KEN, LEFTY, HOPINGLARRY (any others that want to weigh in.) At first blush, or through a certain time of video poker play, is there an indication of integrity or lack thereof in a VP machine, or for that matter a bank of them in a particular casino? The question is made beyond the basic understanding of the per cent return of the machine. But, even then, what verifies that within the per cent return there is not a manipulator, altering the stated frequency.
I don't know other than perhaps state statutes. I have been playing video poker for a little over 6 years at Choctaw Durant. I track my hands and have played about 2.2 million there. In the early days of this I have tracked the number of 4 of a kinds I got for about 400K hands out of curiosity. I was within a few of the number I should have gotten at one every 424 hands playing Jacks or Better. I also for a couple of years tracked how many 4 card starting royals I was dealt. The statistical amount is about every 2764 hands. I was at 2771. This was about 800K hands. I did this because when I started tracking stuff I did not have a royal flush for my first 155K hands, so I wondered. My royals caught up with statistics last year and this year they are backing up again, but I believe the machines are honest.
The simple answer is - Yes. And no. Nearly all VP machines are made in Nevada and must deal randomly upon leaving the factory, remaining that way if they stay in Nevada. Most other jurisdictions require they remain random as well. A casino has little motivation to rig the game because the few extra dollars they make will be dwarfed by their loss of action if word got out to the public that they are cheating. However, the machines you sometimes see in bars, restaurants, and membership organizations (think Moose or Elks Lodges) as well as some tribal casinos are NOT random at all. Some are bingo-based and do not deal randomly, it has already been decided what you're going to win as soon as you deal a new hand, much like a slot machine. If it has decided you are going to get four of a kind and you get that on the "flop" then you can throw it away & it'll give you another 4OAK on the draw. Or so I hear, I wouldn't think of playing those. P.S. Rigging VP machines can be done but it's extremely rare as well as foolish. The American Coin slot route scandal from the early 1990's is one of the most notorious: https://lasvegassun.com/news/1997/nov/12/man-wanted-in-murder-of-slot-tech-arrested-in-ariz/
Thanks Hopinglarry and Lefty. While I'm not a strict tracker like Hopinglarry I am aware of the statistical frequency. And as Lefty stated I am a believer in overall integrity. My question was more to the aberrations indicated by Lefty. And my question was stimulated by two recent sessions(at two different locations) where the statistical frequency seemed to me to be way out of wack. At one of them, I was one of three players whose variance was brutal. And okay, I don't play perfectly ALL the time. But since I began playing VP seriously, less than 5 years back, I have been rewarded well within the frequency, until the two sessions I referenced. Chalk it up to the long run? Perhaps. But in one session, drawing to 20 four card flushes without making one, getting no higher than a straight, in a two hour period, raised, in reflection, my concern.
You can't draw conclusions from small sample sizes. Rare events or sequences of events are part of all true random processes. Being unlikely doesn't mean that these things will never happen. It means that they should happen at the rare frequency that is expected. It also does not mean that rare events can not be repeated within a short time span or happen in the sort term. There is also an expected frequency associated with those. The same applies to hitting a dry spell for events which are considered likely. How small a sample size is too small? For something as diverse as video poker you needs millions of samples (hands played) before you can draw any statistical conclusions. On the other hand, if you never saw anything you thought was rare or surprising, that would be an indicator of a non random process.
Gronbog hit the nail on the head. Small samplings are the bane of statistical analysis. The larger the sampling, the closer the observed results are to the true value. Smaller samplings may exhibit large variances. There is a mathematical method of dealing with large observed variances from the true value known as a "least squares" analysis. Put simply, if a result deviates significantly from either the great majority of results or the calculated true value, a "least square" calculation may allow that particulate result to be discarded from the overall analysis, bringing the rest more into line as they should be. While only loosely applicable to the discussed video poker results, the principle of discarding a result, or event, still applies when determining the accuracy of observed values versus the true value.
Gronbog and Dakota are right, no need for me to expound upon their posts. As far as rigging a VP machine, one would have to be certifiably ignorant to rig common occurrences like flushes. The very few times it has been done, it's been to the rare hands - principally Royals. The statistically meaningful sample size is so enormous that it can go on for years before anyone catches on.
Great input by knowledgeable people in this thread. Bottom line is, bad variation sucks and it DOES happen from time to time. Same thing happens in any game of chance/choice. Billy C
