I have data pertaining to an e-sport, such as player names, their kill count, their death count, their rounds won, their matches won, etc

Currently I'm only able to say playerX wins the map N% of the time when their kill count is >< than K, or teamX wins the map N% of the time when playing opponentY.

I want to offer more than win-rates. I only know basic statistics, so I don't really know what else I can do.

I'd like to be able to predict the outcomes of matches and things like that based on previous data, and generally show more interesting statistics based on Z-scores and such

Basically looking for some input on this from people with more statistical knowledge than myself :)

Thank you!

  • 1
    $\begingroup$ To predict win probabilities I'd probably start by looking at Bradley-Terry type models in a logistic regression. $\endgroup$
    – Glen_b
    Commented Aug 19, 2015 at 0:47

1 Answer 1


"Basic" is a relative word, so I'm not sure what you mean by "I only know basic statistics." I'll assume that you don't know anything about modelling or regression and are simply able to create some additional statistics. Given that, you are fairly limited in what you could do. One thing you might do is form ratios such as #kills/# of deaths or #kills/$round played. This will give you a sense of the expected number kills for each life and the expected number of kills per round.

If you wanted to do something predictive, you'd probably need to use more "advanced" statistics like logistic regression, which could give you the predicted probably that the player wins a round, given their historical statistics of number of matches played, number of kills, number of deaths, etc.

  • $\begingroup$ Great thank you. I'll look into logistic regression. $\endgroup$
    – Plex ASM
    Commented Aug 18, 2015 at 22:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.