What Would It Take???

[New2vp]

Hating a statistical test might be a bit extreme, but I can certainly understand why someone favors a Bayesian approach.  Unfortunately, beyond the allure of incorporating prior beliefs, Bayesian approaches are quite a bit more complicated to implement.I think it also requires a lot more sophistication on the part of the user than chi squared tests. Transforming prior beliefs into a probability density function is generally a requirement (this is called a prior distribution, generally shortened simply to "prior," using the term as a noun instead of an adjective) and that may be more than we are capable of explaining to a lay user of your utility.  This is followed by an integration to determine a posterior distribution or simply a "posterior."  And, often, the idea is to allow continual updating.  Thus proper choice of a functional form is often a consideration to allow for relatively "easy" iterative integration, assuming that your user thinks any integration is easy.Personally, I am much more familiar with Bayesian estimation of parameter values (like means and variances) than Bayesian tests of randomness.  Although my memory is imperfect, I don't recall ever doing a Bayesian randomness test.There have been some studies done checking things like the randomness of lottery drawings, which are likely quite applicable to the video poker problem, but sorry, I have nowhere near enough time right now to work out an example.  Maybe, down the road I can give it a whirl but no one should hold their breath until that opportunity occurs.

[Frank Kneeland]

Hating a statistical test might be a bit extreme, but I can certainly understand why someone favors a Bayesian approach.  Unfortunately, beyond the allure of incorporating prior beliefs, Bayesian approaches are quite a bit more complicated to implement.I think it also requires a lot more sophistication on the part of the user than chi squared tests. Transforming prior beliefs into a probability density function is generally a requirement (this is called a prior distribution, generally shortened simply to "prior," using the term as a noun instead of an adjective) and that may be more than we are capable of explaining to a lay user of your utility.  This is followed by an integration to determine a posterior distribution or simply a "posterior."  And, often, the idea is to allow continual updating.  Thus proper choice of a functional form is often a consideration to allow for relatively "easy" iterative integration, assuming that your user thinks any integration is easy.Personally, I am much more familiar with Bayesian estimation of parameter values (like means and variances) than Bayesian tests of randomness.  Although my memory is imperfect, I don't recall ever doing a Bayesian randomness test.There have been some studies done checking things like the randomness of lottery drawings, which are likely quite applicable to the video poker problem, but sorry, I have nowhere near enough time right now to work out an example.  Maybe, down the road I can give it a whirl but no one should hold their breath until that opportunity occurs.

The book I had used Latin terms , "a priori" and "a posteriori" to describe it and I thought at the time it might be too hard for the standard user. I really had no idea how to get it into Excel. Now I have a second. Motion carried.Chi Squared it is.

[backsider]

OMG! Ive been back home for a short leave, and here I see we have a dream come true for my bud new2vp!

Im ready to begin my personal testing, as long as I dont have to understand Chicago squared or any of the other foreign language.

What do I do? I still havent been at a machine since my unreal deal about six weeks ago. That brings up another question: has Lucky Larry procured a new good luck charm yet?

[Frank Kneeland]

OMG! Ive been back home for a short leave, and here I see we have a dream come true for my bud new2vp!

Im ready to begin my personal testing, as long as I dont have to understand Chicago squared or any of the other foreign language.

What do I do? I still havent been at a machine since my unreal deal about six weeks ago. That brings up another question: has Lucky Larry procured a new good luck charm yet?Welcome back. Have no fear, in its final form all you'll have to do is keep records and input data. The incomprehensible stuff is being handled for you.I'm just finishing the first of three parts, so we're a ways off yet from having it completed.OH...BTW...it's my birth day.~FK

[faygo]

Happy Birthday Frank!       

[backsider]

Here hear! How many candles do we need again?

[Frank Kneeland]

Here hear! How many candles do we need again?That would be 44. Out for rest of day.

[Frank Kneeland]

Update. I've finished the all-dealt-cards part of the test and I'm now adding the drawn cards portion.Next will be the strategy dependent section, which I expect to be the easiest, and I'll end with the Strategy Independent section.Even if no one ever uses this darn thing it's been a marvelous learning experience for me and I'm enjoying making it.~FK

[backsider]

I think, Frank, that the only ones who WON'T use your tool are the ones who fear what they might discover. Why, theyre probably at this very moment constructing make believe excuses as to why they experienced that which they never expected.

Your efforts are much appreciated.

[edog743]

Wow a kinder gentler Backsider. Did not think it was possible