Proof of performance
Nothing on this desk is a private idea. Each model implements an effect that was first documented in peer-reviewed finance research, replicated out-of-sample, and tested across decades and markets. Below is the documented annualised excess return of each effect, the sample it was measured on, and the study it comes from.
This is evidence that the rules have worked over long horizons — not a promise about any single week. Live results always differ: spreads, taxes, timing and a shorter horizon all matter.
Every lens on this desk is a published, replicated finance result — not a private idea. The bars show the annualised excess return each effect produced in its original academic study. That is evidence the rules work over long periods; it is not a promise about any single week.
Leading Markets· Cross-asset momentum+7.8%/yr
Buying the strongest markets and avoiding the weakest produced a persistent premium in every asset class tested — equity indices, currencies, bonds and commodities.
Asness, Moskowitz & Pedersen (2013), Journal of Finance — "Value and Momentum Everywhere" · 1972–2011, 58 markets
Leading Stocks· 12-1 stock momentum+9.1%/yr
Stocks in the top relative-strength decile kept outperforming for 3–12 months. The result has survived 30+ years of out-of-sample data and replication across 40 countries.
Jegadeesh & Titman (1993), Journal of Finance — "Returns to Buying Winners and Selling Losers" · 1965–1989, US stocks (out-of-sample since)
Valuation Screener· Value + quality (profitability)+6.4%/yr
Cheap alone is weak. Cheap combined with high gross profitability roughly doubles the premium versus a naive low-multiple screen — the exact gate this screener applies.
Fama & French (2015) five-factor model; Novy-Marx (2013), Journal of Financial Economics · 1963–2013, US large caps
Insider Buying· Form 4 open-market purchases+6.2%/yr
Company insiders' own-money purchases predict returns, and "opportunistic" (non-routine) buyers deliver about 82bp/month of abnormal return. Cluster buys are the strongest form.
Lakonishok & Lee (2001), Review of Financial Studies; Cohen, Malloy & Pomorski (2012), Journal of Finance · 1975–1995, all US insiders
13F Superinvestors· Best-idea institutional holdings+5.5%/yr
A manager's highest-conviction position outperformed both the market and the rest of that same manager's portfolio — which is why we rank by conviction, not by portfolio size.
Cohen, Polk & Silli (2010) — "Best Ideas"; Wermers (2000), Journal of Finance · 1980–2005, US managers
Seasonality· Calendar / Halloween effect+4.5%/yr
Nov–Apr returns exceeded May–Oct in 81 of 108 markets over three centuries. Weak on its own, useful as a tilt on top of another signal — which is how the desk uses it.
Bouman & Jacobsen (2002), American Economic Review; Jacobsen & Zhang (2013) · 1694–2009, 108 markets
Dark Pools· Off-exchange informed flow+3.1%/yr
Off-exchange volume carries measurable information about future price direction, but only in relative terms versus a stock's own history — never as an absolute level.
Comerton-Forde & Putniņš (2015), Journal of Financial Economics; Zhu (2014), RFS · 2008–2013, US equities
Premia are as reported in the cited papers, gross of costs, over the sample windows shown. Live results differ: real spreads, taxes and a shorter horizon all matter. Past evidence is not a forecast — the point is that each lens rests on documented, repeatedly tested behaviour rather than opinion.
Want to see how the rules read the market this week?