Kyle Cranmer

+1 google scholar, mendeley, impactstory

I couldn't find mentions of open graph social previews in the documentation. I just enabled this with [sphinxext.opengraph ](https://pypi.org/project/sphinxext-opengraph/) and a config like this ``` sphinx: extra_extensions: - sphinxext.opengraph config:...

Note: very small numbers so asymptotics may not set in yet. The systematic variations `ucs_StopRHadron_1100_100000` are both less than the nominal, which may cause some strange behavior. Is there a...

Ok, thanks! [It would be nice to write an example interface for pyhf as well: https://pyhf.readthedocs.io/en/v0.6.3/ ] @oschulz Can I ask an unrelated question (sorry to abuse the GH issue)....

Ok, talking with Philip it sounds like maybe you can write the model to a file with built in Julia serialization. Looking at https://github.com/JuliaIO/JLD.jl I see > For lossless storage...

Seems nice, but I'm wondering about "metrics". It seems like it is introducing a new term or concept. In https://scikit-hep.org/pyhf/_generated/pyhf.infer.hypotest.html#pyhf.infer.hypotest the term `tail_probs` is used, which seems good, though CLs...

Will close #965 which is duplicative.

I'm fine with considering CLs a "modified p-value" I'd also be fine with options being `pvalues=['p_sb', 'p_b','CLs'']` where p is used for the normal p-values and CLs is used for...

p-value as a term doesn't imply any particular distribution, it's the generic term for the tail area of some test statistic. The subscript in p_b and p_sb are indicating what...

Nice.. @lukasheinrich Note, unlike CLs, the -2-sigma band for 95% UL is close to 0. If one wanted the -2 sigma band for the 90% UL it may "go negative"...