probtum

Probtum is a neologism used in probability theory and Bayesian reasoning to denote a unit of evidential strength. It is not part of standard statistical nomenclature and has no official definition in major statistics handbooks. The term appears in informal discussions and educational materials as a way to describe how strongly data support one hypothesis over another.

Definition. A probtum is defined as the natural logarithm of the Bayes factor comparing two competing hypotheses H1 and H0 given observed data. In Bayesian updating, the Bayes factor multiplies prior odds to give posterior odds; the probtum measures the log-scale change in evidence produced by the data. By convention, one probtum equals ln(BF) = 1, which corresponds to a Bayes factor of e ≈ 2.718. Positive values favor H1, negative values favor H0.

Relation and interpretation. The probtum is conceptually related to the log-likelihood ratio and to information-theoretic measures, but it remains informal and is not standardized. It provides a continuous sense of evidence strength, unlike binary significance tests. Its numerical interpretation depends on prior choices through the Bayes factor, so cross-study comparisons require careful alignment of assumptions.

Origin and usage. Probtum appears as a descriptive term in online discussions and teaching materials rather than as an established statistical unit. Critics note that, without standard conventions for priors and reporting, probtums can be misleading if compared across different analyses. Related concepts include Bayes factors, likelihood ratios, and log-odds.