Interlude for Behavioral Economics

A lucid, witty tour of experimental game theory. Friend-or-Foe (~45% cooperate regardless of stakes; tit-for-expected-tat; demographic expectations becoming self-fulfilling). The 'spoiled' Prisoner's Dilemma (secretly tell P2 what P1 chose) shows folk-superrationality beats fairness: cooperation falls from 37% to 16% once P2 learns P1 cooperated, because ~21% will only cooperate under uncertainty about the other's move. Ultimatum Game (stakes-invariant rejection of unfair offers, the Machiguenga exception, the testosterone correlate). Dictator Game (28.35% mean over 41k games; the Dana/Cain/Dawes finding that a third will pay $1 to make the recipient vanish unaware — generosity is partly about being seen). Idea-dense and entertaining, but primarily curation of others' experiments.

Quality 73 · Strong. Strong-tier exposition — memorable, fun, informative, with the genuinely sharp spoiled-PD framing that distinguishes superrationality from fairness. Capped at 73 because it is an explicit 'interlude' reporting published findings (Art of Strategy, meta-analyses) rather than original work.

Claude’s paradigm shift 38 · Slight. Low novelty — these are reported empirical findings, not Scott's own ideas; the value-add is presentation and the superrationality-vs-fairness framing of the spoiler experiment.

Real-world impact 2 · Minor. A memorable, fun tour of experimental game theory with the sharp spoiled-Prisoner's-Dilemma framing that distinguishes folk-superrationality from fairness. Conceptual/pedagogical influence within rationalist discourse, reporting published findings, no material change — low RWI.