A four-factor composite — momentum, value, quality, low volatility — tested on 113 months of survivorship-free, point-in-time US equity data. It has no measurable ability to predict returns, and its top quintile does not beat simply holding every name in the universe.
And the honest other half: this sample could not have detected a realistic edge anyway. The smallest active premium distinguishable from noise here is ~0.41%/month (4.9%/year) — an enormous number. So we have not proved there is nothing to find. We have shown that we cannot find it, that the point estimates sit within a hair of zero, and that every reason we previously gave for the model "almost working" turned out to be a bug in our own data. All seven hypotheses, and how each one died →
The scan still runs monthly, and here is what it surfaced. Treat it as what it is: a way to narrow ~2,000 names down to a handful worth an hour's reading. It is nota set of calls, and the model that ranked them did not beat a coin flip. Nothing here will ever say "buy".
Scan of 2026-06-30 · 2,450 names scored · prices as of that date
100.0th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
100.0th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.9th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.9th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.8th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.8th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.8th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.7th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.7th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.6th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.6th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.5th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.5th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.5th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.4th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.4th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.3th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.3th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.3th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.2th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.2th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.1th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.1th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.1th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.0th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
99.0th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.9th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.9th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.9th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.8th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.8th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.7th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.7th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.7th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.6th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.6th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.5th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.5th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.5th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.4th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.4th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.3th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.3th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.2th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.2th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.2th percentile. Ranked mainly on 12-1 momentum. This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.1th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.1th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.0th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.0th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
98.0th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.9th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.9th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.8th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.8th percentile. Ranked mainly on 12-1 momentum. This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.8th percentile. Ranked mainly on value (book-to-market + earnings yield). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.7th percentile. Ranked mainly on quality (ROE). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.7th percentile. Ranked mainly on 12-1 momentum. This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.6th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
97.6th percentile. Ranked mainly on low volatility (a defensive tilt). This ranking has no measurable predictive power: mean information coefficient 0.017 (t=1.21) over 113 months, and its top quintile did NOT beat a plain equal-weight portfolio even at zero trading cost (t=0.14). Tested on survivorship-free, point-in-time data. It is a starting point for research and NOT a buy signal.
