This is actually what is typically done. On Poker Stars they do this once before the hand and the shuffled deck is fixed. On Full Tilt the stub is shuffled before every street. But they both set up a deck array and randomize it (and FT also randomizes a large set of stubs that are called on as needed). Full Tilt says their method is to prevent superusers since the next cards are never known until time to deal them. Poker Stars says their method is just like using a real deck. Both are equally random.

The best part in thoses threads is usually the one where OP explains he would rather play vs good players than fishes who call raises with ATC.

That's another way of doing it that would probably be used by a Casino. It's actually less efficient if you only need a small number of cards per hand (e.g. for certain simulations). This is because it involves sorting the array - not a problem for dealing cards for real games but a potential bottleneck in simulations.

To actually answer the original question (which was how you get cards from an RNG) using the sort method:

Set up an array of pairs of numbers, one element of the pair being the position in the array and the other a random number. Sort the array on the basis of the random number and then use the index numbers as your cards (assigning a number of 0 to 51 to each possible card).

There is actaully a rather subtle problem with that approach (that should certainly be corrected by a casino) - anyone see it?

I'm having trouble understanding the bolded part. In your array of paired numbers, what are the random numbers? Are they the numbers 0 to 51 randomly sorted somehow? Or are they random fractional numbers between 0 and 1, maybe? I'm not quite able to follow this without a little more explanation, so I'm definitely not able to see a subtle problem with it at the moment.

It's hard to know at what technical level to pitch this.

Typically, RNG's will provide 16, 32 or 64 bit integers (basically just a random string of bits of the specified width).

Obviously, for a pack of cards, you would use 52 elements but let's just use four for this example. You set up an array of pairs of numbers, the first of which is the index into the array and the second is a random number. e.g.

0, 62543
1, 27346
2, 153
3, 2883

Now you sort the array on the basis of random numbers so it looks like this:

2, 153
3, 2883
1, 27346
0, 62543


Now you have the four numbers 0-3 in a randomised order.

Scale that up for a 52 element array and you can deal a pack of cards.

Okay, now I see what you meant. The immediate problem is that you might, by chance, generate the same random number more than once, in which case your sorting algorithm wouldn't work as expected and the "randomized" deck wouldn't quite be perfectly randomized.

The approach I had in mind was to make a single pass through the deck and swap each card with another card, selected randomly. I saw this implemented in a short C program a long time ago and it seemed extremely elegant to me.

My personal preference would be for the swapping method since it seems simpler and less compute-intensive, but I think both algorithms would work fine once all the theoretical gotchas are ironed out.

Yes, that would work very nicely. And it would save the sort.

So that's three ways of getting cards from an RNG. Should be enough to be going on with.

If they worked on my computer then I would.

Woops. Another problem is that there is an inherent bias on one end of the deck or the other. It looks like you were generating 16-bit random numbers to sort. There are 65,536 possible numbers using 16 bits, and since we want each of our 52 cards to have equal probability of being selected, we need to get 52 equally sized ranges of numbers from our RNG. But 65536 is not evenly divisible by 52. The largest 16-bit number evenly divisible by 16 is 65,520. That means 16 cards will have a slightly greater chance of being selected than the others in the deck.

The swapping algorithm also has this problem, and there might be more gotchas in both algorithms I'm not seeing.

The simple solution for this problem is just to reject any returned integer greater than 65,520.

Online blackjack is not rigged either.

.

.

I don't think his assessment of his skills or that of the other "decent" players is terribly accurate.

But it sounds like more of the same - patterns, not random, Bernie Madoff, etc. Oh, and the part where he doesn't have any evidence too.

Where's that from, Sooper? Stephen Meares' new blog?

I just googled "Is Pokerstars Rigged" to have a read of what people think outside of this forum.

Seems like there s thousands of so called "rigtards" out there. Maybe Im not so crazy afterall.

Thats a rigged hand. As if the richwalter player doesnt have access to what cards are coming.

Thousands of people keep thinking they can get a Reparations Credit on their income taxes but that doesn't make them right.

What do you make of this post Mark?

Or are you just going to look for the tiny part of the post you can make fun of?

Somehow I doubt he does...



Hedid have a good score where he won a Stars MTT once and managed to show a profit in in games tracked by OPR. However his last MTT there was in 2005, and he only cashed once in his last 24 MTT's. In fact, he only cashed 10/79 tries there.

In 2007 he must've gotten a poker bug up his ass and tried FTP, who apparently aren't as generous with goldern accounts thus he never had much success there at all.

So, if he had access to turn & river cards, why did he do so badly in smaller tournaments on Stars and all of 'em on FTP? Also, why did he stop playing at all?

I think it's worthy of being made fun of, if that's what you're asking. But luckily, there's a lot to choose from!
That's a pretty small sample size, I'd say. Also, if he's cashing that infrequently, I doubt he's as good as he thinks. Probably doesn't understand ICM/Pushbotting - which is really a skill "tight" players often lack.
I'd love to see these hand histories. These stats are dubious at best.
these are the notable beats out of 136 games? wow, he really has a mind like a steel trap.
if your boy will cough up the HH's and this is true that would be fishy. somehow I bet he never does though.
again, I don't really put a lot of faith in what this fella says, but his sample-size is a joke, so whatever. And like the word "fact," I would reckon he might be having some trouble with exactly what counts as a "coin toss."

Got a link where this came from? Maybe we'll make a field trip...

I'm shocked, shocked to find that gambling is going on in here! The favorite doesn't always win???



If you believe this:



Then you are acting like this:




So have some more of this and you too can be a true believer:

I think he might like to see some PT stats

Rigtards on the internet? Surely not


I await your logical analysis of this sooperfish. Do you now see how producing corroborating evidence is much more powerful than just smugly stating spurious opinions as fact

http://digg.com/software/Is_PokerSta...he_Real_Answer

link

Sharkscope is rigged IMO to hide house players/house bots true results.

There is a name for the syndrome where you include more and more parties in the conspiracy to keep it plausible.

Let me google that for you.

FYP

That's funny, there are also thousands of idiots roaming the planet as well. Coincidence?

Just because a lot of people agree on a conclusion does not mean it is correct. How about all the 'flat-earth folks', or people that think the moon landings and 9/11 were conspiracies.

There are millions of people who have opinions on subjects that they haven't the slightest clue what they are talking about.

Notice how little statistical analysis was done in the long post about people's bad beats.

Never attribute to conspiracy when stupidity of the observer will suffice.