Factor Decay Is Real: How Published Quant Signals Lose Half Their Sharpe and Why Pod Shops Trade Them Anyway
Goldman clocked systematic equity managers down 4.2% in summer 2025, yet Millennium, Citadel, and Point72 closed the year in double digits.
Goldman’s prime brokerage clocked systematic equity managers down 4.2% between June and late July of 2025, during a calendar year when the S&P 500 closed up 16.4%.
They were running models that perform well in a backtest, perform acceptably in production most months, and then surrender a year or two of accumulated gains across a handful of sessions when too many desks try to leave through the same door.
Join 12,500+ quants reading daily breakdowns of hedge fund mechanics and market structure:
That distance between a strategy that validates on historical data and a strategy that survives a crowded exit is the entire subject of today’s piece.
What follows is an account of which classic quant strategies have decayed, what the academic record says the decay actually amounts to, and the structural reasons that Millennium, Citadel, Point72, and their peers keep allocating risk to those strategies long after the standalone edge has thinned.
The short version is that a degraded Sharpe ratio still earns its place inside a multi-manager book, for reasons rooted in correlation, mandate constraints, and regime protection rather than raw return.
The detail behind that, along with what the 2025 crowding unwinds confirmed about the real failure mode, runs below.
