So you think this his post/claim is correct ?
"You don't analyse the gazillion possible ways cards can be dealt. You analyse the sample that was actually dealt. We already know exactly what the entire population looks like, that takes zero analysis to know with 100% certainty, and the size of it is irrelevant. "

He is right on these areas. You are wrong. You are 100% wrong. Unquestionably you are wrong. You may believe you know what you are talking about, but you are wrong. Really, really wrong.

Again, if you think otherwise then post on the forum I suggest which is dedicated to exactly the topic you think you know what you are talking about. They will tell you that you are wrong, and my hope is it will be amusing to watch you debate with experts on a topic where you are totally wrong, yet believe you know what you are talking about. No secret what my agenda is on this, and I have no expectations you will post in that forum, because deep down you know you are 100% wrong.

In case you missed the link, here it is again to the stats forum

https://forumserver.twoplustwo.com/25/probability/


All the best.

Further proof that you're in way over your head on this one.

Also another great example of:

https://en.wikipedia.org/wiki/Dunnin...3Kruger_effect

Of course it is.

I'm a stat guy & I agree that 10^30 is bigger than 10^4 and harder to analyze.

There's prolly more variables when the numbers are so big.

I knew this handicapper who made his first bet when he was fifteen years old,
and he always made money. But he didn't bet like you or me.

You know, havin' some fun with it, crap like that. He bet like a f'ing brain surgeon.

He'd get the wind velocity so he could judge the field goals. He even figured out the different bounce you got off the different kinds of wood they used on college basketball courts, you know?

Season after season, the prick was the only guaranteed winner I ever knew. But he was so serious about it all that I don't think he ever enjoyed himself.

Definitely not rigged.

Here's an interesting read for you: https://en.wikipedia.org/wiki/Sample_size_determination

Go to the various linked pages therein, too.

This is for a sample size determination of two possible events.
A binomial distribution with yes or no.

Lets say we break it down to 17 cards dealt that is a magnitude of 10^16
Then we cut out identical events due to permutations i.e. 6 permutation for a flop and so on.
Also we don't go with a too fine comb at it and put events in groups an categories.
Lets say everything break it down to a magnitude of 10^12 or less.

Simple approximation is this:
https://en.wikipedia.org/wiki/Bonferroni_correction
Our m is going be 10^12, that means if we aim for a 0,05 error we have to calculate a 0,05 * 10^-12 error.
Which leads us to the simple calculation n = 400*10^12 = 4*10^14 sample size for a chance of 1 out of 20.
Pokerstars has a promo that announced that they are reaching 10^11 Hands dealt.

And that is vague assumption for a minimum amount.
There are still several issues I didn't put into consideration.

I do this for a living, have dealt with much simpler systems then this before and they were already a huge pain in the ass.
A system of this size is a total nightmare.

Just common sense should point you in this direction that you can not analyse a thermodynamic system with multiple parameters like a several flips of a coin.

You are still wrong. You still have not posted your wrong in the stats forum to get comments from the non rig thread experts, which shows you know you are wrong. Continue to be wrong here for as long as you like, many deluded expert riggies before you chose that path.

Here is a link again to the stats forum. I know you will not sue it, but it is fun to remind you and see you continue to avoid the topic.

https://forumserver.twoplustwo.com/25/probability/

All the best.

What do you do for a living?

I'm not sure what your calculations are supposed to show, so let's bring it all back down to my simple level.

If we want to test how many pairs of aces we received (in the hole) against how many pairs of aces we "should" have received, (or any other test we want to try), by inspecting the 100,000 hands we have seen dealt, are you saying that whatever the answer is it's meaningless because there a lot of ways that the rest of the cards in the deck could have fallen if they had been dealt?

*

Edit: It's a pity you weren't around when was posting. You and he would have had some interesting chats.

It would have been ostensible.

All this handwaving still shows that you don't understand statistical analysis.

Ah, yes, good old thermodynamic systems like poker games.

So mods, is this the ostensible back again in a new form after his last one was banned? If so I assume he has not obtained the same status as the Nash guy who gets banned whenever he creeps back in.

There is virtually no chance that a repeat of the entertainment had by the ostensible riggie can be duplicated (even though he did it a coupel times over a span of years in the stats forum), especially with the same zealot, and I doubt he would answer truthfully if asked.

Note, I am still jealous that Mike Haven subjected the stats forum with the ostensible riggie, because I have tried many times to get deluded riggies to post there for amusement, and he got a guy to create a hall of fame thread (that is even stickied there) on pretty much his first try, so I will be embarrassed if I have been trying to hurl the repeat of that same guy to that same forum!

There's no obvious behind the scenes link, and there's an awful lot of overlap between ban and join dates.

There are some posting similarities.

I am a physicist.
I have been trained in programming statistical simulations and analyzing data, also I do it on a regular base.

Just trying to show how complicated that issue really is.

On that example with those Aces.
Not all Aces are the same and that fact could also get abused.
- Some Aces are easier to play some are tougher.
- Some Aces can generate better value then others.
- Some win some lose
...

The answer is it's not meaningless, but to get really good quality we would have to go further into detail, which we might not going be capable of.

How are you going to get all the information ?
How are you going to get such a huge sample size ?
How are you going to deal with cards we don't see ?
Not every hand and board is going to be revealed to us.

You also have to assume that someone who is rigging data might not just put in arbitrarily values and could also run some kind of evaluation to check his output.

Just look back to the superuser scandal on UB back then.
They might never found out about it if the guy would not been so reckless and just constantly checked and adjusted his stats.
He just had to went into his own database look up some stats of several winning players and try to imitate them.
Also it was super lucky that they send them his hand history.

In other words you could totally get ****ed over and there is nothing you can do about it

If this is true, then you are still comparing things that are not similar. The kind of simulations you are almost certainly referring to, are samples of which the population definition is not known and you are trying to infer information about it from a sample.

In poker, or any card came, we know exactly what the population of possible deals and hands looks like, and we know how frequently everything occurs in the long run. Tests for a biased deal are generally tests to see if a sample is likely to be randomly dealt from that population or not. You are looking to make a determination that sample X has Y% chance to be part of the normal random deal population. This isn't complex.

These are simple tests that a high school kid can do on paper without a computer, assuming he has some sample statistics to work with. Those come straight out of any poker tracking program or from hand databases.

And no, we don't need to know the unseen cards to do tests of this kind.

What do you think poker players do when they perform a statistical analysis on a sample of hands? All it is is a series of binomial distributions.

- Given that I was dealt a suited hand and saw the flop, how often did I flop a flush draw?
- Given that I was dealt QQ and saw the flop, how often did I flop at least an overpair?
- Given that I was dealt into the hand, how often was my hand precisely AA?

...and so on.

The thing is that the filtering functions in programs like PT4 and HM2 make this analysis trivially easy.

That guy posted his crazy stuff years apart in the stats forum, and openly acknowledged he was posting similar drivel on a bunch of different message boards. Was fun to watch people try to talk sense to a genuinely strange human, but if this person is just a repeat then that falls fully in the meh, whatever, been there done that category.

That time it was using the word ostensible a lot. This time he claims to be a physicist talking about thermodynamic systems. Instead of ?/3 it is 10^16. Yawn/0

Not much can be done with him if it is him, and it is always disappointing to see someone come back and produce a much lower quality thread. Even the guy who had the epiphany from watching the movie "Casino" will ideally run out of quotes soon in reply to him.

Given how he does not really speak directly to people (much like the other account), I will have to assume it is him again doing his message board rounds with his new/tweaked belief. Who knows, maybe he will find a new twist to make it fun to read again, people already seem to be trying to have a genuine conversation with him to point out his errors. That should go well...

Great. It doesn't change the fact that your posts have been full of flaws and bad assumptions.

Did you read his post? He was talking about analyzing whether a player is getting more or less pocket Aces than they should be. For such an analysis, all Aces are the same.

You've explained yourself quite poorly here once again, but I think I understand what you're getting at. It sounds like you want to change this to a rig where they strategically give more or less Aces based on the situation. Yes, that could be complicated to detect. It would also be quite complicated to program in a way that would generate more profit.

No idea. What are we trying to prove?

Well, I don't think a rig would make any sense unless there was some evaluation behind it. An arbitrary rig would be pointless.

I think that's a little oversimplified, but I'd definitely agree that he/they could have been much better at avoiding detection.

Yes, but it wasn't needed. It was obvious there were super using accounts without the hand histories.

No question there could be rigs that aren't detectable, but I'm not convinced that a rig that is profitable enough to be worthwhile could be undetectable. As for super users, someone who is very careful could be very difficult to detect - but that also requires a site to be complicit or very, very negligent. But yes, I'd agree that's possible.

I'd be much more on the lookout for colluders, bots, and hackers. But that's for other threads.

They are correlated you don't have one single event you are looking for like do I have a flush or not.
You look at a chain of events and will get a multivariate distribution.
One of those bad boys in n-dimensions.

Yes, you are looking at one single event. One at a time, anyway. But you can look at as many different single events as you have time for. Players have done exhaustive analysis on the distribution of a multitude of different events. Then, they post their results and subject them to scrutiny from their peers. No one who has done this kind of analysis has found any evidence of rigging.

Even if this has any relevance in analyzing a database of poker hands (and count me as skeptical), it doesn't make Mike Haven's link about sample size determination irrelevant. You take a sample of a population (in this case, all the hands you've ever played on poker site x), analyze it for a multitude of different events, and use the data to make inferences about the population as a whole (in this case, all of the hands that have ever been played on poker site x). I do wish you'd take Monteroy's suggestion and go over to the Probability forum and explain to those guys why their methods are inadequate.

I won't run out of quotes but at this point i'm entertaining myself more than others and the guy didn't take the bait.

junglemit more fun b/c he actually attacks and insults you

If you can't see why a event with two possible outcomes is not the same like a event with a trillion possible outcomes, then you are either to stupid or to ignorant.
Either way a discussion is pointless.

Seems like that Casino scene was far more matching the situation then I realized on first sight.

But the conversation is about poker, where a player's hand has just 169 relevant outcomes. The rest is just irrelevant permutations of those.

Even when considering the full board and multiple players hands, the outcomes relevant to poker are still magnitudes below "trillions".