The instinct problem
Walk into any brokerage app and human nature pulls the same direction: buy the thing that's down 40%, because it's "cheap," and avoid the thing at an all-time high, because you "missed it." It feels like shopping for bargains. In practice it's often catching falling knives while refusing to board moving trains.
Momentum, the tendency of recent winners to keep winning over the next several months, is one of the most persistent effects documented in the academic literature, showing up across countries, decades, and asset classes. One well-known study (George and Hwang, 2004) found that nearness to the 52-week high was itself a powerful predictor: stocks near their yearly high went on to outperform stocks far below it.
Why it keeps working
The most convincing explanation is boringly human: anchoring. Investors treat the old high as a ceiling ("it can't go higher than that"), so good news gets absorbed slowly, and the stock grinds upward for months as reality overrules the anchor. A new high also means every single holder is sitting on a profit: there is no trapped seller overhead waiting to "get back to even," which is the supply that caps most recoveries.
This is the same reason breakout systems like the Turtles' Donchian channel entries made money for decades: every enormous trend, without exception, must pass through new highs on its way up. Buying the high is buying a ticket that every big winner is guaranteed to punch.
A plain set of rules
- Universe: liquid stocks or sector ETFs; skip the illiquid junk.
- Entry: buy when price closes within a few percent of its 52-week high (or makes a new one), ideally with the broad market above its own 200-day average.
- Initial stop: a fixed level below entry that risks no more than 1% of your account.
- Exit: trail a stop under the trend: below a recent swing low, a 20-week low, or a falling moving average. Let the market, not your comfort, end the trade.
- Expect dry holes. Many breakouts can fail quickly. Close them by rule and keep each planned loss small. A sustained trend can be much larger than a planned loss, but no future winner is guaranteed to offset prior attempts.
What can go wrong
Momentum's dark side is well documented too: it crashes hard at sharp market turns, when the previous winners get repriced fastest. That's why the two guardrails above matter: a market-level trend filter and fixed fractional risk. Neither eliminates loss, and gaps or poor liquidity can make a reversal cost more than planned. The purpose is to define and limit exposure, not guarantee the fill.
Want to survey whole fields instead of single wells?
Dual momentum compares broad asset classes once a month and stands aside when none qualifies.