Play Money And Reputation Systems

A focused mechanism-design essay on play-money and reputation prediction-market systems. Two core design axes: relative vs absolute accuracy, and zero- vs positive-sum. The sharp payoff is the Susan/Randy example — in Metaculus's positive-sum system, a lazy grinder answering 360 questions can out-score a superforecaster who nails one, because the maximally-correct answer isn't rewarded 360x more; hence 'reputation systems' don't actually produce reputation (nobody drops their Metaculus score in conversation). He then defends zero-sum relative-accuracy play money (Manifold) as a reputation-in-disguise, dissects the mispricing problem (the wrestler-9%-president, his own book-review conditional markets where people 'vote' by buying yes but never no), and proposes a per-market interest-free loan to unlock long-term-market arbitrage without inflating reputations. Genuinely insightful forecasting-platform design.

Quality 68 · Strong. Strong-floor (68). Original, focused mechanism-design reasoning (the positive-sum-rewards-grinding critique, the reputation-doesn't-produce-reputation insight, the loan fix); a self-contained analytical essay well above the roundups, but narrow.

Claude’s paradigm shift 48 · Moderate. Moderate (48). Sharpens real design frames (relative/absolute, zero/positive-sum) without originating them.

Real-world impact 1 · Negligible. 1 — a design-analysis blog post (Manifold did later add a loan feature); negligible durable effect from this post itself.