> Always be wary of provided random number generators

Pseudo-random number generators

There are several concerns when it comes to PRNGs:

1. They should be very fast. An application should not spend its time generating random numbers.
2. The output should have robust statistical qualities. Bits should appear to be independent, and closly follow the desired distribution.
3. For cryptography, it is important that an observer can’t learn anything meaningful about the PRNGs internal state from its output.

For simulation and analytical workflows, only the first two properties are important.

References

[1] Press, William H, Numerical recipes 3rd edition: The art of scientific computing, Cambridge university press, 2007.

• Finding the Best 64-bit Simulation PRNG