PM Atlas PM Atlas home

Concept · information-theory

Longshot Bias

The tendency for low-probability outcomes to be systematically overpriced in prediction markets · bettors pay too much for tail outcomes (relative to true probability), so longshots have negative expected return. Originally documented in horse racing; appears in modern PMs too, though 2026 research argues a confounding "yes bias" may explain much of what's been attributed to favorite-longshot.

Key insights

In their words

5c contracts win only 4.18% of the time.· Ranger Global
Takers disproportionately buy YES longshots, accepting returns 64 percentage points lower than equivalent NO positions.· Jonathan Becker
Whales are not the sharpest participants: heavily capitalized traders systematically bleed expected value to small-order traders.· Deleep et al.

Where it matters

Longshot bias is the single most exploitable systematic mispricing in PMs · and the 2026 research suggests it's been underestimated because it's been entangled with YES/NO labeling effects. If a PM builder wants to claim "accurate prices," longshot bias must be measured per-platform and per-category. For aggressive traders, the bias is renewable arbitrage (sell longshots, especially YES longshots, on Kalshi entertainment/media categories). For Dekant, the design implication is more subtle: a distribution-market curve doesn't inherit YES/NO labeling at all, so the entire longshot-vs-favorite asymmetry has to be re-examined in continuous space · but the same retail behavior (overpaying for tail outcomes) will reappear as overweighting the tails of the drawn distribution.

Connections

Platforms linked to this concept

Related concepts

Sources

Open in the interactive atlas →