Further to Carter’s point below, we should be very clear there is no *best* pseudo random number generator. By their very nature the algorithms they use are deterministic and there is always a trade-off between speed and quality.

In this proposal we have managed to find a way of producing random numbers a *thousand* time faster and of passing almost all the quality tests that are available. Not only that but the underlying algorithm (thanks to Guy Steele et al and Oleg Grenrus) has really good splitting properties which we have also thoroughly tested.

I urge anyone interested to run the quality tests available here (https://github.com/tweag/random-quality) and to run the performance tests available here (https://github.com/lehins/haskell-benchmarks).

Dominic Steinitz

dominic@steinitz.org

http://idontgetoutmuch.org

Twitter: @idontgetoutmuch

On 2 Jun 2020, at 19:32, Carter Schonwald <carter.schonwald@gmail.com> wrote:

The proposal has a number of nice ideas. And a number of things that I’m reflecting on how to tweak or align with my own thoughts because I think it’s wrong but I’m not 100% on what approach is best yet. Just that the current one proposed or that random currently has is not !