Prediction tracker
Every locked call, and how it settles. Outcome accuracy and calibration are only meaningful once events resolve — until then this is the honest scoreboard: what was predicted, and what is still pending.
Locked
2
predictions
Settled
0
events resolved
Awaiting
2
pre-catalyst
Outcome accuracy
n/a
needs settled data
Calibration
Reliability of outcome confidence — populates as events settle
Calibration measures whether stated confidence matches reality — do events called with medium confidence actually happen about as often as that implies?
It populates as predictions settle. 0 of 2 settled so far — the dashed line is perfect calibration; points will plot predicted vs. realized frequency once there is enough history to bin.