Early-assignment prediction
Open for collectionFor each short option position that crosses an ex-dividend (for calls) or interest-rate-sensitive window (for puts), we record the model's prediction (assign / not-assign) and the actual outcome. The result is a confusion-matrix calibration of our early-exercise incentive heuristic.
Methodology
On the day before ex-dividend, the model computes the early-assignment incentive (dividend minus remaining extrinsic value for short calls; interest income minus remaining extrinsic for short puts) and predicts whether early assignment will occur. The actual outcome is recorded on the following business day. False positives (predicted assign, did not assign) and false negatives (predicted no assign, but did assign) are explicitly captured for downstream model recalibration.
Dataset schema
The observation schema is published as a versioned JSON Schema: early-assignment.json. Every observation contains the fields listed there. The schema is governed by the methodology version and bumped semver-style when materially changed.
Public exports
The complete observation dataset is exportable at any time:
- CSV: /research/dataset?study=early-assignment-prediction&format=csv
- JSON: /research/dataset?study=early-assignment-prediction&format=json
- Live observation count: /research/stats?study=early-assignment-prediction
How to contribute
The observation ingest endpoint is gated by a server-held bearer token (held by the editorial team). External contributors can email [email protected] with proposed observations (matching the schema) and we will review and ingest. We do not accept self-reported results without source data.
Version history
- v1.0 (2026-06-23): Initial schema published. Data collection opens.
Suggested citation (once published)
Dedhia, A. (Early-assignment prediction). OptionIncomeTools Research. Retrieved from https://optionincometools.com/research/early-assignment-prediction/
See also: Research program · References & citations · Corrections log · Disclaimer.