Biased Traders, Honest Prices: The Psychology Inside Prediction Markets

Every trader who opens a prediction market brings the same equipment: a human brain, with its factory-installed biases. We fall for the favorite-longshot bias, paying too much for improbable outcomes because the payoff feels exciting. We are overconfident in our own read of the news, we overweight whatever happened this morning, and we herd into whatever price is already moving. If markets were made of our individual judgments, the prices would be a mess. Mostly, they are not.

The biases are not theory; they are measurable in the tape. In a 2026 study of 72.1 million Kalshi trades, contracts bought at one cent won just 0.43% of the time against a 1% implied probability, meaning longshot buyers overpaid by more than half. At five cents, the win rate was 4.18% against 5% implied. There is also a YES optimism tax: traders disproportionately buy the affirmative side, and YES contracts underperformed NO at 69 of 99 price levels. Dollar-weighted, YES buyers earned excess returns of -1.02% while NO buyers earned +0.83%, a gap of nearly two points paid for the privilege of agreeing.

The losses are not evenly shared; they flow in a predictable direction. Karl Whelan's study of more than 300,000 Kalshi contracts found an average pre-fee return of roughly -20%, concentrated among takers, the traders who cross the spread with market orders, and among longshot buyers. Makers, the traders who rest limit orders and wait, earned positive returns. On Polymarket, across 588 million trades, the top 1% of accounts captured about 76.5% of the profit, predominantly by providing liquidity. The crowd's biases are, in effect, a standing transfer from the impatient to the patient.

So why does the final price stay honest? Because the mechanism taxes exaggeration. A poll costs nothing to answer; a position costs money to be wrong, and that single fact disciplines overconfidence, anchoring and herding better than any survey design. Biased flows also cancel: for every recency-driven buyer chasing the headline there is a maker happy to sell into the spike. Kalshi Research, grading 686,272 trades, found the market's closing prices scored a Brier score of 0.063, accuracy that rivals professional forecasting operations. The bias lives at the tails and in the taking. The price in the middle belongs to whoever was willing to be bored and right.

The lesson is not that prediction markets have cured human psychology. It is that they have built a machine that runs on it. The probability on the screen is the crowd's mind after its worst impulses have been charged a fee, and the toll is paid, trade after trade, by whoever insists on being the hero of the story instead of its bookmaker.

Sources: Becker Kalshi trade study (2026); Whelan, Makers and Takers, CEPR; Kalshi Research; Polymarket maker-taker study, SSRN
#infiniti #infinitimarkets #predictionmarkets #behavioralfinance #polymarket #kalshi
web: Infiniti.w3ai.io
Infiniti Prediction Markets
https://infiniti.w3ai.io/intelligence/biased-traders-honest-prices-the-psychology-inside-prediction-markets