Sections Introduction Likelihood and sprt Rating Distribution Uniform Prior Use of the Distribution Outcome Models Appendix
to do Five outcomes of a game pair
For a random variable with five possible outcomes, where is the probability of observing outcome i and , the pentanomial distribution gives the probability of sampling the random variable n times and observing each of the outcomes times:
where .
Given some observations and a model that is a function of the rating difference diff, we can consider
The quantity is the likelihood, ...
to do Likelihood ratio
to do Sequential probability ratio testing
Sprt requires two hypotheses to compare – that the rating difference is or some particular – but you may only care that a patch is better at all.
There is implicitly a uniform prior?
We may also be interested in finding
where
leaving only the prior to be determined.
This is a distribution, where is the probability that the difference in rating is x given the observed outcomes obs.
We can use as a prior a uniform distribution
over some interval , and in particular, we can set and for some L so that over the support.
Then
and L cancels, giving us
which is simply normalized to have unit area over the support.
We can then consider the limit
which converges. Then converges, and in this sense it is reasonable to talk about a “uniform prior over all the integers”.
It is, of course, not necessary to use a uniform prior, but in the absence of any reason to assume a different prior it is perhaps the best default.
to do Conditional distribution function
to do Stopping when the cdf at zero is less than some threshold
The cdf at t is
and you might say “stop when the probability that the rating difference is negative is less than 5%” or whatever.
to do Equivalence of exit advantage and rating difference
to do Trinomial game outcome
to do Pentanomial game pair outcome
Model is perhaps more properly probability of pair outcome given rating difference, book bias, and book drawishness.
to do Calculating multinomials