This may have been covered in the past, but my memory leaves me cold (seems to be happening more and more these day…UGH!)
What are thoughts about using the “auto-play” option of VPW for long term simulations (10’s to 100’s of thousands of lines of play), to faithfully represent “real” casino results?
Thanks.
… bl
Any program likely relies takes advantage of an RNG algorithm coded into the computer’s operating system. So you’re not relying on the program developer’s expertise so much as a thoroughly vetted operating system.
RNG algorithms undergo serious statistical analysis. These days, RNG concerns lie chiefly with the prospect that someone might purposely manipulate RNG output, rather than with the RNG results themselves.
In short, there’s good reason to trust Video Poker for Winners when running auto-play simulations for personal use and interpretation.
—In vpF…@…com, <bornloser1537@…> wrote :
This may have been covered in the past, but my memory leaves me cold (seems to be happening more and more these day…UGH!)
What are thoughts about using the “auto-play” option of VPW for long term simulations (10’s to 100’s of thousands of lines of play), to faithfully represent “real” casino results?
Thanks.
… bl
Thank you.
What gave (gives) me pause is the fact that most of these “programs” (VPW included) are usually touted as “training” programs, designed to enhance a user’s ability to play “error-free” in a real casino environment. VPW even allows the user to select a “Training Style”, within which one can select a “Level of Difficulty” (i.e., Mixed, Beginner, Intermediate, Advanced).
My “wondering” revolves mainly around how simulation results may vary when choosing different “levels of difficulty” during individual long-term simulations in “auto-play” mode.
… bl
bl writes: “What are thoughts about using the “auto-play” option of VPW for long term simulations (10’s to 100’s of thousands of lines of play), to faithfully represent “real” casino results?”
Different pseudo-random number generators. And “thousands of lines” is generally nowhere near enough, you generally need millions. A fun test: do a run of 10 thousand plays on VPW from reset and record the number. Shut down the app and restart it and run again: what do you get? Same number? You will get the same number if VPW resets its prng to the same seed every time and doesn’t randomize the draw selections. Do a web search for “problems with random number generators” for some more details. Also check out: “mersenne twister random number generator”. I don’t know what algo IGT uses in their slots, I’m sure it’s an industry secret to avoid exposing exploits.
also check out: “russians engineer a brilliant slot machine cheat and casinos have no fix”.
For fun (in response to nightoftheiguana2000’s message)…
Running 15 simulations of VFW Spin Poker at quarters, each run 6,667 hands, starting with $6,000 (mixed difficulty):
Simulation
Bankroll
Final Amt
1
$6,000.00
$10,046.25
2
$6,000.00
$3,367.50
3
$6,000.00
$5,642.50
4
$6,000.00
$3,313.75
5
$6,000.00
$3,192.50
6
$6,000.00
$3,315.00
7
$6,000.00
$4,211.25
8
$6,000.00
$3,897.50
9
$6,000.00
$5,790.00
10
$6,000.00
$5,335.00
11
$6,000.00
$1,193.75
12
$6,000.00
$3,313.75
13
$6,000.00
$3,427.50
14
$6,000.00
$2,628.75
15
$6,000.00
$4,892.50
There must have been a RF or two during that first simulation , I would think…LOL
… bl
Small sample size for any Monte Carlo type simulation yields lots of Standard Error in the results because of the high standard deviation of Video Poker and payoff skew that comes from the Royal Flush and bonus features. When we run a simulation for Blackjack, despite the fact that 21 has mostly even money payoffs it still takes 100 million rounds or more to get the EV and Win% to settle on a number and still the Standard Error is 0.03% to 0.05% and the exact edge of extreme counts is unreliable due to the small number of hands in those buckets. Now that simulators are faster 400 million rounds is the new “standard”.
“What are thoughts about using the “auto-play” option of VPW for long
term simulations (10’s to 100’s of thousands of lines of play), to
faithfully represent “real” casino results?”
. Really, I am just playing around, having nothing better to do. I am not involved in anything theoretical, in which I want to get “accurate” results for some high-brow gambling treatise. I was just running a few simulations to get a “feel” for what happens during 6,667 hands of $0.25 Spin Poker without being at a casino. Dunbar’s short-term ROR calculations say that I seldom should lose $6,000. Yes, i only ran 15 simulations, and did not lose $6,000 during any one of those. I know this does not mean that I will NEVER lose $6,000. But, I guess this sort of confirms that it should be rare. (And, yes, I know about small number statistics, or at least I read about them.)
To that end, in my first posting on this subject, I was just wondering how well the VPW auto-play option might mimic what really might happen.
(Maybe I should have stopped at Simulation 1…LOL)
… bl
Any practice mode that alters card selection for a dealt hand from a “random” deal would be invalid for any type of simulation run (where you look to parallel casino play outcomes).
The “mixed level” is simply another way of saying that the card frequency is unaltered from a random deal. The other two difficulty modes (beginner/advanced) filter dealt hands so that the EV difference between the best and 2nd best holds are above/below a desired threshold. These modes aren’t random deals.
—In vpF…@…com, <bornloser1537@…> wrote :
My “wondering” revolves mainly around how simulation results may vary when choosing different “levels of difficulty” during individual long-term simulations in “auto-play” mode.
Thank you.
… bl