I don't understand that after a year of doing the hot table drawings it has not been realized that the RNG that the bravo software uses is an ineffective way to get a random drawing. The probability of so many repetitive drawings happening has far surpassed a matter of coincidence. Especially right now where the array is not changing because bravo is not being used. There should not be back to back duplicates every day the drawings are held. The has to be a better way to do the drawings if the company/developers of the bravo suite don't want to fix it.

The odds of the same table hitting 3 consecutive hours is 1:8000. Getting a lottery drum or some other solution would be a far better solution.

Sent from my SM-N950U using Tapatalk

How did you come up with that number? Could you show your work?

Math is hard, but assuming 20 tables. 20x20x20

Sent from my SM-N950U using Tapatalk

Actually it'd only be 20x20, or 1:400, since any table that hits simply needs to repeat twice.

It is 1:8000 for any single table to hit 3 times.

Sent from my SM-N950U using Tapatalk

Only for a specific table. If any of the 20 tables needs to hit 3 times it's 1:400. If you're going to be condescending with "math is hard" comments you might want to use the correct math.

Every random number genrator ever made has had some math enthusiast come in and declare it broke with some whack math.

I didn't reply to you did I. I wasn't the one that decided to find an interpretation of my post that would make it incorrect. The message was clear that the **** isn't random and whether you want to exclude the odds of the first drawing or not is irrelevant to the discussion.

Sent from my SM-N950U using Tapatalk

That may be. But the odds of 1 out of 39 tables hitting on 3 consecutive hours are far lower than that.



Think it again. We are not talking about let's say table 1's odds of hitting in 3 consecutive times. We are talking about one of of all the active tables. It's far less than 1 in 8000.

I think my point is being missed and that is fine. I have benefitted from this (I have actually been at a table that hit 4 consecutive hours which is at least 1:8000) but I hear the grumblings at the tables and think that there are better ways to do it. I did not make the post to argue about what the odds are for any table vs YOUR table.

I have seen enough of the drawings to know that the amount of repeat drawings does not reflect actual randomness. I understand that a poker room management software company would not invest much time or money into a quality RNG which is why I was suggesting a different method.


Sent from my SM-N950U using Tapatalk

Ok, but what we are telling you is this. You gave us a number that showed that an event that seemed to happen a lot shouldn't happen as much as it seems. The first answer is that the event you described isn't as improbable as you said it was.

The second answer is that there is a selection bias in play. What looks as if it's abnormal tends to stick with people even though there's nothing necessarily out of the ordinary about it.

My third answer is this. When I have to choose between the possibility that a computer RNG is faulty or the possibility that a human has a faulty perception of randomness, my guess is that the second possibility is is infinitely far more likely, because the human brain isn't really equipped to deal/comprehend with randomness. After all, the human brain through evolution is designed to find patterns even when those patterns don't exist.

Having said that. Is there a possibility that the RNG is faulty? Maybe, could be. I am not qualified to answer. But it's far more likely that the people grumbling that the RNG isn't really random are probably at fault.

I suggest reading a book called the Drunkard's Walk or Taleb's Folled by Randomness. Both books deal with precisely those issues.

dont tap the glass my friends.

nvirish, seems like youve found a pattern here, you should table change to the table that wins the previous drawing to increase your odds!

You guys are so smart.

There is precedence. Washington state gaming commission determined that the genesis gaming RNG could not be used for lottery drawings in their state.

Sent from my SM-N950U using Tapatalk

Or, as I learned in school, no matter how improbable, if it is possible, it can happen.

Once it does, the odds for that discrete event to occur were 100%.

Seems to me that the RNG in Bravo is flawed and it is dependent on 2 elements. one is the number of tables in use and the other is the number of seconds that it runs.
So if the number of tables does not change and the number seconds it runs does not change it keeps selecting the same table. The only time the table selection changes is when a new table is added.
To remedy this, if the number of tables does not change, is to increase the number of seconds the RNG runs at each and every drawing.
I hope this is as clear and I wanted it to be.

All of the subsequent comments to you express very well the misconception you make that produced that ridiculous number. Math does seem to be hard for you. I majored in math and I think I have a much firmer grasp on probabilities than you demonstrate.

Much like finding a marked card that nobody else believes is marked, the easiest way to convince people is to start calling it out.

Five minutes before the "random" selection, go up to the floorman and tell him what table is going to be selected. And it will knock his socks off when he realizes that you've called it correctly.

Do it three times in a row and I guarantee you they will yoink the system before you do it a fourth time.

Alternatively, make some money off of it - offer the floorman even money on a guess. You guess the 1 in 20 table, he pays you $100. You miss, and you owe him $100. See how much financial pain he can endure before admitting that you've got the system solved.

I got 20 out of 8000. There are 8,000 ways to pick a set of 20 objects 3 times. Only 20 of those instances we care about, the ones where they all match, so our 20 out of 8,000 reduces down to 1 out of 400.

Let me make sure I understand what is being said here. there are hundreds of ways to create effectively random numbers from a psuedo-random generator (randomizing against the clock, for example). You are telling me that your theory is that, in this day and age of mature and advanced computer applications, you think that Bravo has written code so simplistically bad that it picks the same table 3 times in a row consistently?

I think it would actually be much more difficult to have an RNG with such a specific skew.

Could it be possible that this is just a case of confirmation bias, and that you ignore all the negative outcomes that do not support yourconjecture?

No, the likelihood of the (any) same table hitting twice is 4%. The likelihood of a specific table, say table 10, hitting twice (a priori), is 0.16%.

And this assumes there are only 2 drawings. If there are more, the probability of any table hitting twice in a row at some point in the day is even more probable.

You're right. I realized while in the shower I erred and deleted my post immediately upon getting the chance. I'll rethink this.

I should never post before coffee...

Same and specific may not be equals until after a first successful drawing. When I play there is always a 'hand of the day'. Yesterday it was K4 .. showed up many, many times. I'm still waiting for the day when AA is the 'hand of the day'. (Yes, I realize that there are 16 combos of K4 and only 6 of AA).

Hopefully random does, in fact, mean random. Is it amazing that the slot players win 'way' more drawings than the players in the poker room? This is tilted by 'entries' ... It's possible that the Bravo RNG is influenced by something else in the logs and leaning more towards the player's table with those 'other' factors influencing the choice. GL

PS .. Why are 3 and 16 'always' showing up in the history of a roulette table and yet I never play them (or the game period)?

Random is not binary. There's a scale.

Let's say you're timing a marathon. Is the last digit of the number of seconds random? Now, let's say you're timing the 100 m dash. Is the last digit of the number of seconds random? The former is way more random than the latter.

You can generate numbers that seem pretty random from pseudorandom methods - for example, taking a number between 1 and x, multiplying it by y (where y and x are relatively prime and y is large relative to x), and taking the remainder when diving by x. The sequence of number is entirely predictable if you know the seed (first) number but it's probably good enough for some applications.

Not everything needs military grade, cosmic microwave background radiation level randomness. I believe that the Excel RNG has a flaw in it, see how long it takes for you to guess the next number.

I concur with this. Randomizing against the system clock, or hell, just using the psuedo random numbers generated is more than ample for the vast majority of applications. While I think that over very large sample sizes, certain applications of psuedo random numbers may develop a bias, I absolutely, 100% do not think you are going to see a discerinble bias with any predictive value in a function designed to pull a random number between 1 and 40. Yes, you may see clustering, but that is not bias, but a function of variance.

I will trust psuedorandom number generators more than I trust hand shuffling or machine shuffling. And I trust hand and machine shuffling just fine.

Machine shuffling is using PRNG. At least shuffle masters are.