From time to time I enjoy making some theorycrafting and implementing some calculators for games and my decisions in them.

But this time another theorycrafting ended up on computational side of things way more than usual. So, I decided to also publish my results here.

League of Legends ARAM champion selection math & statistics

This is the article about congitive biases in random champion selection and about the model that is used to estimate likeliness of certain events in a sequence of games played. Used model is DP over number of games and a vector of current distibution of champions, that uses hypergeometric distibution to calculate transitions. Model itself is impelemented in JS inside browser page to make it easily customizeable for everyone.

I'd be happy to see criticism and suggestions, especially about mathematical and implementational part (see "Approach and model details").