Across nearly all price levels on resolution markets, the crowd's implied probability overstates the true win rate. In plain terms: YES is systematically a touch too expensive. This is calibration bias, and it is persistent because it is behavioral, not a one-off glitch.
Our Calibration Bias signal blends a theoretical curve with our own resolved-market history, sharpening automatically as more outcomes settle. It fires hardest in the mid-range where no other signal disputes it.
When Calibration Bias is the top contributor to a score, lean NO or sell YES — but always confirm liquidity and expiry first.
Does it work on every market? No. Thin or terminal markets break the assumption; the signal self-suppresses there.