I made a thread:

http://www.fullcontactpoker.com/poker-foru...pic.php?t=48912

Copernicus and I debated on which strategy would be the best.

Copernicus argued that:

Being all-in for your whole stack early on for coinflip and doubling up and then only being all-in for 1/2 your stack thereafter on coinflips would be the clear winner.

If I don't have the strategy right, please correct me. I would think that you would usually have to at least double up for your whole stack first since it's unlikely that you will get a chance to be all-in for 750 when you only have 1500 (your starting stack).

I argued that:

Whenever you are in coinflips, always double up for your whole stack each time and pass up opportunities where you can't double up for most of your stack.

So using what programming I know, I wrote a program that would simulate these two theories. My strategy won out each time. I ran each scenario 100,000 times (100,000 tournies) and my strategy reached 23,000 chips or more 1000 times more often on average than the other strategy.

With the 1/2 strategy, you really have to pick a cutoff point at which you will now be completely all-in , right? Maybe I picked the cutoff point too high. I picked the cutoff point at 1500 which was your starting stack in the 180 man SNG's. I'll see what I find with a lower cutoff point.

If anyone wants the program, I can give it to them. Or if you want to see the output, I can get you that, too.

Yes, send it to me, since my spreadsheet came up with very different results. paulsgoalie@aol.com

A difference may be that I modeled it as HU play..ie the bet was a function of the smallest stack, not just the players stack. I dont recall how often I increased blinds, but the bet would of course be all in if it was less than the blind.

I didn't take into account the blinds. I don't think that is necessary. I wrote it in C++ . I'll send you the program, the output and the source code here in a little bit.

You know, I didn't take into account the blinds but you are right that the blinds are another factor. But the blinds would only be a significant factor in this modeling theory if you were already shortstacked and in that case, you would be all-in no matter what. I don't know how you were modeling blinds,etc on Excel but my program is very simple.

In general, though, I would think that you would want to be in coinflips as few times as possible, so therefore getting as much in preflop as possible would be the best way to go I would think.

Ok, I just sent you the program. Let me know what you think. It's a really simple program.