Results

What gets graded

Every projection the board actually served is logged when it is made and reconciled against the final result once the game settles. What appears on Results is that record, per market, over a window you choose — open without an account, and it stays that way whether or not you ever subscribe.

Misses are in there. Not as a footnote: the record is graded / hit / missed as one shape, so a surface cannot show wins without showing the misses beside them. When a week goes badly it is in there too.

A day that has not finished grading says so rather than showing a partial number, and thin samples are labelled with their sample count instead of being presented as a rate you can lean on.

Calibration: the question the record cannot answer

A hit rate tells you what happened. It does not tell you whether the number meant anything. Calibration does: for each market, the graded rows are grouped by the probability we printed and compared against the share that actually happened. Does the group we printed a given probability on land at about that rate?

The Calibration section on Results draws that only where a market has enough graded rows and a fitted curve, and lists the rest with the reason. It is not a hit rate and not a return — it is whether the number meant what it said.

A market without a curve yet shows the model’s raw estimate, and the board says so on the row. That is the honest state for a new league, not a hidden one.

Closing-line value

On the markets where we track it — the MLB batter markets — the record can compare the price at the time against the price at kickoff. That is closing-line value, and it measures one thing only: whether a read moved in the direction the market later moved. It is not a return, and it is not a claim about money you would have made.

The graded record is open to everyone; the closing-line view is free with an account. Every other market’s record is on the page either way.

Why no numbers on this page

Deliberately. Any figure printed here would be a snapshot of the day it was written, and the one you want is today’s. Read it on Results, where it is computed rather than typed — and where you can pick the window yourself rather than taking ours.

How the model changes over time, and why that is stated as a mechanism rather than a promise, is on Methodology.