022 - Andreas Clenow - A Most Private Discussion on Building Long Term Wealth through Trading

ELI5/TLDR

A long-running quant trader walks through how he thinks about three different audiences for the same job — running a hedge fund, running a family office, and now running a phone app for people with ten dollars to invest. The thread that ties it all together is that no single strategy is a finished product. Trend following, momentum, all the rest — they are building blocks. The hard part is assembling them into something that survives a lifetime, including the boring middle years where nothing exciting happens and you have to keep showing up.

The Full Story

The same models, three different jobs

Clenow has spent fifteen years writing the same book three different ways. Following the Trend explained Futures trend following. Stocks on the Move explained momentum in equities. Trading Evolved showed people how to build the back-test environment for themselves in Python. The common thread is what he keeps repeating in this conversation: the models in the books are educational artefacts. They are designed to teach the phenomenon, not to be deployed as-is.

“Never take anybody’s model and trade it. That doesn’t matter who built it — you don’t trade somebody else’s model without understanding what it is and what it does.”

He notes that the rules in the 2023 edition of Following the Trend are simpler than the original 2012 version. On purpose. The first book got misread by some people as a recipe; the simpler version makes it harder to mistake the textbook for the meal.

Why the big trend funds stopped being pure trend funds

The conversation gets unusually candid when the hosts ask why the famous trend-following CTAs from the 2000s mostly turned into multi-strategy shops. Clenow’s answer is not about alpha decay. It is about who writes the cheques.

If you run a billion-dollar fund and pitch to pension funds and insurers, you are no longer selling returns. You are selling volatility control. A pure trend program has long, ugly drawdowns and lumpy years. The people allocating fifty-million-dollar tickets care more about smooth lines than fat ones.

“It’s not that there’s a big problem where you can no longer make money in trend following — that’s ridiculous. The problem is you’re running a business and you need to look at more factors than just what’s the average annual return.”

The funds that stayed pure — Jerry Parker, Mulvaney — are running smaller books for clients who specifically want classic trend as one slice of their own portfolio. The big shops drifted into multi-strat because that is what the institutional buyer wanted to buy. It is a business question disguised as a strategy question.

What he actually does to measure momentum

For Stocks on the Move he wanted a momentum measure that did the same job as standard technical indicators but with cleaner mathematics. A lot of older indicators were designed for people doing calculations by hand. They simplified the math so it could be solved with pen and paper. There is no reason a modern quant should inherit that constraint.

His version: run an exponential regression on a stock’s price history. The slope tells you the annualised percentage return implied by the fit. Then multiply that slope by the R-squared — the coefficient of determination, a measure of how well the data actually sits on the regression line. A stock that drifted sideways for a month and then doubled in a week has a steep slope but a terrible fit, so the R-squared penalty drags its score back down. The ranking favours stocks that have been climbing in a tidy, persistent way.

The hosts noted it has been informally christened the “Clenow indicator”. He says this was originally a joke about how everyone in technical analysis wants to name something after themselves. The name stuck anyway.

Building blocks vs. total solutions

This is the framing Clenow returns to most often, and it is the part of the conversation that distinguishes it from the standard trend-following talk.

A hedge fund strategy is a building block. It is one slice of someone’s portfolio. The client buys it because they have decided trend exposure, or momentum exposure, or vol-selling exposure is something their wider book needs. Nobody — except possibly Jerry Parker — sincerely believes a client should put all their money into a single building block.

A total solution is the whole portfolio. It is what you actually need if you are the end investor, not the supplier. A total solution diversifies across building blocks. It includes things you do not expect to make money on most of the time, because their job is to be there when everything else cracks.

“You can enhance the portfolio greatly by including things you don’t expect to perform… Most of the time it slowly loses money over a 10-year period. The point of including a small allocation of that is that when the other things take a hit, suddenly you get a big spike in this.”

He describes this as an insurance premium. Nobody refuses house insurance on the grounds that their house has never burnt down.

Hush — a total solution for someone with ten dollars

The bulk of the second half is about his new venture, a US-only app called Hush (get-hush.app). The pitch is straightforward: institutional-grade asset management for retail savers, with no minimum balance, capped at five dollars a month in fees, completely free for small accounts.

The mechanical trick that makes it work is fractional shares. A ten-dollar account can buy 0.01 of an ETF. That means a customer with ten dollars can hold roughly eighty positions, each weighted properly, instead of being shoehorned into a single all-in-one fund. Below the surface, Hush is running the same kind of multi-asset allocation Clenow has built for institutional clients — just scaled down to the cent.

The portfolio architecture is a six-level hierarchy. At the top: stocks, bonds, alternatives. Each splits into sub-categories, which split again, until you reach the investable universe of individual ETFs. Risk — not exposure — is allocated across the tree. Weekly, the algorithm adjusts each position based on recent volatility, long-term momentum, and the live correlation matrix. The portfolio is always 100% invested. It never sells anything to zero; it just shifts weights.

The “noise” framing — which is where the show title comes from — is partly a positioning statement. Clenow is openly hostile to the retail trading culture of tick charts, news flashes, leverage, and constant updates. He sees most of it as marketing dressed up as activity. Hush is meant to be the opposite: weekly rebalances, no buttons to press, no daily P&L stress, just a habit of putting twenty dollars in every week for thirty years.

The unsexy point about regulation

A digression that is worth pulling out. Clenow describes the compliance and legal cost of running a regulated US-based venture as a significant ongoing line item. He is not complaining about the rules. He is complaining about the asymmetry: a regulated player spends real money to follow the rules while unregulated influencers post returns and tips on the internet with no consequences. Regulation creates a moat for incumbents and a barrier for innovators, and is patchily enforced against the people most likely to mislead retail.

The Australian hosts add that their regulator has the rules but not the resources to enforce them. The pattern seems to be widespread.

Why he thinks Python is now table stakes

Asked about his fourth book — Trading Evolved, the Python primer — Clenow makes a strong claim. If you want to do systematic trading at any level, you need to learn to program. Not as an expert. Just enough to verify what your tools are doing.

“Imagine a doctor or a lawyer like 50 years ago — you think they could type on a typewriter? No, that’s a low-level job. Nowadays imagine a doctor or a lawyer who can type his own things on a keyboard like one finger at a time. Come on. He’s a helpless baby.”

He uses zipline-reloaded, the open-source fork of the back-tester originally built by the now-defunct Quantopian. The thing he singles out as unusual is how it handles Futures: instead of feeding it a continuous contract (the standard simplification), you feed it every individual contract with its real start and end dates. The engine constructs the continuation it needs for ranking, but trades the actual contracts. This lets you back-test curve strategies, term-structure trades, and roll logic the same way you would actually trade them.

The novel

He has also written a noir-ish novel called A Most Private Bank. Set over seven days in Zurich, fictionalised but with more truth in it than he is willing to specify. His pitch: you will learn more about how high finance actually works from this book than from another textbook on Futures back-testing.

Key Takeaways

The textbook models in trading books are educational, not deployable. Their job is to teach a phenomenon, not to be a recipe.
The big trend funds drifted to multi-strat for business reasons, not because trend stopped working. Pension allocators buy volatility control, not returns.
Single strategies are building blocks. Diversification across building blocks is the total solution. Most retail investors confuse the two.
A good total solution includes positions you do not expect to make money on. They are insurance premiums; they pay off when the rest of the book cracks.
Clenow’s momentum measure: exponential regression slope multiplied by R-squared. The R-squared penalises ragged trends and rewards smooth ones.
For a long-horizon equities portfolio, 25–30 stocks is enough to get most of the diversification benefit. Going from 30 to 50 adds very little.
Position sizing has to be re-checked constantly. If you hold a 5% position for two years without rebalancing, you no longer have a 5% position. Risk drifts.
Rebalancing frequency is governed by trading cost, tax regime, and account type. Switzerland’s lack of capital gains tax simplifies the problem.
Fractional shares enable institutional-grade portfolio construction for very small accounts. This is the structural change that makes apps like Hush possible.
Levered and inverse ETFs are dangerous unless you understand option Greeks. The retail-facing ones are often expensive structured products in disguise.
Never trust an ETF’s name. Read the rules.
Regulation in finance creates a moat for incumbents. The unregulated competition — influencers, anonymous tipsters — frequently makes louder promises and pays no cost for them.
Learning to program is now baseline literacy for systematic trading. Open-source tools like zipline-reloaded are good enough to start.
Zipline-reloaded handles Futures by individual contract, not by continuation. This unlocks curve and term-structure strategies that other back-testers cannot represent honestly.

Claude’s Take

This is Clenow in late-career reflective mode. There is less new technical content than in his books, but the framing — building blocks versus total solutions, who pays for volatility control, what business considerations actually drive strategy choice — is the kind of perspective that only shows up after twenty years in the chair. The Hush pitch sits well inside that frame: a deliberately boring product for an audience that the rest of the industry either ignores or actively preys on.

The score reflects what the conversation is and is not. It is not a deep technical session — there is no new model, no fresh research, no real benchmark numbers. But the texture of how a senior quant thinks about business design, audience segmentation, and the gap between a hedge-fund strategy and a household portfolio is genuinely useful, and rarely articulated this clearly. The Hush concept is also one of the more honest retail-finance products in recent memory: fixed cap on fees, real diversification, no gamification. Whether it survives the customer-acquisition economics is a different question. The strategy is sound; the unit economics of selling five-dollar-a-month subscriptions to ten-year-olds are not obviously favourable.

The weakest section is the one on regulation. Clenow’s complaint is legitimate but does not land hard because the conversation does not push on it. The strongest is the candid section on why trend funds turned into multi-strat shops — that is the kind of insider observation that almost never makes it into the textbooks.