Markets with the largest recent logit-space move versus their own history.
How to read this table
Each market's recent move is measured in log-odds and standardized against its own history. Z = +2 means the market just moved further than ~95% of its typical moves — red = up, green = down. The bar has a zero baseline and a symmetric ±4σ scale. A large |Z| is a research flag, not a recommendation.
Needs a few more collection cycles of history before z-scores populate.
Current probability versus the approximate volume-weighted crowd cost basis.
How to read this table
The mini chart is a volume-at-price profile. The gold bar is the POC — the level where the most contracts changed hands, i.e. the crowd's approximate cost basis. The blue bar is current price. Stretch is the distance between them; conviction is the share of volume within ±5¢ of the POC.
After similar historical unusual moves, how much of the move was later given back within the next window.
How to read this panel
Every historical |Z| ≥ 2 unusual move in collected data is tracked one window forward: did the probability retrace or extend? The split bar is the observed base rate; avg retrace is how much of the typical move was given back (negative = moves kept extending).
Reversion sample still building — this panel appears once enough comparable unusual moves have been observed (terminal settlement moves are excluded).
Current hourly logit volatility compared with similar markets in the same lifecycle and expiry bucket.
How to read this table
σ logit/hr is each market's hourly volatility in log-odds. The × column compares it against the median of markets with similar time-to-expiry, since volatility naturally rises near resolution. 3× = three times as jumpy as lifecycle peers.
How often markets settled YES by final price bucket.
How to read this chart
Settled markets only: contracts priced ~80¢ should resolve YES ~80% of the time. Bars below the dotted diagonal mean that price range settled YES less often than priced; bars above, more often. n = bucket sample size.
perfect calibration
1%25%50%75%100%
actual YES rate when sample is sufficientdotted line = perfect calibrationexpected/implied value when no samplelow sample bucket
Hover any bar for bucket value and sample size. 2 of 100 buckets currently have enough settled samples; muted bars preserve the full probability scale without overstating thin data.