This page used to present FilingDrift's score as a return factor: sort companies by how much their filing language changes year over year, and the most-stable-language group appears to outperform. The in-sample numbers were real, and they reproduced under a range of factor adjustments.
We went back and stress-tested that claim properly against survivorship bias — the fact that a universe of companies that still file quietly excludes the ones that failed and delisted. When we rebuild the universe with the delisted names added back and model their outcomes honestly, the apparent edge does not survive as a tradeable, all-cap return. What remains concentrates in the smallest, least liquid names, and the ranking inverts above roughly $300M in market cap. That is not something we're willing to sell as an investment signal.
So we retired the factor/return positioning entirely. We'd rather correct it in public than leave a number up that we no longer stand behind.
What FilingDrift actually is: a transparent, reproducible screen for unusual year-over-year language change in SEC filings. It surfaces the filings worth reading and shows you exactly which passages moved — a first pass for credit, distress, and diligence work, not a prediction of returns. See the methodology and honest scorecard, or the case studies.
Not investment advice. We analyze language; the interpretation is yours.
Questions: hello@filingdrift.com