guessableestimable

Guessableestimable is a coined term used in statistics and related fields to describe a quantity that is both guessable and estimable. In this usage, a quantity is considered guessable if, given relevant context, prior information, or heuristic knowledge, there exists a reasonable initial guess or decision rule that yields a nontrivial chance of being correct. It is considered estimable if there exists a statistical estimator based on data that converges to the true value (or to a well-defined target) under standard assumptions, such as consistency or unbiasedness with finite variance.

Formally, for a parameter theta, estimability refers to the existence of an estimator theta_hat_n based on a sample of size n that converges to theta as n grows. Guessability adds a practical criterion: using domain knowledge, a rule of thumb, or priors that produce a usable guess before or alongside data collection, with risk or error bounded in a meaningful way. When both properties hold, the quantity is said to be guessableestimable, meaning that informed guessing and data-driven estimation reinforce each other.

The concept highlights a pragmatic view of inference: useful parameters are those for which prior understanding or intuition can guide initial judgments and where data can subsequently refine those judgments through reliable estimation. It is relevant in model selection, parameter tuning, and interactive estimation tasks where quick guesses are valuable but should be backed by formal estimation.

Limitations include that not all estimable quantities are guessable without substantial prior knowledge, and not all guessable quantities yield robust estimators if data are limited or model assumptions fail. See also estimability, identifiability, and predictive inference.