---
title: "Did the Chater/Oaksford camp reply to Brighton's 2006 cross-validation reanalysis of the take-the-best result — and does the reversal stand?"
type: "question"
status: "open"
date_raised: "2026-08-01T00:00:00.000Z"
tags: ["rationality","ecological-rationality","rational-analysis","take-the-best","model-selection","verification","corroboration"]
---


[[claim-brighton-2006-reanalysis-reversed-take-the-best-result-against-chater-oaksford]]
records that Chater, Oaksford, Nakisa & Redington's 2003 test of
[[entity-gerd-gigerenzer|Gigerenzer]]'s take-the-best heuristic against complex
"rational-norm" models was reanalyzed by Brighton (2006), who — using
cross-validation instead of fit-then-test-on-the-whole-sample — reported the
original verdict reversed, with take-the-best now winning across all sample
sizes. That account is sourced entirely from the ecological-rationality side
(Gigerenzer & Brighton's own 2009 "Homo Heuristicus").

## Why it matters

This is the empirical leg of the *contested, not compatible* reading
([[observation-ecological-rationality-and-rational-analysis-are-contested-not-compatible]]).
A reanalysis that overturns a result, narrated by the camp whose heuristic wins
under the new method, is exactly the kind of one-sided claim that wants an
independent check before the vault treats the reversal as settled. The
methodological substance — whether cross-validation is the right criterion, or
whether Chater et al. had a defensible reason for their in-sample protocol — is
a live model-selection question, not a foregone conclusion.

## What I'd need to answer it

- Read Brighton (2006) directly: H. Brighton, "Robust inference with simple
  cognitive models," AAAI Spring Symposium (Between a Rock and a Hard Place:
  Cognitive Science Principles Meet AI-Hard Problems), 2006 — confirm the
  reanalysis method and the reversal in his own words.
- Locate any reply from Chater, Oaksford, Nakisa, or Redington to Brighton
  2006 / to the "Homo Heuristicus" characterization — a rejoinder, a later
  paper defending the original protocol, or a commentary. The 2009 "Homo
  Heuristicus" issue of *Topics in Cognitive Science* carried open peer
  commentary (e.g. Hilbig 2010 on the "less-is-more" claims); check whether any
  came from the rational-analysis camp.
- Read Chater et al. (2003), "Fast, frugal, and rational: How rational norms
  explain behavior," *Organizational Behavior and Human Decision Processes*
  90(1), 63–86, directly (only cited via Gigerenzer & Brighton so far).

## Candidate next move

Route to the topic queue as a corroboration task. On resolution: if a reply
exists and contests the cross-validation framing, that is a contradiction note
in its own right; if the reversal is independently confirmed or uncontested,
discharge the `[unverified-mechanism]` flag on the claim-note additively.