Trend Following in Stocks: Why It Doesn't Work (And What Does)

Andreas Clenow wrote a blog post titled “Trend Following Does Not Work on Stocks.” It got people’s attention. It also got him some angry emails from people who thought he was dismissing the whole concept of following price momentum in equities. He wasn’t. He was making a far more specific point, and it’s one that matters a lot if you’re trying to build a stock-based momentum system.

Andreas is a hedge fund manager and the author of two books, Following the Trend and Stocks on the Move. He spent years at Reuters and Equus International before launching his own fund, and he now serves as chief investment officer at Asus. His background straddles quant modeling, IT, and active fund management, which gives him a particular vantage point on the gap between how retail traders think about position sizing and how institutions actually do it.

In this episode we dig into why you can’t simply copy a futures-based trend model onto stocks, what momentum in equities actually requires, and how professional risk allocation differs from what most traders practice.

Watch the full episode below, then read on for the complete breakdown.

Table of Contents

Why traditional trend following fails on stocks

The core argument Andreas makes is straightforward: trend following as a methodology was built for diversified futures trading, where you’re running a model across oil, bonds, currencies, grains, and metals simultaneously. The diversification across uncorrelated assets is what makes the strategy viable. Most individual positions lose. A small number of big winners offset everything else.

When you apply that same model to a basket of stocks, you lose the diversification benefit almost entirely. Stocks move together. A biotech and a utility don’t behave like oil and soybeans. They share enormous beta exposure to the same market index. If the S&P drops 3%, almost everything in your stock portfolio drops with it.

“You can’t just take a trend model built for diversified futures trading and apply it on stocks,” Andreas explains. “You have to take into account that the stocks are very different. They behave very differently, and they behave very much the same.”

That last line captures the paradox. Individual stocks are highly varied in their sector exposure, volatility, and business fundamentals. But at the portfolio level, they’re heavily correlated. A standard trend model running on random stocks produces, in his words, “horrible performance.”

Momentum vs. trend following: the practical difference

Andreas uses the term “momentum” rather than “trend following” for his stock strategies, not just for semantic reasons. The implementation is genuinely different.

In a traditional trend following approach, you hold a position as long as price continues moving in your direction. You exit when it hits your stop. Whether a position is performing better or worse than other positions isn’t really the point.

In a momentum strategy for stocks, Andreas is constantly ranking stocks by their adjusted momentum score and rotating into the best performers. A stock can be rising and still get removed from the portfolio because other stocks are rising faster. “You move out of stocks that are not performing among the best stocks anymore, and you move into the best ones.”

This creates a continuously rebalancing portfolio that’s chasing relative strength rather than absolute direction. The mechanics are more complex. The trading frequency is higher. And stop-loss points work differently. Andreas shows in his book model that you don’t necessarily need individual stop-losses because the rotation mechanism itself provides an exit when a stock’s momentum deteriorates relative to alternatives.

Market state: the factor most stock trend models ignore

One thing that separates stock momentum strategies from futures trend following is the need to account for overall market direction. When you’re trading a diversified futures portfolio, you’re naturally long some things and short others across uncorrelated markets. The direction of any single index doesn’t matter much.

With stocks, you’re almost always net long. You can’t run the same strategy in a rising market and a bear market and expect similar results. “You can’t trade the same whether the index is going up or down,” Andreas says. “Normal trend following, you have no such factor to care about.”

His approach in Stocks on the Move incorporates a market filter that adjusts positioning based on the state of the broader index. When conditions deteriorate, the model moves to cash rather than fighting against a falling tide.

Risk allocation vs. cash allocation: the institutional approach

This is where Andreas gets into territory that surprises a lot of retail traders. The standard approach to position sizing in a stock portfolio is to divide capital evenly: 20 stocks, 5% each. It’s simple and intuitive, but it’s not how professionals do it.

The problem with equal cash allocation is that it ignores volatility. A highly volatile biotech moving 4-5% a day and a utility moving 0.5% a day represent completely different risk exposures if you give them the same dollar allocation. Your portfolio risk ends up concentrated in the volatile names by default, even if you didn’t intend that.

Andreas allocates by risk, not cash. He measures the volatility of each stock and sizes positions so that each one contributes approximately equal risk to the portfolio. “Cash is not risk here,” he says bluntly. A smaller cash position in a volatile stock can represent the same risk as a larger cash position in a quiet one.

Position pyramiding gets similar treatment. The idea of doubling down because a position has moved in your favor and you’re now “playing with the bank’s money” doesn’t make sense to him from a probability standpoint. “The basis for the trade hasn’t really changed. It’s not like you have a higher probability of future gains now.” The gain on a position doesn’t change what the trade is likely to do next.

Rebalancing frequency and trading costs

As volatility changes over time, position sizes need to adjust. Andreas does this roughly every two weeks in the book model, with a 5% threshold for triggering a rebalance. If a position size hasn’t changed by more than 5%, he leaves it alone. More frequent rebalancing runs into trading costs, and costs are not trivial at the retail level.

At the institutional level, where transaction costs are minimal, you can rebalance more aggressively. Retail traders need to be more careful. There’s also a tax dimension: frequent rebalancing can trigger short-term capital gains treatment in some jurisdictions, which changes the math on whether constant size adjustments are worth it.

The broader principle is that position sizing should respond to actual risk, not to convenience or round numbers.

What hedge funds actually look for (that retail traders often miss)

Andreas makes an interesting observation about the gap between institutional and retail risk management. There’s a common assumption that hedge funds take enormous risks to generate their returns. His experience is the opposite. Funds like the one he manages target much more modest return profiles than most retail traders pursue, but they do it with tight risk controls.

“We aim for much lower returns, I would say, than a lot of what I read about that retail traders use.” What they give up in headline returns, they recoup in consistency and survivability over long periods.

The institutional approach isn’t built around finding one great system and riding it. It’s built around managing risk per trade, portfolio-level exposure, and drawdown control in a disciplined, repeatable way.

Key takeaways from this episode

Futures trend following and stock momentum are different strategies with different mechanics. Don’t copy one onto the other without adapting.
Stocks are highly correlated, so you can’t use diversification the same way you can across futures markets.
Market state matters for stocks in a way it doesn’t for diversified futures portfolios. You need some mechanism to reduce exposure in bear markets.
Equal cash allocation is not equal risk allocation. Size positions by volatility, not by dollars.
Position pyramiding based on unrealized gains doesn’t change the underlying probabilities of a trade.
Rebalancing frequency should be determined by trading costs and the size of volatility changes, not by habit.

Get the show notes & transcript

Related episodes

Want to learn more about building better trading systems? Subscribe to the Better System Trader podcast for weekly interviews with the world’s top systematic traders and quantitative researchers.