Inspiration
Corporate disclosures describe changes in a company’s financing, leadership and operations. We wanted to investigate whether that information could help investors decide when to collect option premium through covered calls or purchase downside protection through puts.
Massive’s 119 disclosure categories gave us a structured starting point. Our question was whether these labels add useful information beyond prior stock prices and market conditions.
What we built
We built a disclosure research atlas and developed three financing hypotheses: debt issuance, credit facilities, and debt issuance accompanied by an underwriting agreement.
Each hypothesis tests two strategies: funded shares with a short call, and funded shares with a protective put. We compare event-period performance with stock ownership and earlier ordinary periods for the same company.
The covered-call comparison asks:
$$\text{Event-specific increment} = (\text{Call strategy} - \text{Stock}){\text{event}} - (\text{Call strategy} - \text{Stock}){\text{ordinary}}$$
This helps distinguish a disclosure-related opportunity from the usual effects of owning stock and selling calls.
How we built it
We used Massive historical disclosures, stock prices, option data and corporate-action records. We committed the economic hypotheses before implementing the strategies, prepared data locally, and ran scientific calculations through scheduled HiPerGator jobs.
Our backtests include transaction-cost assumptions, liquidity checks, dividends, modeled assignment and portfolio limits. Verified results return to the computer with file hashes and execution records.
What we learned and the challenges
The broad stock research produced exploratory leads, but better prediction does not automatically mean profitable trading.
Historical option coverage and matched comparisons proved challenging. The completed pilot left too few usable comparisons to establish alpha. Our statistical method also failed its synthetic calibration, so we withheld significance claims.
The project delivers implemented strategies, completed pilot backtests and a reproducible framework for further testing. Its current conclusion is inconclusive, with the evidence and remaining work recorded explicitly.
Log in or sign up for Devpost to join the conversation.