52 week high backtest

What the numbers mean

This is trend-following, and the record carries the usual asymmetry: most closed trades lost money, and the total still finished the window ahead because the few that worked ran far longer than the many that didn't. The low win rate is the design, not a fault. Entry wants a close within 5% of the 252-day high with price above the 200-day average, a condition that flickers on and off in choppy advances, and the only exit is a close under the 50-day average: no stop, no target, no holding limit. A name gets bought near its high, shaken out on an ordinary dip, bought back weeks later, each round trip booking a small loss.

The universe is twelve large US names, the book holds at most five, a fifth of capital each, so when signals cluster the portfolio is concentrated and internally correlated: the deep drawdowns are market drawdowns hitting every holding at once, not one bad pick. And a trailing average is not a stop. The gap between price and the 50-day widens as a position runs, so a sharp reversal hands back a large open gain before the exit is ever met.

Read the win rate and the profit factor as a pair; neither says much alone. Check whether the drawdown is shallower than holding these twelve names outright over the span, which is the case for the 200-day filter. Weigh the trade count against a decade and twelve symbols: a high count means the 50-day exit cycles often, which is where unmodeled trading costs would land.