Every stock is scored using a multi-factor quant model that grades five key dimensions:
Each factor receives a letter grade (A+ through F) based on where a stock ranks relative to the full universe. The weighted grades combine into an overall quant rating from 1 to 5.
A stock must clear specific minimum grade thresholds to be considered a buy:
The portfolio holds up to 50 positions, each new buy sized around 3.5% of the book. We add the single best qualifying name each cycle and let winners run — conviction adds on names up 30%+, with weight caps to manage concentration. New picks are evaluated every other week.
We don't use fixed holding periods or trailing stops. A stock is sold when its quant rating falls below 2.5, with a full exit below 1.5. Big winners (up 60%+) may be partially sold to protect gains, and positions that stay underwater for 150+ days with a weak rating are force-sold to clear out stale losers.
Prices are refreshed daily after market close. Full fundamental data (financials, estimates, revisions) is refreshed weekly. The system continuously rescores the entire stock universe so that grades reflect the latest available data.