@user10478 It depends on Method you choose in SeedRandom. Closest to bad traditional C PRNGs would be the "Congruential"method. The default setting should pass typical randomness tests, though.

Yet another mystery delivery...

@user10478 The default ExtendedCA generator is quite good, as @kirma mentions. Some good info in this post: mathematica.stackexchange.com/questions/3208/…

Awesome, thank you both very much!

Some seeds can produce suspicious autocorrelation tests, but I am not sure how important this is and I don't really know a lot about prng testing

    SeedRandom[1234, Method -> "ExtendedCA"];
rands = RandomReal[{0, 1}, 1000];
Tally[AutocorrelationTest[rands, #, "ShortTestConclusion"] & /@
  Range[999]]
(*{{"Do not reject",473},{"Reject",526}}*)

@ydd "ExtendedCA" is definitely a bit of a black box in the academic sense. Nonetheless, using 1234 as the seed has only 5 nonzero bits of 5120...

I would assume that these PRNGs are seeded from a strong operating system entropy source by default.

I am also probably just going to far with the lag all the way out to `Length[rands]-1` also. I don't really use autocorrelation tests ever so I could be doing something meaningless by going out that far in the lags. Another seed that is poor is 12345. The automatic seedings appear to do much better with the autocorrelation tests

It is probably best practice though to not assign a RandomSeed[] value though I guess? Unless you want your prng to be reproducible which I guess you would have to specify a seed

SeedRandom you mean? It's slightly concerning that I think some functions have small default values for random seeding in order to have reproducible results. I don't believe that those algorithms would have guarantees in producing unbiased samples in general, though, because it's often more useful to find instances than to guarantee statistical properties of those instances.

Yes sorry SeedRandom

no problem :)

For instance, FindInstance has a RandomSeeding default of 1234, specifically.

At the same time FindInstance is not going to return unbiased instances (although it's not explicitly stated) when a large set of instances is requested, *Sample functions exist for that purpose where applicable.

Same default seeding for NetInitialize