Quote:
Originally Posted by Weevil99
This is a thought experiment. In the originally quoted post, I asked qpw and Markusgc how they would respond if...
Suppose they do have the data, although I'm not sure how it could be verified. If it were UB hand histories, for example, I guess they could politely ask UB to supply you with their HH and see if that worked.
But let's assume they've provided this data, that you've somehow verified its validity, and that you've verified their calculation that what it shows had a probability of .0001315 of occurring.
How do you respond?
Assuming that this happened, then the next step is deciding whether a 1 in 8000 event proves something. Realistically given the number of hands dealt I think it is safe to say that it does not, as many 8000-1 situations will occur every day, especially if you allow people to define their own scenarios after they experience and observe the data.
The key will be to create a theory and test it without using the cherry picked historical data.
The odds of someone guessing a correct number between 1 and 10,000 three times in a row is nearly impossible (1 trillion to 1), however if after seeing the numbers of
8,734
974
5,317
one can say how amazing it is that those 3 numbers were chosen as the odds were 1 in a trillion that it would happen. Even though a 1 in a trillion event just took place, it actually proves nothing. Do the test again and if you get those same numbers then you have something fairly significant as evidence that something is wrong.
So if someone takes their 100,000 hands and finds some things that are very lucky or unlucky among all the ways of analyzing the hands, I would suggest that is nothing special. Maybe flush draws hit a lot, maybe not enough. Maybe sets hit a lot, maybe not enough. You can create an unlimited number of ways to analyze data, but if you selectively choose how you analyze data in hindsight you have not really proven anything.
Unless, you then apply a similar test on future data and show that the same thing happens again and again when it should not. At that point you are onto something.
Quote:
Originally Posted by Weevil99
The set over set example I'm using is something I pulled out of thin air, not something out of my or anyone else's HH. I made that clear in the original post months ago. But this is a thought experiment. Let's pretend it's real and that the probability is as quoted.
How do you respond? Also, what would your private reaction be? I didn't ask that one before, but it's an interesting question.
Test it again with unbiased data that has not already been studied to find where the luck quirks took place. Define your conditions and then gather the data, not the other way around. If your set over set theory is valid then it should appear again and again, so do more testing to show this is the case.
I can tell you the winning numbers of the lottery all the time the day after the draw. That does not prove it is rigged. Most of your "what if" scenarios sort of fall in this approach.
That make sense? If you see a logic error here feel free to point it out (as that was part of your plan I believe).