Energy
Factor t-stat 3.87, decayed 56% since publication. The anomaly that refuses to die — still significant out-of-sample and across countries and asset classes. Its weakness is tail risk: violent crashes in 1932, 2009 and 2020 when beaten-down stocks snap back.
Factor t-stat 4.2, decayed 100% since publication. Statistically enormous and statistically real, but the turnover is ~1200%/year. After realistic bid/ask and market-impact costs the profit is close to zero for anyone trading size. It is a liquidity-provision fee, not a free lunch.
Factor t-stat 2.03, decayed 95% since publication. Largely subsumed by the value factor once that was formalised in 1992; as a standalone signal it has been weak for decades.
Factor t-stat 1.07, decayed 116% since publication. 'Low-vol' became a marketed product category (min-vol ETFs) after ~2011, which arguably crowded the trade and led to a sharp 2016–2020 underperformance.
Factor t-stat 1.68, decayed 94% since publication. The effect is concentrated in the very short leg — expensive, hard-to-short small caps — and much of it overlaps the low-beta and profitability stories. Weaker since the mid-2000s.
Factor t-stat -0.59, decayed -57% since publication. The formal BAB factor is post-2014, but the low-beta effect was documented in 1972 and has been remarkably persistent. Critics (Novy-Marx & Velikov) argue the AQR construction leans on tiny illiquid stocks.
Factor t-stat 6.08, decayed 31% since publication. Moskowitz–Grinblatt claimed industry momentum explained most of individual-stock momentum; Grundy–Martin (2001) pushed back hard. As a standalone it has been weak since the late 1990s.
Raw numbers behind the scores
Scores are 0-100 percentiles within the current 496-stock universe, oriented so 100 means most exposed to whichever side of that factor has historically earned the premium — a description of where this stock sits today, not a prediction of where it goes next. Click a factor above for the full replication: its history, decade-by-decade record, and whether it survives multiple testing.
Scores computed from price history only. This screener does not yet cover value, quality, size, or dividend-yield style factors — see the screener page for why.