I'm pretty sure I get what you're saying but I want to push it one step further to make sure I understand.

Let's say we take 5,000,000 real deal hands. The deals were all preshuffled but at strategic times the site switches the order of the particular deals. However, by the end of 5,000,000 dealt hands, all 5,000,000 would have been used (that is: they didn't deal 6,000,000 and discard 1,000,000). I would think on a study like the one you're doing on your site, that you wouldn't see anything amiss. Am I right?

Now, some individuals might have distributions that are slightly off, but would it be enough not to be chucked to variance? Depends how often they would do it I guess.

I'm not saying it would be undetectable, but I would think it would be much harder to detect.

Please correct me, since I'm sure I'm missing something.

Thanks for putting in good english what I was trying to say.

What you are missing, is that to be undetectable all 5 million hands must be played by the same set of players, and the only distribution examined must be the entire set of hands. The distribution is only normal for the entire set as a whole, not for any subset of it if those subsets are chosen non-randomly.

Do you fail often at irony? With all due respect, which in your case is admittedly not very much, I'm not the one who has posted "I am a rigtard" over 800 times in this thread alone. And that's not to mention that bone thread you started a couple of weeks back, about online bots, and whatever garbage makes up the remainder of your posts.

Are you even aware that there are actually threads about poker on this forum?

I was responding another way, but I think I get you: unless your study is going to have every hand ever dealt, given that you're likely only going to have a subset it will be detectable. So your billion hand study would still show anomalies as long as there were a lot more hands actually dealt by the site. If you managed to get your hands on every deal, it could remain undetectable by that method, but still would show up in individual hand histories.

This is why I do like this thread. You can actually learn something! (well, the riggies don't, but I do!)

No, you missed the other condition. It would always be detectable if you examine player subsets or time subsets (which is defacto always the case).

I think that's what I said (though I didn't phrase it well). When I said "that method" I meant your billion hand study: that is if you had every hand ever dealt, in that aggregate study nothing would show up. but if you looked at individual player's hand histories, anomalies would show up.

Am I getting it?

Nah, I'm saying any large, reputable online poker site rigging the deal is about as plausible as magnet grids on a golf course.

Do you plan to make player and time subsets analisys in your study spade (or have already made it)? because it seems to me that this is the most crucial point of any analisys as anyone wouldnt be so dumb to manipulate the total frequency of dealt cards.

Every study is a player and time subset because you do not have access to every hand ever dealt.

This is why they can't store every combination of 52 cards:

Every study works with a sample. You can analyse the entire sample or you can analyse subsets of this sample. I was asking about the billion hands sample spade is analysing, is he going to make subsets analisys of that sample?

But there's no need to, unless you think that those billion hands are being manipulated and no others. If you think the site is rigging the deal, then the set they're rigging is going to be much bigger than those billion hands. The billion hands will be a sample of the full set of x billion hands dealt by the site.

I think you lost some posts, go a few posts back and you will get it.

It only takes 1 well manipulated hand to bust a player out of a tournament or crack his bankroll. They dont need to rig every hand. Just 1 to knock you out and another one or 2 later if you didnt take the bait. This keeps the statistics within variance.

It works on the same principle of discipline that is needed for success in the game. You can play 99.99% of your hands perfectly, but fk up once and none of the other hands matter. Replace fkn up with getting rigged against and its the same thing.

And what's the reasoning here? And what criteria is used to determine whose turn it is? Be as creative as possible here.

I didn't lose any posts, and I believe I get it just fine. But I welcome anyone to correct me if I'm mistaken, pointing out the flaw in my logic.

lol google pro golf rigged with magnets

i quit FTP long ago

im starting to think there are just massive amounts of undetectable superusers planted by the house.

pooflinger

Your new site isn't going to accept US players. What are you going to do now?

Have you ever met a paranoid schizophrenic that knew he was ill?

I had my 3rd full house beaten in about 500 hands this week the other night and I decided that was it. I just cant keep doing it to myself. I watched anything from 10 high to flopped king pair win for an hour as I got nothing only to finally get a playable hand and run into the bigger full house and broeak my roll for the night. I'll stick to trading rigged foreign exchange markets thank you very much.

Does that mean we will no longer have the pleasure of your entertaining and carefully thought out musing on this thread?

As if.

This is an interesting side problem, I don't think there is a clear answer to it as it depends on some definitions.

52! is less than 2^225 so you might say that you can store a deck of cards with 225 bits, which is theoretically true, but the problem then is to 'decrypt' what the deck is from the string of 225 bits, you would have to look up the string on a table containing every hand - there is no 'logical' way to do it so that an algorithm can determine the deck from the string.

To get 249 bits you say, presumably, we start with a deck in some known order, Ac,....Kc,....,As,....Ks or whatever, then move the first card to some position from 1-52, it takes 6 bits to uniquely define this position. It takes 6 bits to move the next card too, until we get down to 32 cards left when it only takes 5, etc. Total 6*20 + 5*16 + 4*8 + 3*4 + 2*2 + 1 = 249.

However, surely 248 bits is enough to uniquely define a deck with that method, not 249? The last card can only go in one place anyway so we don't need to say where it is.

Also, this seems like a fairly simple algorithm - it is good because it is totally clear what is going on but there might be one that uses less bits and can still be decrypted in reasonable time. Anyone have any ideas, I'm curious.