12-1 momentum strategy: buying the winners, and what the big number hides

A momentum strategy doesn't ask whether a stock is rising, but whether it's rising more than the others. The classic version, "12-1", ranks the entire universe once a month by the return of the last 12 months excluding the most recent month (which tends to reverse), buys the leaders in equal weight, and rotates every month. This is one of the most well-documented phenomena in academic research, since Jegadeesh and Titman in 1993, and still one of the most dangerous.

In our tournament it achieved +188% Over three years, versus +95% the index. Before you get excited: +114% came from in a single year, and the universe chosen today is exactly the universe momentum loves. The explanation below.

The rules we ran

• Ranking: At the start of each month, the return from 12 months ago until one month ago (12-1).
• Portfolio: The top 10 stocks, 10% each, full sell and rebuy on every rotation.
• Regime: When the S&P 500 is below its 200 average, the portfolio is in cash. Without this, momentum tends toward sharp crashes at the start of recoveries.
• Costs: 0.05% per side, 344 trades. The measurement: 184 American stocks and ETFs, daily candles, May 2023 to September 2026 (840 trading days), execution at the open after the signal, cost 0.05% per side, long only and without leverage. The universe is today's and not 2023's — stocks that collapsed and disappeared are not included, and this biases every result upward. The market was rising for most of the period.

The result, and what's behind it

• One year did it all. Without 2024, the year of chips and AI, the strategy lost to the index every other year. A result that depends on a single year is not an advantage, it's an event.
• Survivorship bias here is maximal. The universe is chosen today, and therefore it contains the big winners of recent years and not the losers that were deleted from the indices. Momentum buys exactly the stocks that this universe "knows" will succeed. The true number on a clean universe is significantly lower.
• The one-third decline is the price. A portfolio of the top 10 stocks is a portfolio concentrated in a single sector most of the time. When the sector turns, they all turn together.

Momentum crashes: what the research does agree on

The literature documents an average momentum advantage over decades and many markets, and also documents the other side: "momentum crashes" at turning points, such as 2009, when the previous year's losers jumped hundreds of percent and momentum portfolios were wiped out. The regime filter we ran reduces this, it doesn't eliminate it. Anyone running momentum needs to be prepared for a year in which it loses to the index by a wide margin, and to keep going.