Andreas Clenow — Trend Following Is About Taking A Lot Of Bets On A Very Large Number Of Markets

ELI5/TLDR

Two trend followers, Andreas Clenow and Meb Faber, sit down for an hour. The single line that holds it all together: trend following is not a strategy, it’s a portfolio. You take small, identical bets on as many uncorrelated markets as you can find, lose on most of them, and wait for the few that keep running to pay for everyone else. The entry rules barely matter. Diversification and position sizing do almost all the work. Most retail traders get this exactly backwards.

The Full Story

The thesis in one sentence

Clenow has written the same idea three times now, each book aimed at a different audience. The sentence underneath them all is this: a trend on one market is a coin flip; the same trend on fifty markets is a business.

“To run trend following on a single market is just begging to get hit badly. Trend following is basically about taking a lot of bets on a very large number of markets independently.”

You buy what’s moving up. You short what’s moving down. About seventy percent of the time you get stopped out with a small loss. The other thirty percent pays for everything, plus a margin. Run that on one market and you can bleed to zero before the big trend ever shows up. Run it on gold and soybeans and bunds and currencies and equity indices simultaneously, and on any given week something somewhere is trending. The diversification doesn’t smooth the curve as a side effect — it is the curve.

Why the rules don’t matter as much as you think

Clenow’s first book, Following the Trend, spent eighty percent of its pages on one strategy: a moving average filter with a breakout. Simple enough that most readers assumed it couldn’t be a real edge. Then he added a reverse-engineering chapter — took that toy model, ran it against the biggest CTAs in the world, and got correlations of 0.7 to 0.9 just by nudging risk dials and the commodity-versus-financials mix.

“It’s not that complicated really. You can explain the performance using a simple model.”

His point isn’t that the famous funds are frauds. His point is that the entry and exit rules are the least interesting part of the system. Once you accept that trends are trends, the only question is how many you can stack in parallel, and how much you size each one. Retail traders argue about whether to use 50-day or 200-day moving averages. Institutions argue about correlation matrices and vol targets. Only one of those conversations is load-bearing.

What risk actually is

The retail-conference set-piece Clenow keeps performing: somebody asks him what percentage they should risk per trade. He asks back: per what? Per day? Per week? Per year?

“Risk is a matter of valuation change per unit of time. If those components are not in it, it’s not risk.”

Without a time dimension, “risk per trade” is just a number people repeat at each other. Most of what gets sold as money management — pyramiding, percent-risk-per-position, fixed-ratio sizing — is gambling math dressed up in finance vocabulary. The institutional version is volatility-targeted position sizing, which is just risk parity for systematic traders: size each market so its expected contribution to portfolio vol is roughly equal. A low-vol position has to be large to matter. A high-vol position has to be small to not blow you up.

That last point lands hard in the Swiss franc story at the end of the interview, so hold onto it.

Markets versus marketing — what changed since the 80s

The most interesting new thread in this conversation, compared to the Quantopian talk earlier today, is Clenow’s clarity that the strategy and the business are two different problems.

“The biggest thing I’ve come to realize is that it’s not about the strategy as much as it is about the business.”

When the Turtles and AHL were starting in the 80s, trend following was a small strategy with room for new entrants. Anyone with a working model and a stomach for drawdowns could raise money. By 2019, it’s commoditized. The big shops have eaten the category the way large banks eat mutual funds — distribution and brand do the lifting, not the underlying edge.

His read: if you’re running your own money, trade what works. If you’re building a business, you need to be different from the billion-dollar players, because “slightly better than them” is not a fundable pitch. He’s allocating institutional capital from Zurich, and what he buys around the trend-following core is interesting — equity-market-neutral managers for stable cornerstone returns, US alternative lending (securitized in Europe) for uncorrelated yield, and at one point a saltwater disposal business in Arizona cleaning frack water. The point isn’t the saltwater. The point is that any portfolio big enough to matter has to keep hunting for things that aren’t stocks dressed up as something else.

The portfolio role of trend, and why he doesn’t love equities

Trend following sits in his book as a core building block, not a satellite — it tends to perform best in high-vol regimes and severe bear markets, which is when the rest of a portfolio is screaming. Returns have come down over the past decade. He doesn’t think the strategy is broken; he thinks low yields compress everything.

Then Meb pulls a real opinion out of him on equities:

“Equities are still one of the worst asset classes. If you look at the volatility-adjusted returns over time on buy-and-hold equities, it’s still one of the worst-looking ones.”

Five to seven percent total return with dividends, against drawdowns of fifty to sixty percent, twice in the last two decades. The construction of cap-weighted indices makes it worse — the S&P 500 is really the S&P 20, with 480 rounding errors trailing behind. Equal weighting the same 500 names outperforms over time. The 1970s innovation wasn’t market-cap weighting per se; it was lower fees than active. Almost any other weighting scheme beats cap-weight if you back-test it honestly.

He still tells listeners to own equities. He just thinks they’re more dangerous than the index brochures suggest.

Long-flat versus long-short

Meb asks the question most retail allocators eventually arrive at: if shorting futures has net lost money over the past decade, why not just run the long side?

Clenow’s answer is the cleanest treatment of negative correlation as a portfolio input I’ve heard from him.

“A component in a product doesn’t need to have an attractive positive return. It can even have a negative return if it has sufficient negative correlation to the other part of the portfolio.”

The short book, in isolation, might be a drag. Stapled to the long book, it cuts portfolio vol enough that risk-adjusted return goes up. You’re not buying the short side for its returns. You’re buying it for its behavior in the exact moments when your long book is in trouble. There are no prizes for highest gross return. You win on risk-adjusted return, full stop.

Meb’s tell: if forced to pick one personally, he’d go long-flat. As a diversifier inside a bigger book, long-short earns its keep.

Stocks need momentum, not trend

The bridge to the second book, Stocks on the Move. People kept asking if trend-following worked on equities. His answer was: not really, but momentum does, and the difference matters.

Trend works on futures because you can trade everything from cocoa to palladium to bonds — true diversification. Run the same logic on stocks and you don’t get diversification, you get beta. Fifty stocks trending up are just one bull market with extra steps. Momentum on stocks has to be conditioned on the regime of the overall market. You can’t realistically expect to make money buying stocks in a bear market, and shorting stocks for weeks or months is a specialist’s game.

The Swiss franc story

The closing anecdote is the one to remember. The day before the SNB abandoned the EUR/CHF peg in January 2015, Clenow needed to do a large personal FX conversion. His bank quoted him a 3% spread. He spent two hours yelling up the management chain trying to get it tightened, gave up in disgust, walked away. The next morning the peg broke. The franc moved roughly 30% in his favor overnight. Best trade of his year, executed by losing his temper.

The serious lesson he extracts is more useful than the funny one. Trend-following CTAs got crushed on the franc because volatility-targeted position sizing had been telling them to load up — the artificially suppressed vol made the position look small, when the real risk was a one-sided cap waiting to break.

“Some funds lost enormous amounts because they took massive positions on the wrong side of an artificially inflated asset and didn’t take into account that the low volatility is also artificially created by central banks.”

Don’t blindly follow your rules in special situations. If a central bank is holding a line against market pressure, your vol model is lying to you about what your position size means. It’s the deepest practical point in the interview, and it’s about model-versus-reality humility, not about trend following per se.

Key Takeaways

The strategy isn’t the entry rules. It’s the diversification across markets plus the position sizing.
Risk without a time dimension isn’t risk — it’s a number. Always denominate by unit of time.
The short book of a trend system can lose money standalone and still be worth running, because of negative correlation in stress regimes.
Cap-weighted indices are barely indices. The S&P 500 is dominated by its top 20 names. Equal-weight beats cap-weight over time.
Trend works on futures because the universe is genuinely diverse. The equivalent on stocks is momentum, and it requires a regime filter on the overall market.
Vol-targeted position sizing breaks when volatility is artificially suppressed by a central bank or peg — the position looks small but the tail is enormous.
Building a trend-following business in 2019 is different from building one in 1995. The category has commoditized and brand-name distribution dominates.
Clenow’s writing audience is himself, ten to twenty years younger. That’s why the books read like a senior engineer leaving notes for an intern.

Claude’s Take

This is the more polished, more discursive Clenow. The Quantopian talk earlier today was the lecture version — diagrams, position sizing math, the chimp metaphor. This is the magazine-profile version, with a sympathetic host who already agrees with him. You learn more about the business of running money and less about the actual trading mechanics.

What’s genuinely new compared to the talk: the SNB story, the strategy-versus-business framing, the alternative-lending vehicle, and his unusually flat take on equities as an asset class. The franc anecdote is the keeper — it’s a clean illustration of model-versus-reality that goes beyond trend following.

What’s recycled: the seventy percent loss rate, the moving-average-plus-breakout simplicity, the institutional-versus-retail framing on risk. Fine to hear twice; they’re load-bearing ideas.

Where to push back. The same n=1 problem applies that I flagged on the Quantopian talk. Trend following’s reputation rests heavily on 2008 and a few other regime breaks. The interview itself admits returns have come down for a decade. Clenow’s response — “we don’t see any sign that it stops working” — is honest but unfalsifiable. If the strategy continues to underperform for another decade, the same defense will be available. That’s not a reason to dismiss it; it’s a reason to size it as one component, which is exactly what Clenow himself recommends. The marketing risk is when allocators take “core building block” too literally and stack 20-30% in trend after a hot run, then panic-sell after a drawdown. The strategy works on people who can sit through the boring middle years. Most can’t.

The publishing-economics digression is the most candid part of the conversation. He admits the books make almost no money and that the first publisher took everything. Self-publishing the second and third was a business decision masquerading as a lifestyle one. Fair enough — and refreshing to hear an author say it out loud.

Score: 7. Solid, but a partial duplicate of denser material elsewhere in his catalog. If you’ve only got time for one Clenow hour, take the Quantopian talk. If you’ve got two, this is the second one.