Information on calculator Random Number Generators from PPC Journal articles

[Dieter]

(11-07-2016 11:06 PM)Gene Wrote:  And here's a link to your R-D RNG in PPC Journal V7N6:

V7N6 Valentin's R-D RNG write-up

The text says that the seed must not be pi or an integer multiple thereof. In this case the R-D function would return a multiple of 180° and thus a fractional part of 0 – and all subsequent "random" numbers would be zero as well. But this is not the only limitation. The seed also must not be equal to any value in radians that is equivalent to 1°, 2°, 3°, ... 180°, 181°, 182°, ... . A part of these will not round to an integer angle after R-D, so these do not return 0 as the next random number, but a very small value like 1E–8 may be returned which leads to a predictable sequence of near-zero random numbers following this one.

As already mentioned in previous posts I agree with Valentin in that we should see the random number generation for this application (games) a bit more relaxed. That's why I suggested the frac(x+pi)² method. Even though it may have its weak points in terms of mathematical properties, it worked fine for me over the last decades. And it can be used even with integer. negative and zero seeds.

Dieter