Postfit

Postfit is a term used in statistics and data analysis to describe results obtained after a model has been fitted to data. The fitting procedure typically involves maximizing a likelihood function or computing a Bayesian posterior, yielding best-fit parameter values and their uncertainties. Postfit results reflect the information contributed by the observed data and any prior assumptions used in the fit.

In high-energy physics and related experimental fields, postfit results are commonly presented as post-fit yields and post-fit distributions. These show how the expected counts for signal and background processes, as well as the shapes of distributions, have been adjusted after comparing the model to the data. The post-fit view often includes revised background normalizations and shape parameters that are informed by the observed events in different analysis categories.

A key feature of postfit analyses is the treatment of nuisance parameters, which encode systematic uncertainties from experimental conditions, detector effects, and theoretical models. After the fit, nuisance parameters are constrained by the data, producing pulls and reduced uncertainties that combine into the post-fit error estimates and correlations. This contrasts with pre-fit expectations, which are based on nominal values before observing the data.

Postfit results are used to interpret statistical evidence for signals, estimate backgrounds, and quantify uncertainties. They depend on the chosen model, priors, and fit configuration, so results are typically accompanied by diagnostic information and discussions of fit quality to avoid over-interpretation.