Organizing a project to determine which sites are legit or rigged

[nineinchal]

[ QUOTE ]
You also have the problem that there is nothing to compare it to. The only thing to compare it to is statistically info. I would imagine if you took a million hands from a B&M it would be off INSAINLY LARGE from what it should do to human error. That being said the randomness is thought of as perfect there.(by the people that dont like online randomizers)

In short this is a decent idea however ill keep all my HH to myself.

[/ QUOTE ]

Again, we would not be analyzing HH to see how you play. It is to view outcomes of hands to determine whether or not the site is rigged in any, way, shape, or form; or whether sites have biases over one type of player or game to another.

[Henry17]

[ QUOTE ]

The people who believe the sites are fair (as I do for the most part) have a great incentive. That is, to ensure that the sites do in fact offer a game equivalent in all aspects of live poker except for showing up at the casino.

[/ QUOTE ]

I don't think that is sufficient to motivate most people. Given a finite and very limited number of hours in a day for someone who is already 99.9% sure the sites are not rigged I can't see how this would be a priority.

[train.]

[ QUOTE ]
You also have the problem that there is nothing to compare it to. The only thing to compare it to is statistically info. I would imagine if you took a million hands from a B&M it would be off INSAINLY LARGE from what it should do to human error. That being said the randomness is thought of as perfect there.(by the people that dont like online randomizers)

In short this is a decent idea however ill keep all my HH to myself.

[/ QUOTE ]

If there is enough interest in this process to get rolling it won't matter whether or not you keep your HH to yourself. There'll be enough HH provided by others that contain your play that if someone want to analyze your play you withholding your personal stash won't matter.

[MicroBob]

If you get the hands and get people to analyze them then I think all you will is a bunch of people arguine back and forth whether the data proves the cards are 'off' or not.

This kind of started happening when some people were looking at onlinepro's stats in that big thread.

Also, FWIW, I suspect you could just take 200k of your own hands from a given site and mostly be fine.

Myturn2raise evidently believes he found some problems with NL all-in hands across 2 or 3 different sites. I already have some theories on the errors he might have committed in coming to his conclusions if indeed he really did do a study and isn't just leveling everyone.

I won't go into great detail on how silly I think the project itself is. Just pointing some other issues with it if you seriously want to go about doing it.

[ryanj247]

seriously, does this really take 100 million hands? it seems to me that most likely this test could be run just on the data i have. i just want to see someone run the same test MT2R did on some set of data and see if they come to the same conclusion. or perhaps identify what about the testing process itself is flawed.

in any case, we need someone to step up and offer to do the statistical analysis on a set data. then, we need specific instructions on what kind of data that person needs (i.e., should we filter out all in confrontations from our HH? etc), and where to send it if we have it.

[Josem]

if it helps, i'm currently collecting a bundle of HHs for some absolute poker stuff.

i have 3million hand histories before that, so i suspect we can use that as a start.


i'd also suggest using holdem manager, since it seems to handle bigger databases easier and more comprehensively.

[MicroBob]

ryanj - I really don't think it should take that many zillion hands to do a study.
The bigger issue imo is in trusting that the people who are interested in taking on such a project can do so competently.

The fact that there isn't even agreement on something as basic as how many hands should be required for such a project kind of goes to my theory that various parties are going to come to vastly different conclusions from the same data anyway and such a study is likely to end up not proving very much.

[nineinchal]

[ QUOTE ]
[ QUOTE ]
You also have the problem that there is nothing to compare it to. The only thing to compare it to is statistically info. I would imagine if you took a million hands from a B&M it would be off INSAINLY LARGE from what it should do to human error. That being said the randomness is thought of as perfect there.(by the people that dont like online randomizers)

In short this is a decent idea however ill keep all my HH to myself.

[/ QUOTE ]

If there is enough interest in this process to get rolling it won't matter whether or not you keep your HH to yourself. There'll be enough HH provided by others that contain your play that if someone want to analyze your play you withholding your personal stash won't matter.

[/ QUOTE ]

There ain't nothing so special in my personal stash that can't be found in 2+2.

[nineinchal]

[ QUOTE ]
seriously, does this really take 100 million hands? it seems to me that most likely this test could be run just on the data i have. i just want to see someone run the same test MT2R did on some set of data and see if they come to the same conclusion. or perhaps identify what about the testing process itself is flawed.

in any case, we need someone to step up and offer to do the statistical analysis on a set data. then, we need specific instructions on what kind of data that person needs (i.e., should we filter out all in confrontations from our HH? etc), and where to send it if we have it.

[/ QUOTE ]

I wouldn't know exactly the extent of testing.

I assume that there would be some sort of process that could be developed that would indicate that there was some sort of rigging that was built into a site that would in fact be conclusive and used as valid info.

[nineinchal]

[ QUOTE ]
ryanj - I really don't think it should take that many zillion hands to do a study.
The bigger issue imo is in trusting that the people who are interested in taking on such a project can do so competently.

The fact that there isn't even agreement on something as basic as how many hands should be required for such a project kind of goes to my theory that various parties are going to come to vastly different conclusions from the same data anyway and such a study is likely to end up not proving very much.

[/ QUOTE ]

Here, here.

I would want to conduct this in a way that any conclusions would prove whether or not a site is rigged. This information would serve our community by guiding us towards quality gaming sites and away from fraudulent operators.