Problem

We have some situation, and want to calculate the probability of some outcome / some value ,but we have no analytical model for it.

We want to have some answer to the given question, and don't want to spend 1 week/month/year on developing some analytical model.

Idea

Use "Monte Carlo" (famous casino) methods.

All that you need:

• code that describes the logic of the problem (game logic, math problem, assets prediction)
• some Random Numbers Generator (python built-ins)

Process:

1. Set the amount of loops, it must be big for the good simulation: 100.000, 1.000.000, 10.000.000 (amount larger $\rightarrow$ prediction better, not linearly, but logarithmically)
2. In each iteration:
   1. generate random values for the problem
   2. use them for calculating the solution for this numbers
   3. add the result to global result set
3. Compute the mean for the global result set after all iterations

Pros and Cons

Pros:

• easy to code - simply write down the problem logic and simulate it with random numbers many times
• good for cases, when you need only the answer, and not interested in the generalizing
• fast for cases, when you want a "good but not necessary exactly" approximation (~95-99%)