A cutoff date is the date after which no information is used in producing a figure. Every estimate is as of its cutoff date, and a figure restated after that date is a new figure with a new cutoff.
The cutoff gives a number its place in time. Records retrieved, labels extracted and features computed are all limited to what existed on or before the cutoff. A court that revises a docket afterward does not change a figure already issued. It creates the need for a new one.
Cutoff discipline is what lets two parties talk about the same number. It also makes duration drift measurable, since the movement of a distribution between two cutoff dates is exactly the change in what was known.
A cutoff date is not the date a figure was published. A figure can be published weeks after its cutoff and still be exact about what it knew.
Working a file that turns on this term?
Working through a diligence process?
Institutional partners evaluating a position against this platform's outcome and duration models are welcome to reach out directly.
