Deal Calculator
 

I made a Palm application to calculate reasonable deal splits:

DealCalculator

Any questions or comments welcome,
Craig

PS: Note:
Not guaranteed bug free.
Uses the independent chip model, which slightly favors small stacks, but is MUCH better than the McEvoy model. Perhaps sometime I will also include the better random walk model. Errors are typically <~1%.
Only does up to 8 players (each additional player increases time by about x5, so full tables are not really Palm compatible (my laptop is fine for up to ~18)).
Uses integers for much of the data, so add zeros on the end of Prize$ numbers if you desire more significant figures in EV.

Re: Deal Calculator
 

Thanks for doing this! I am no computer genius but I couldn't get it to open. When you say Palm application what does that mean?

Re: Deal Calculator
 

Its runs on a machine running PalmOS (most likely a PDA although you can run it on your desktop with a simulator). If you have such a device then put in your download directory (or however your PDA management s/w works) and sync with it.

Re: Deal Calculator
 

Is there any way to run it without a palm? Thanks for doing this Craig.

Re: Deal Calculator
 

Several months ago I posted links to C source code for several of these, they are at
FourWay and IC source

The code is ~= atrocious, but they should work (user interface is also worse than the Palm App).

I coded the Palm stuff because I am trying to learn the operating system for a different project, and I thought it would be nice to be able to do it live. However, there are certainly device limitations,
Craig

Re: Deal Calculator
 

Just downloaded. Looks great. Could explain the fields a little bit and the output?

Stack seems pretty self explanitory. Just enter the size of each remaining player's stack.

What is Prize $? I tried entering different values, but that doesn't seem to effect the P1 column,

Great work. I can barely turn my palm on so the thought of writing an app is awe inspiring.

Re: Deal Calculator
 

I'm glad someone was able to get it and run it.

The Prize$ column is the payout schedule (i.e. 1st=$500, 2nd=$300, 3rd=200).

EV is expected value of that player's stack given the distribution of stacks. (EV = sum PnFinish*Prize$)

P1Finish is the probability that that player would finish in that place in a freezeout among equal players. (This is only an approximate result, but a good approximation that yields exact results for all stacks equal as well as zero and gigantic stacks and for 1st place. It also never yields patently false results, such as greater EV than 1st place prize money, which some other methods can.)

Calculate P1 calculates P1Finish and EV for player1, while calculate all also calculates EV for all the other non-zero stacks.

The program reorders the list of stacks so that all non-zero stacks come before zero stacks, and negative stacks become zero stacks.

Craig