Prediction‑market scoring probe

Prediction markets turn real-world questions into tradable prices, and on Polymarket a 60-cent share means traders collectively price that outcome at about 60%. Polymarket says its prices are set by supply and demand on an order book, not by the company itself. (docs.polymarket.com) The scoring method in the recent analysis starts with a simple gap: actual win rate minus implied price. If contracts priced at 40% go on to win 45% of the time, that five-point spread is the “edge” the model is trying to measure. (xrticles.com) The same write-up says it tested five formulas across about 72 million trades and found a +1.12% mathematical edge for the probability-based approach, versus a -1.12% baseline for discretionary “vibes” trading. The figures come from the analysis itself, not from an independent audit or a Polymarket filing. (xrticles.com) Before any bet is sized, the framework uses an expected-value filter, which is a pass-fail check on whether the estimated payoff is positive after comparing your probability estimate with the market price. On prediction markets, that usually means buying only when your estimated odds are higher than the price implies, after accounting for trading costs. (marketmath.io) (docs.polymarket.com) It then uses the Kelly criterion, a bankroll formula published by Bell Labs researcher John L. Kelly Jr. in 1956, to decide how large a position to take. Kelly aims to maximize long-run growth, but even supporters often use fractional Kelly because full-Kelly sizing can produce large drawdowns. (princeton.edu) (wikipedia.org) That matters on Polymarket because fees and execution can erase thin edges. Polymarket’s current documentation says some categories charge taker fees while geopolitical and world events markets are fee-free, which means the same signal can have different economics depending on where it is traded. (docs.polymarket.com) (polymarket.com) The broader claim is not that markets are easy to beat, but that crowd prices may be slightly miscalibrated in repeatable pockets. A system that scores those pockets, filters for positive expected value, and sizes positions systematically is trying to replace intuition with a repeatable rule set. (xrticles.com) (docs.polymarket.com) Skeptics would still want to know how the backtest handled slippage, changing liquidity, market selection, and whether the signal survives out of sample. The published article describes the formulas and headline results, but it does not carry the weight of a peer-reviewed paper or an exchange-certified performance report. (xrticles.com) For traders, the practical takeaway is narrower than the headline: a small statistical edge only matters if the probability estimate is better than the market’s and the bet size does not blow up the bankroll. That is the whole logic of the probe — find the mispricing, clear the expected-value bar, and wager small enough to survive variance. (xrticles.com) (princeton.edu)