vpFREE2 Forums

Risk of Ruin Before Royal in OpVP

The "Session Bankroll" feature that was added in Optimum Video Poker 1.0.8 gives reasonable approximations for the risk of ruin before hitting a royal for large bankrolls but fails for small to moderate bankrolls. The problem is that most players who are interested in session bankroll requirements are concerned more about small bankrolls.

The failure is because of the use of the Poisson Distribution instead of a direct calculation. Being based on cycle probabilities, it is not good for bankrolls less than sufficient to play at least one cycle without hitting a royal.

Therefore, the feature is being replaced with the use of the Sorokin/jazbo formula for a Risk of Ruin Before Royal function that gives an accurate result for any size bankroll. A window identical to the current Risk of Ruin feature is used, providing the capability of calculating the bankroll needed for what you consider to be an acceptable risk, or the risk with any given bankroll. Also, a slot club rebate can be included in the calculations.

My thanks to Steve Jacobs for his correspondence on this subject. He spent considerable effort convincing me that the Sorokin/jazbo formula is the way to go, but we still have not resolved the question of why my simulation gives such different results, especially for small bankrolls. I would appreciate direct correspondence with anyone who has written a simulation program for RoRBR.

OpVP 1.0.9 will be sent out to all registered users within the next few days. If you have a legal copy of Optimum Video Poker and did not receive the 1.0.8 update last month, it's probably because I don't have a good e-mail address for you. Please provide an e-mail address where you can receive a large attachment. Alternatively, I will still send a complete updated version on CD for $5.00.

I have had no problems sending a StuffitX file to Macintosh users, but Zipped files sent to Windows users are sometimes bounced. It seems that some users and/or ISP's have paranoid virus checking, rejecting any attachment containing an executable program, even though it's compressed, and in a few cases rejecting any attachment. In some cases an alternative address has worked, but a better solution is to make use of a feature offered by most ISP's and most e-mail programs. Set up a list of e-mail addresses from which you will accept anything, and include my address in that list. Since both the Windows and Macintosh applications are the direct output of the compiler, neither could possibly contain a virus. Also, it would be nearly impossible for me to send anything containing a Windows virus because I work only on a Macintosh. To test the Windows version of OpVP, I transfer it to a PC on CD, with no direct connection.

Dan

···

--
Dan Paymar
Author of best selling book, "Video Poker - Optimum Play"
Editor/Publisher of VP newsletter "Video Poker Times"
Developer of VP analysis/trainer software "Optimum Video Poker"
Visit my web site at www.OptimumPlay.com

"Chance favors the prepared mind." -- Louis Pasteur

I uploaded source code for a bust before royal simulation about a week
ago to <http://www.wildlife-pix.com/vpoker/bustorroyal.c>. There's a
graph of results for Jacks at
<http://www.wildlife-pix.com/vpoker/fbust.png> and my previous
discussion at <http://groups.yahoo.com/group/vpFREE/message/51002>.

What do you mean that your simulation gives "such" different results?
The only simulation result I remember you comparing was the bankroll
for 50% RoR - your result of 740 betting units is close enough to the
"right" answer of about 732 (IIRC). Remember when you estimate a
probability from a frequency the standard deviation of the frequency
is sqrt(p*(1-p)/N). For p=1/2 and N=10,000 the 1 sigma uncertainty in
the risk of ruin estimate from a simulation is +-0.5%.

Mike

···

--- In vpFREE@yahoogroups.com, Dan Paymar <Dan@O...> wrote:

My thanks to Steve Jacobs for his correspondence on this subject. He
spent considerable effort convincing me that the Sorokin/jazbo
formula is the way to go, but we still have not resolved the question
of why my simulation gives such different results, especially for
small bankrolls. I would appreciate direct correspondence with anyone
who has written a simulation program for RoRBR.

You're welcome. I'm glad that my efforts resulted in an improved
product for your customers, as well as a clearer understanding
of RoRBR in the VP community.

···

On Sunday 13 November 2005 11:47 am, Dan Paymar wrote:

My thanks to Steve Jacobs for his correspondence on this subject. He
spent considerable effort convincing me that the Sorokin/jazbo
formula is the way to go, but we still have not resolved the question
of why my simulation gives such different results, especially for
small bankrolls. I would appreciate direct correspondence with anyone
who has written a simulation program for RoRBR.