RPP algorithm from someone who knows what he's talking about

Poker Discussion
164

Wow! A zombie thread … this should be fun!

Having now read the OP, I have to say that it was very well written. There is one consideration, at least, that @spivak didn’t consider though.

Anyone who has read my ramblings on other threads knows very well that I do not believe and strongly oppose people who do believe that RP is rigged.

Having said that, it occurred to me that writing a non-random PRNG would be extremely difficult but simply using a deck of 62 cards, for example, rather than 52 would be trivial.

Since it is predominately the low stakes players who think RP is rigged, we will assume that the algorithm is something like “if low stakes, use tricky deck”. See how easy that is to write? lol

Our “tricky deck” will be a normal 52 card deck plus, say, 3 aces, 4 kings, 1 queen, 1 jack, 1 ten. This will comfortably allow for AA getting cracked more frequently, more straights, more high full houses and so on.

No need to mess with the PRNG or dealing algorithm.

Et voila, more “exciting” games on the low stakes tables.

Regards,
TA
(who strongly believes that RP is totally random and completely fair!)

2 Likes
165

Then, how would we handle it when someone’s hand held 2 or 3 Kings of spades? It would simply have to happen. Personally, I’d run for the door, screaming all the way.

2 Likes
166

:crazy_face: I’d be screaming, “Witchcraft!” when I should probably scream “Rigged!” I lay this at the feet of them that raised me.

3 Likes
167

Ah! That’s a good point … since I was being whimsical I was quite obviously a bit too lazy with my thinking! Oopsie!

However …

This could be overcome in the dealing algorithm, still leaving the PRNG alone, but I don’t know what the computational cost would be since we would have to check every hand for illegal combinations.

The first work-around that occurs to me would be to have the computer scan ahead a maximum of 23 cards ( = 9 players x 2 hole cards each + 5 community cards). A potential 5 of a kind would be managed by simple substitution. That is, discard and redeal any card that might cause 5 of a kind. Provided we weren’t too heavy handed in salting the deck, this will be a rare occurrence anyway and even more rarely would the redeal result in the same situation. We’d have to check for it, obviously, but I don’t think that would be prohibitively expensive computationally.

If we have eliminated the chance of 5 of a kind, we have to worry about duplicate cards. I think that I would manage that through “suit substitution”. If, for example, there are 2 king of spades, we could change one to hearts or whatever suit will not cause a problem.

I don’t think that either of these solutions would be particularly difficult to code or expensive to run.

The question that I keep coming back to when people talk about RP rigging the game though, is the question of what gain RP achieves compared to running a fair game.

In real world money, the cheapest option is to operate a fair game using, as other people have pointed out, freely available, very well tested PRNG and dealing algorithms. Any deviation from that results in a real world cost that RP have to recoup. If the best that RP can do is simply recoup the costs of writing and running the “rigged” algorithms, it would be a waste of time.

The risk to RP by running a rigged game is not insignificant! I can certainly imagine that the financial penalties that would apply when they get caught will be quite large. Setting aside the question of jail time, the income from running a rigged game, as opposed to a fair game, must not only cover the eventual penalties that come from getting caught but provide a sufficient gain to replace the income that would be lost from having the site shut down. Once convicted, remember, the opportunities for the owners of RP to make some alternative income is substantially limited.

And this is where I always get stuck! It may be possible to make the numbers work in a real money casino because the income is determined by the rake and, obviously, more high cards results in larger pots and therefore a larger rake. I’m not at all convinced that this is the case but I can see the potential to make an argument along those lines.

I just can’t make any defensible case for RP to cheat though. There is no risk vs. reward scenario that I can conceive of that results in even a weakly positive argument for cheating.

My conclusion is that it is far easier and far more profitable for RP to run a totally fair game.

Regards,
TA

1 Like
168

I kinda doubt there would be any real legal concerns because it’s a play money site. The reputation costs are another matter.

What do we mean by “rigged” anyway?

The more the game is juiced to provide, ummm, “excitement,” the less realistic it becomes. Maybe there is a short-term market for such nonsense, but it seems like long-term suicide to me.

The alternative would be to rig it to benefit a specific player group, but who should you favor?

Rig it to favor those who buy chips and you end up selling less chips, though it might help your retention rate and be a slight net positive overall. Still, this doesn’t seem to be worth the reputation risk that could put you out of business.

And if you rig it against those who buy chips, you lower your retention rate and don’t make as much anyway, while still risking your reputation.

The only sane, sustainable business model is to provide a venue that reflects real-world poker as closely as possible, and that’s exactly what Replay does.

The problem is that a lot of new, casual players have no idea how real life poker works, and exaggerate the number of times their blah blah lost to that yadda yadda. It is what it is and will never change.

1 Like
169

Any time someone makes quads an opponent with a full house is on the cards, because there must be a pair among the community cards, which is a prerequisite for a boat. When you make quads, you always want your opponent to have a boat and no life jacket.

But of course you knew that.

2 Likes
170

I have seen various “reviews” of RP on Web sites, and the chief complaint sseems to be either that the game is rigged in favour of those who buy chips, or that the game is rigged against those who buy chips.

Having never bought a single chip myself, it could be said that I have not had the chance to see one half of the phenomenon for myself, but on the other hand, I have definitely seen enough to prove to my satisfaction that RP is biased in my favour as a non chips buyer.

Furthermore, when you look at all the players who have hundreds of millions of chips, or billions of chips, it seems unlikely that they have bought many of those chips. I believe a billion playchips would cost something like $30,000 which is a lot to pay to play free poker. Most people would use the spare cash to buy a car, or something like that.

I have actually noticed that there may be some bias against chip buyers, because those players who buy 1 million chips and then enter a 1-million chip tournament invariably lose, and usually they are out in the first hour.

The reason for this is that most of the players who play the 1-million chips tournaments have already won dozens or scores of tournaments against the best players on RP and have accumulated multimillionaire bankrolls in the process,so they do not need to buy chips.

For example I was recently up against a player ranked something like 330,000 on RP who was in a 1-million chip tournament. I was in the BB early in a tournament and happened to have AA. This player, who was in early position shoved shoved 5000 chips preflop to steal blinds to the value of 150 chips.

Now, you do not risk 5000 chips to win 150 chips at the start of a tournament. Ah, you might say, if you have a good hand and shove it, then some fool may call and you will double up right away and take the tournament by the scruff of the nexk. However, if you are going to do this with AQ, the hands most likely to call you are AA, KK, QQ, and AK. AQ is not that great a hand and is just a slightly jacked-up improvement on AJ. Shoving preflop with AQ to win 150 chips is not a winning strategy, though admittedly it might occasionally pay off, and you double up. Most likely the loser will be another player who bought chips to get into the tournament who thinks that calling off their entire stack preflop with AJ is a good move.

Anyway, I will tell you for free that the players with multimillionaire or billionaire stacks did not get there by jamming AQ preflop from early position at the beginning of tournaments.

No, the truth is that while RP is rigged against everyone, meaning that every player bar one in a tournament will get knocked out–almost always due to incredible bad luck on the river–it is particularly heavily rigged against bad players who will lose again and again even when they get good cards preflop. In fact it may be the case that bad players get hole cards just as good, on average, as good players,and that even when they do win pots, they don’t win as many chips as they might have done.

There! I have gone and said the unsayable. One of the main causes of chip loss is poor play.

1 Like
171

@spivak are you still active in this thread? I realize that you started it 4 years ago.

I don’t think RP’s RNG is rigged.

But, I do feel that something has changed ever since the recent major change in format about 1.5 years ago. I’ve been asking this question a number of times in the Fairness Debate thread. My first post on the forum was in that thread around July 2020. I was noticing a number of unlikely hands given the few hands I had played.

I’d like to ask your opinion of the frequency one should expect of observing QUADS while playing NL Holdem. My complaint was, I was observing quads being dealt whether to me or someone else at my tables more frequently than I would’ve expected. Recently, I mentioned that the frequency was almost on a daily basis for me which is even more frequently than when I first mentioned it. Observing the number of QUADS being dealt has been my baseline on fairness of the deal.

Just today, I had QUADS but I wasn’t called but, in the very same tournament I was just knocked out by QUADS. And, please realize that I did not play 000’s of hands in that tournament.

Again, I’m not insinuating that anything is rigged because I can’t see how it would be possible to favor one player vs another.

Since, you’re presenting yourself as an expert in probability and very familiar with RNG. If you had observed the high frequency of QUADS being dealt as I have, would this seem odd to you given that it’s occurring so frequently when I’m playing?

I’m not saying observing two QUADS in the same tournament is impossible since it has just occurred to me but this was not the first time for me.

172

in the end we all play I lost quads to a royal

please can you be clear - you feel something has changed and things seems odd - the RNG certificate is there to be seen and think you are best asking replay direct if they have changed anything rather than players as I think they are best qualified to answer your concerns

Best
Tiggs

1 Like
174

Four of a Kind Odds.

“The odds of hitting quads on the flop are a minuscule 0.2 percent or 407 to 1. By the river, it goes up slightly to 2.1 percent but the odds are still very much against you”

So, over a sufficiently large sample, you would expect see quads 2/1000 through to 2/100 hands depending on whether you played your pair through to the river or not.

For the sake of an example, let’s say that a “sufficiently large sample” is 10,000 hands and they are always played through to the river. We expect, then, to see quads 200 times. We should not expect these hands to be evenly distributed. There is no reason at all why the very first 200 hands should not be quads.

Provided we continued dealing to our sample size, the fact that the first 200 hands were all quads and then no more quads were dealt is meaningless. What we need to care about is whether the data fits the expected value which, in this case, it most certainly would.

Hope this helps,
Regards,
TA

1 Like
176

mvac @tiggyxxx

Hey, what matta you?

Btw, I thought by posting my concern in these threads was asking replay in a direct manner. I thought replay reads these posts.

177

OP here, mostly just to say that I suddenly got a huge mass of emails about this thread.

I am no longer active on RPP, either the forums or the games, or anything else;

I stand by my 4 year old post, but I will make no claims at all about what might have happened since then: It is possible that things changed since then.

There are way too many new posts to reply to each in turn, so I think that is probably my sign to peace out.

178

When I posted that it seems there are a lot of big hands for excitement, I didn’t say, or imply it favored players. In and of itself, creating excitement would be part of a good business plan.

1 Like
179

That’s a fair and responsible comment but “creating excitement” very strongly implies, perhaps demands, an understanding that the game is not played using a randomly shuffled, standard 52 card deck.

The push-back is because it doesn’t make sense for Replay to do anything other than use a randomly shuffled, standard 52 card deck.

Even if, in some alternate universe, we could concoct some plausible reason for Replay to “create excitement”, doing so using an algorithm that any clown with zero statistical training can detect would blow up most spectacularly! The site would die within weeks or months of the proof of “cheating” coming to light - and you must remember that if the “cheating” is so easy to detect, there would be a number of players with solid evidence of this “cheating” that would be published on the interwebs high and low. RP could not possibly survive such a thing.

Therefore RP is not “creating excitement” however you want to dress it up or RP is using a highly sophisticated algorithm that can only be detected by detailed analysis, by a statistician, of hundreds of thousands, perhaps millions of games.

I know of 4 players who have done some fairly simple analysis over samples of a few tens of thousands of hands who concluded that the games that they sampled met the expectations of a randomly shuffled, standard 52 card deck. That is a long way from being conclusive but it is, in my mind, a strong indicator that there is no “creating excitement”. What makes this weak evidence somewhat stronger is the complete absence of data from those who claim otherwise.

In the absence of contradictory data, I feel very confident in saying that RP use a randomly shuffled, standard 52 card deck in all games at all times. There is, I believe, no “creating excitement”, no magic “let’s beat those quads with a royal flush” shenanigans.

Regards,
TA

1 Like
180

Great Post I’ve always wondered about how the cards were dealt. Thank You

1 Like
181

This is how the deck is shuffled prior to being dealt

This is gatzby’s explanation. At the time he was the ‘go to’ guy at Replay
gatzby

Sep '15

We do take our random number generator seriously, enough that we’ve had it certified: http://www.gamingassociates.com/certificate/ReplayPoker_Certificate.pdf

While we do make money by selling play chips, it doesn’t make sense to rig the game – nobody wants to play an unfair hand of poker, least of all us. We genuinely enjoy the game and want to make sure other people can too. No fair games, no players, no poker, right? That would be super boring!

In our system, random is truly random. The dealer program has no regard for player history or what cards you’ve been dealt in the past, nor does it account for the cards that have been dealt previously at the table. It does not calculate the types of hands that will be dealt and it certainly does not favor one player over another.

Here’s how the dealer works: For each hand, we create a new deck deck of cards ordered lowest to highest. A card is then randomly selected from that deck and put it into another deck. This process is repeated until every card has been randomly selected from the first deck and moved to the second deck. After the second (entirely random) deck is complete, the dealing process begins from the top of the deck just as you would have in real life poker. As mentioned before, this process has absolutely no knowledge of anybody’s hands or previous decks.

What you might be seeing is the scale at which the dealer is working, which is certainly much larger than any one dealer in a casino. Say you’re dealing with a million hands per day – that’s still 320,400 AK flips per day.
Obviously that math isn’t the best, since the chances of getting AK pockets are going to vary too, but it’s still a heck of a lot of hands, most of which one player couldn’t sit through without getting bed sores!

2 Likes
182

I think you’re the right person to ask this question to. Is it true, that in algorithm, if you fold your hand, the board cards might be different then if you had played? Also, would the speed that we play affect the board cards? By the way, thanks for your helpful posts

183

If you are dealt cards, then fold, it will not affect the board. If you are sitting out and not dealt cards, it will.

For example, if 3 handed, the flop will contain the deck’s 7th, 8th, and 9th cards. If one of the 3 players is sitting out, the flop will be the 5th, 6th, and 7th cards.

Since the whole deck is randomized (shuffled) before each hand, the speed at which one plays should not affect the cards at all.

Technically, I would guess that the duration of each hand does affect the pRNG for the following hand, but not in any predictable way.

184

Very interesting stuff. So we really can’t look at the board cards & know if we would’ve lost or won for certain. like we can with a deck of cards. Thank You & keep up the great posts

185

If you were dealt in and fold, the board won’t change, so sure, you can know if you would have won.

They put the entire deck into a random order, and after that, it plays just like a real deck would play.

1 Like