Some material about Statistics for HEP:

My BLUE page is here.

Unfolding:

• My posts on the AMVA4NewPhysics blog:
• Lost in unfolded space - part I
• Lost in unfolded space - part II: "and then I accidentally divided by zero..."
• Lost in unfolded space - part III: "Let a hundred flowers bloom, let a hundred schools of thought contend" (Mao Tse-Tung)
• Lost in unfolded space - last episode: "Thou shalt not unfold"

Links:

• Barlow's tools (asymmetric errors, etc.)
• P.Bhat, Multivariate Analysis Methods in Particle Physics (2011)
• G.D.Lafferty, T.R.Wyatt, Where to stick your data points: The treatment of measurements within wide bins
• Data Science @ LHC 2015 Workshop - Reading materials
• Proceedings of the Higgs Machine Learning challenge
• Bob Cousin's review of the approaches to deal with physical constraints
• Unfolding by Machine Learning (see also CMS-only talk here)

CMS-specific links:

• CMS statistics questionnaire
• StatComm FAQ
• Poisson error bars in ROOT
• StatComm recommended references
• MVA recommendations
• Search-related:
• Search procedures
• Recommendation for the definition of the sensitivity and reach of a CMS analysis
• Look-Elsewhere effect
• TOP PAG blinding recommendations
• Unfolding:
• TOP PAG unfolding wiki
• Bob Cousin's interim recommendations on unfolding
• Topical CMS Statistics Committee Meeting, Dec.11, 2015
• Topical TOP PAG Meeting, May 13, 2016
• DIANA meeting, Nov. 13, 2017
• Talk on Unfolding by Machine Learning

ROOT commands for 1-tailed conversions (zvalue: "number of sigmas"):

zvalue = sqrt(2)*TMath::ErfInverse(1 - 2*pvalue)  
pvalue = 0.5 - 0.5*TMath::Erf(zvalue/sqrt(2))