Doomswitch? Part 6 Last one!

[bigredlemon]

[ QUOTE ]
[ QUOTE ]

Probably not. The actual number of possible deck shuffle combinations is 52! If you are not familiar with how big this number is it exceeds the number of atoms in the known universe. Really. Something like 8x10 to the 67th power.


[/ QUOTE ]

Not quite. This post is largely off topic but with the nit scientist in me I feel I have to address this.

52! is indeed 8x10^67 but is nowhere near the # of atoms in the universe. Avogadro's number is 6.02x10^23 which represents the number of atoms in a mole of something. How much is a mole? A mole of water for example would weigh 18g or have a volume of 18ml, not very much.

Now consider that the amount of water on Earth is about 326 million cubic miles, which translates to, well, a heck of a lot of atoms[1] (three atoms in a molecule of water on top of that). And that's just counting water...on one planet.

/geek mode off

[1] I did a quick calculation and got about 2.38x10^48 atoms just in earth's oceans.

[/ QUOTE ]
The difference between 2x10^48 and 8x10^67 is huge however. One is about 40,000,000,000,000,000,000 times bigger than the other. A quick googling shows that there's 1.33x10^50 atoms on the entire earth. There are 5x10^68 atoms per galaxy. So in conclusion: in the grand scheme of things, nothing you or I do will matter. Even the Hitlers and Ghandis of the world are less than a spec of sand in the beach of the universe.

[RustyBrooks]

Yes of course, there are a tremendous number of possible deck orderings.

But your claim is that certain situations had subtle modifications, like from 4:1 to 3:1. Even if the modifications were much slighter, from 4:1 to 3.9:1, it would not take very many hands to figure this out.

I mean, people who say a site is rigged are not saying "some decks appear more often than others" right, they are saying "big hands don't hold up as often as they mathematically should". This is a very easy assertion to test.

But even so, you do not need to examine a significant number of deck shuffles to see if the distribution is random.

[SGspecial]

[ QUOTE ]
I mean, people who say a site is rigged are not saying "some decks appear more often than others" right, they are saying "big hands don't hold up as often as they mathematically should". This is a very easy assertion to test.

[/ QUOTE ]
Actually I don't think you've got the common assertion right yet. The assertion most people use to claim site X (or online poker in general) is rigged is that their big hands don't hold up as often as they should, or the corrolary that their draws don't come in often enough. The typical claim is by those who feel they are good players and play correctly, and that the sites are rigged in favor of the fish so they don't go bust as quickly and will pay more rake. Thus you can't disprove conspiracy theories like this by showing that AA holds up over KK 82% of the time since it could be rigged that they hold up 90% of the time for the "fish" and only 50% for the few good players.

[RustyBrooks]

[ QUOTE ]
[ QUOTE ]
I mean, people who say a site is rigged are not saying "some decks appear more often than others" right, they are saying "big hands don't hold up as often as they mathematically should". This is a very easy assertion to test.

[/ QUOTE ]
Actually I don't think you've got the common assertion right yet. The assertion most people use to claim site X (or online poker in general) is rigged is that their big hands don't hold up as often as they should, or the corrolary that their draws don't come in often enough. The typical claim is by those who feel they are good players and play correctly, and that the sites are rigged in favor of the fish so they don't go bust as quickly and will pay more rake. Thus you can't disprove conspiracy theories like this by showing that AA holds up over KK 82% of the time since it could be rigged that they hold up 90% of the time for the "fish" and only 50% for the few good players.

[/ QUOTE ]

This assertion is provable too, as long as you can establish the rules for what you might think the site thinks a fish is. Make 4 new accounts, play like a fish, then repeat my experiment. [Edit: or recruit 4 fish]

Look, this is an assertion about a statistical anomaly, that has been empirically "observed" albeit anecdotally and not particularly rigorously. It is eminently provable, or disprovable.

If I was a richer man I'd do the James Randi thing and give a prize for proving it.

[SGspecial]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
I mean, people who say a site is rigged are not saying "some decks appear more often than others" right, they are saying "big hands don't hold up as often as they mathematically should". This is a very easy assertion to test.

[/ QUOTE ]
Actually I don't think you've got the common assertion right yet. The assertion most people use to claim site X (or online poker in general) is rigged is that their big hands don't hold up as often as they should, or the corrolary that their draws don't come in often enough. The typical claim is by those who feel they are good players and play correctly, and that the sites are rigged in favor of the fish so they don't go bust as quickly and will pay more rake. Thus you can't disprove conspiracy theories like this by showing that AA holds up over KK 82% of the time since it could be rigged that they hold up 90% of the time for the "fish" and only 50% for the few good players.

[/ QUOTE ]

This assertion is provable too, as long as you can establish the rules for what you might think the site thinks a fish is. Make 4 new accounts, play like a fish, then repeat my experiment. [Edit: or recruit 4 fish]

[/ QUOTE ]
That's the key to the problem. The proof would have to be based on subjective terms, like what constitutes a fish? How could you even recruit 4 fish, if they have to sign up themselves? I guarantee you most of the fish don't think they're playing badly, they think they're good players and the site is rigged against them too. I honestly don't believe there is anything more nefarious going on here than a systematically shallow understanding of probability and statistics, and the fact that 80% of poker players are fish who think the other 80% are fish.

[RustyBrooks]

[ QUOTE ]

That's the key to the problem. The proof would have to be based on subjective terms, like what constitutes a fish? How could you even recruit 4 fish, if they have to sign up themselves? I guarantee you most of the fish don't think they're playing badly, they think they're good players and the site is rigged against them too. I honestly don't believe there is anything more nefarious going on here than a systematically shallow understanding of probability and statistics, and the fact that 80% of poker players are fish who think the other 80% are fish.

[/ QUOTE ]

I agree with you (sites are not rigged).

It's up to the person who thinks the site is rigged to tell me what constitutes being rigged. Just like the Randi foundation, we then come up with a concrete set of rules for what would determine success or failure of the experiment, and try it.

As for recruiting fish, I certainly would not say "Hey, I need 4 fish, someone come sign up", I'd try to contact players that met the criteria, and ask them to participate.

Anyway, the site would have to recognize fish by some kind of pattern, if it is indeed biased towards fish, so if you think you know what that pattern is, you can recreate it yourself.

Even without recruiting players, running tests, it would be pretty trivial to get millions of hands from a bunch of players at 2+2 and crunch through them.

[chucky]

Wouldn't you need 1 million showdown hands? That would take a while longer to compile.

[bigredlemon]

Here's a pretty easy way to determine whether the site is rigged:

Dig up your entire hand history for the site and look for situations where you have a 4 flush on 4th, and you saw the river. Determine the percent of times it hits.

You would only need something like 30 hands to say with 95% confidence that your expected hit rate equals your actual hit rate. With a few thousand hands of the above scenario, you'd probably be able to say if your hit rate was within 2% of expected with 99% confidence.

Getting the hit rate of major losing players will probably be a lot tougher, but this would give you fairly good results.

On a side note, I kept track of how often I won with rolled trips. After two months I felt my lose rate was around 70% with them. My actual lose rate was only 40%... which while is still pretty bad, isn't nearly as bad as I had thought. So if you feel your flush hit rate is only 30%, it might actually be 45%.

[RustyBrooks]

[ QUOTE ]
Wouldn't you need 1 million showdown hands? That would take a while longer to compile.

[/ QUOTE ]

Only if the rigging is very, very subtle. So subtle that you'd never notice it yourself. As the poster above me points out, you don't need the ridiculously large number of samples people think you do, in order to detect malfeasance.

But whatever. People who want to believe the site is rigged are never going to let anyone pin down a set of criteria that would prove that it was or was not.

[SuitedBaby]

"It is rigged" is one of the many money safety issues all of the sites face every day. Yet any site could publish all of it's deals any time they want. For confidentiality they could just publish what was exposed during the hand. Anyone could check their hand history against hand numbers for veracity. The showdown hands alone would have an enormous amount of information easily adequate to determine if things were rigged. But the sites don't make this available. Hmmm.

Patty

Sorry Andy if this has gotten off track. My last post in this direction. A bit of geeky humor to follow to make up and lighten things a bit.

Years ago I used to occasionally quiz some of my smarter friends who would sort of scoff at my card playing hobby. Many were science types. Sometimes I would quiz them as to how many different shuffles were possible with a 52 card deck. Many knew it was 52! Then I asked, "is 52! greater than the number of grains of sand on the beaches of the USA?" Maybe 50% said yes and 50% said no. Everyone that said "no" agreed however that it was not greater than the number of grains of sand on all the beaches on Earth. Of course the number of grains of sand on all the beaches is incredibly, incredibly, undescribably small compared to 52!.

Then one day I asked a PhD physicist friend of mine the same questions. He said "oh my yes 52! exceeds greatly the number of grains of sands on Earth." In fact he said the number was probably something like 1x10 to the 55th power. When I told him the number was actually something like 8x10 to the 67th power he was aghast and obviously embarrased. To quote him he said "Oh my God, I was way off". That is even though he was actually "infinitely" (as used in common parlance) closer than anyone else who had confidently estimated the number and probably ended up felling they had been pretty close.

The science of large numbers is a fascinating thing and can play with your mind a bit.