Early-assignment prediction

Open for collection

For 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.

Status: Open for data collection. No published findings yet. We will publish results only when the observation count is sufficient and the dispersion / confidence intervals can be computed honestly.

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:

How to contribute

The observation ingest endpoint is gated by a server-held bearer token (held by Akash Dedhia, Founder & Lead Analyst). External contributors can email hello@optionincometools.com with proposed observations (matching the schema); Akash reviews and ingests personally. Self-reported results without source data are not accepted.

Version history

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.