It may worth adding that usually in computer science
pseudo-random numbers are generated. Truly random (?) values are only to be
found in nature. And pseudo-randomness may, strangely(?), be closer to what you expect in statistics.
In my computer programming class, we're task to random out numbers and sometimes, tutorials are saying that random numbers mostly comes from the current time you have in your computer's clock. In this case, the system is basing the random number which is not actually random at all. Applying it to the context of gambling, random chances could possibly have a basis of it being random and might not be random at all.
The rationale for gambling is different from the problematic of RNG in computer science, for gambling you still want something that will have poisson like distribution to have a "fair" game where the chances of winning are still known.
Chaotic mechanical processes in the end tend toward poisson distribution if the parameters stay constant, like dice or roulette or coin tossing, even at quantum level on long enough time you can find average with good statistcal match.
Turing machine are not good at creating randomness they are not made for this, on the contrary. They are perfect exemple of closed system with very low entropy.