/sci/ - Science and Mathematics

You can always flip a coin 10 000 times :^)

>>3058

Sure.

It's called Monte Carlo method.

You just need to understand that your error will be about N^(-1/2).

Yes, there are so many ways to bypass computation with probability theory. Stirling's method is an old favorite of mine, but there are many, many neat tricks (like the Box-Mueller Transforms) that help make computationally difficult things simple.

>>3058

>>3061

One of the biggest problems with Monte Carlo simulations is that it easily misses rare events unless you're using an ungodly number of samples. And an unfortunately large chunk of the time those rare events have large effects.

But hybrid approaches, Monte Carlo combined with other methods for rare cases, is astoundingly powerful.