Comparing probability distributions over discrete random variables

Hi guys

Sorry if the post is not in the right place but I am new here. I am comparing the performance of two monte carlo methods.

The results I got are the marginals for the same set of random variables. Lets say for example:

First Method Second Method
X1 0.62 0.58

X2 0.71 0.75

X3 0.412 0.43

X4 0.57 0.60

I need to measure how different they are since I will be comparing a set of methods.
At first I thought of K-L divergence but it does not apply here since I will have a lot of zeros.

Any suggestion?

thanks in advance