Subject: How long before a published screen “stops working”
Academic evidence: Across 97 anomalies the excess return falls by ≈ 32 % after publication as capital crowds in. It does not go to zero.
SSRN

O’Shaughnessy’s experience:

Micro‑cap/value + momentum combinations (Tiny Titans, SCGV) have retained meaningful alpha for > 25 years because they trade in the least‑arbitraged corner of the market. Capacity is tiny, so the big money has stayed away.

Large‑universe multifactor blends (Trending Value) felt the crowding first; the edge was largely gone within ~5 years, then lay dormant for a decade until the style cycle turned in 2022.

Low‑volatility yield screens never produced huge alpha to begin with, so the post‑publication haircut was small.

Practical take‑aways
Expect a half‑life, not instant death. History suggests a newly public factor loses roughly one‑third of its edge within the first 3–5 years as it becomes crowded.

Capacity matters. Strategies that live in illiquid micro‑caps can keep working for decades because professional money can’t scale into them.

Cycles trump publication. All value‑based screens struggled from 2010‑2020; that drought swamped any publication effect. When the style cycle reversed, so did performance.

Keep re‑testing. Simple tweaks (industry‑neutral value ranks, shorter momentum look‑backs) restored Trending Value’s edge in out‑of‑sample tests
Portfolio123 Blog
. Factor decay is often a signal that refinements—not abandonment—are needed.

Size your bets for volatility. Tiny Titans has delivered the biggest long‑term reward but also the deepest drawdowns (50 ‑ 70 %). Allocate accordingly or pair it with defensive factors.

Bottom line:
Some of James O’Shaughnessy’s screens—especially those that fish in micro‑cap ponds—have kept beating the market for more than two decades after he revealed them. Others, geared to larger, better‑arbitraged universes, lost their edge within about five years but can revive when style winds shift. Expect every published edge to fade, but not necessarily disappear; life‑cycle management (periodic re‑validation, capacity awareness, and factor diversification) is what keeps these quantitatively simple screens relevant today.