[VA] SRC #012a - Then and Now: Probability

[C.Ret]

(10-09-2022 01:29 AM)Valentin Albillo Wrote:  As 23/288 evaluates to 0.079861... then yes, your result is correct, except that you omitted the 8, a simple typo.

Quote:EDIT:
I upgraded my code to avoid rounding errors as much as possible. Now I find P = 1656 / 20736 for R=5 and S=4 [...]

But 1656/20736 immediately simplifies to 23/288, which is the correct result, as stated in my OP. Why would you give your result as a fraction not reduced to its lowest terms ?

Thanks Valentin.

I was very tired last night, which certainly explains the errors in copying, but also my lack of lucidity. I didn't even see that my fraction was reduced to 23/288.
My general condition explains this but does not excuse me.

Today, I wake up in great shape and I will go over my notes and resume my program. I will post it here and explain how it works. The major problem now is its efficiency because in the current version, on my poor HP-71B, it needs no less than 1'27" for (R,S)=(5, 4) and 2'32" for (R, S)=(5, 7).

And of all the authorized machines I own, this HP-71B is by far the fastest.

I expect it to take over 4 hours to calculate (R,S)=(50,60).

I recently got a MATH module, I hope to find the algorithmic way to take advantage of it. For the moment, the calculation is done using arrays and numerous nested loops. It looks very much like a matrix product.

I'm going to have a good breakfast and finish some work in the garden to prepare it for the arrival of winter while thinking about it.

I will post this evening the fruit of my developments, trying to do it before falling asleep and being unable to reread myself...

Best regards.
C.Ret