RNG Questions

[olds442jetaway]

Many years ago nearly 20 at Foxwoods, I was betting 5 quarters on DDS machine. I was dealt a Royal good for 5k. A month later on the same machine, I was dealt a Royal again. The next day, that machine was gone never to be seen again.

[FAA]

Outfoxed no longer!

[FloridaPhil]

I believe that "software updates," when performed, are easily tracked via the machine's memory.

For sure. Modern slots store all kinds of information. Years ago I was standing near by when a manager and an engineer had the administrator's screen open on a multi game machine. I could clearly see the results for each game listed on the screen. I remember distinctly that the Keno return percent was low. This was a long time ago. Most newer slots are networked, so they can retrieve this information from the server. The casino knows the return for each game in each machine. If a machine doesn't perform up to their standards, it is up to the casino to decide what to do about it.

I do not know what or how the regulator's software test works or what it tests. I suspect the RNG control code is a standard chip set of some kind. This would be necessary or the regulators would have to carry a different test program for each type of machine. It would be very interesting if the owners of this website could have a technical representative of a machine manufacturer address this issue for us. I'm pretty sure they are at least partially behind or sponsoring this website. There are a few questions I would like to ask myself. It would go a long way to reduce some player's suspicions.

[olds442jetaway]

Agreed.....especially when I can stay about even playing forever on wolf vp, and continually lose my shirt at casino vp. Both 99 percent identical machines. One would think 20 million lifetime hands should represent long term and there would not be these differences. By lose my shirt, I mean consistently a 2 percent difference over the years with the same paytables

[FloridaPhil]

Agreed.....especially when I can stay about even playing forever on wolf vp, and continually lose my shirt at casino vp. Both 99 percent identical machines. One would think 20 million lifetime hands should represent long term and there would not be these differences. By lose my shirt, I mean consistently a 2 percent difference over the years with the same paytables

Let's look at this objectively. I assume the odds on these games are derived from mathematical computer simulations? We plug in the constants and view the results. In order for these simulations to have meaning, all factors must be consistent throughout the simulation. The bet must remain the same. The odds must never change, no errors can be made and the player never runs out of money. Sounds good.

The problem is no human gambler plays video poker that way. Changing up your bet is a perfect example of this. Unless the change is done at a predetermined and consistent point, a simulation doesn't know when to do it. This is especially true if a bonus is involved. Even without it, how can you ever hope to create a simulation that plays like a human player? The only way this will work is if you assume a player plays the same game exactly the same on every hand with zero errors forever. Good luck with that idea.

Computer simulations are not real casino play. I have no argument with the math. The problem is with the way the game is played. I believe if we played the same game perfectly and constantly exactly like a computer, in time our results would mimic the predetermined odds. I don't know about anyone else, but I'm not a computer. I'm a human player. I make mistakes. I change my bet. I lose my nerve and I run out of money. Every once in a while, I sneak over to the high limit room. That practically never works, but when it does I remember it and I want to do it again.

Playing the best games as computer perfect as you can is the best strategy we have. It will not now or will it ever guarnatee a profit as long as there is a human in front of the keyboard. If you want to argue that someone can successfully mimic a computer forever, that's a different discussion entirely. Sounds like the basis for a new super hero movie to me. Video Poker Man! Anyone?

[Gronbog]

The problem is no human gambler plays video poker that way. Changing up your bet is a perfect example of this. Unless the change is done at a predetermined and consistent point, a simulation doesn't know when to do it. This is especially true if a bonus is involved. Even without it, how can you ever hope to create a simulation that plays like a human player? The only way this will work is if you assume a player plays the same game exactly the same on every hand with zero errors forever. Good luck with that idea.

Changing your bet is a meaningless example of this. Regardless of how, when or why you change your bet, the expected result is coin-in multiplied the return of the game. All changing your bet does is change your total coin-in. Over time the results will still end up as predicted.

Having said that, although it is not necessary, betting patterns can be simulated heuristically and simulators can also be programmed to commit errors at whatever rate you want.

[FloridaPhil]

OK. Let's say you play single coin quarters all day. You get the itch to take a pot shot and you play $20 at max coin dollars. You hit a royal for $4,000 putting you ahead for a year's worth of play. How do you simulate that?

Or the converse, you play max coin quarters for a year. You run out of money and play single coin quarters. You hit a royal for $62.50.

or

You play max coin quarter Deuces Wild all day. You switch to Bonus Poker at 3:00 PM. You hit a quad deuce for $50 instead of $250. Can you create a simulation that can account for that? Do you add up all the simulations to get the expected return? You can't. A computer does not know when a human may change the game or the bet.

[SlatRGatr]

You are trying to simulate a predetermined result? What is the point?

A simulation would be used to determine the odds of that happening. You would set up parameters... for example, play X hands (all day) and then increase bet to $20. Then depending on what you were looking for you could say "what are the odds of hitting a royal on that $20" or "what is my average bankroll after X hands at the $20 level" etc.

For the second example, your simulation could say "if your bankroll drops below a certain amount, switch to single coin." Then "How often will I hit a royal at this single coin level before running out of money" or "How often will I win enough to switch back to max coin based on ___________"

[FloridaPhil]

You can't write the code because you don't know what the human player is going to do when. Your example doesn't work because the player may switch at any time he/she feels like it. There is no logic or consistency to what a player may do or when he/she may do it.

Expected returns are based on a perfect non real world scenario calculated into infinity. We don't play video poker in that world. If you can find a player who plays exactly like a computer forever with an unlimited budget, the expected return is accurate. I don't think that player exists. If it does, he/she is a very rare individual.

I am not disputing the math. I am putting forth my reasons why our results may vary from the expected.

[olds442jetaway]

My only point is this and nothing else. That is why I listed a percentage return difference instead of dollar amounts. In wolf vp, once I make an error that is it. There is no ability to correct it the way I have the game set. To me over a twenty plus time frame, 20 million hands, and having results vary 2 percent every year is significant. Does that mean the machines are rigged? Not necessarily. This is what I think happened and I hope I am correct. When my bets got very high playing for real, I have been very unlucky. That in itself could account for the 2 percent difference. Playing Wolf vp, your bets are capped at 5 coins. Many times over those 20 plus years, I have tried max bet several times a session and never won anything big. There were a couple of exceptions a few years ago, but even then they were not max bet wins. If that is not it, I have no choice but to believe real life gambling produces different and larger losses percentage wise than rng simulation programs.