What is stats.correct_chi2?

[thehrh]

Given the name and the original PR #682, I feel inclined to use this test statistic. Where can one read about it? What is the statistical justification?

The docstring isn't more enlightening than the original PR, and the statistic isn't documented in https://github.com/icecube/pisa/blob/f63627ce9d370b4ac83687126fd5643f41432afb/pisa/utils/stats.md either.

[philippeller]

It's not a chi2; that's bad naming. It is the log-likelihood for a normal approximation of the Poisson dist.