Method &
Investment Principles

01 · The Principles We Want To Hold

We are a single-family office managing only our family's own capital. For us, the credibility of an investment method does not come from complex language. It comes from a way of thinking that can be tested, executed, and corrected when it is wrong.

There is no dazzling technique here. What matters to us is a small set of plain but difficult disciplines: ask where returns come from before judging the asset; define the nature of the risk before accepting it; write down the counter-case and the falsification conditions before settling on a view; stay precise where precision is real; stay honest where the future cannot be known. Then use time, diversification, and discipline to keep family capital from falling behind the real growth of the world's wealth.

02 · Start From First Principles

When any asset class enters our field of view, the first questions are always the same: where do its returns ultimately come from? What are we really betting on? Who bears the risk, who creates the value, and why can investors reasonably share in it? If those questions cannot be answered, historical returns, market consensus, and attractive narratives are not enough reason to commit capital.

Equity: betting on progress, and on companies remaining the vehicle for it

Take equity. We treat public equities and private equity as the same broad asset: ownership of companies. A company is not a mysterious financial instrument. It is an institutional structure through which modern society organizes capital, talent, technology, and risk. Long-run equity returns come from society genuinely progressing, and from society continuing to use investable companies as the main vehicle for that progress. Owning equity is therefore not merely a bet that a market index rises over time. It is a bet that human society keeps moving forward, and that investors can still share part of that progress through the company structure.

Property and bonds: do not mistake past experience for permanent law

Property is another example. It created wealth for many families over the past few decades, so it is easy to treat "property appreciates" as a law. But at the root, property prices are still constrained by supply, demand, income, financing conditions, and demographics. When prices are already unaffordable for more and more buyers, and population growth is slowing, past appreciation should not be mechanically projected into the future.

We are even more cautious with sovereign bonds. Many call bonds issued in a country's own currency "risk-free" because the state can issue money and need not default in nominal terms. But that only means the risk may not appear as default; it does not mean the risk is absent. Our current thinking is that the risk shifts from "will we be paid back?" to "will what we receive still be worth anything?" Inflation, currency value, real rates, and future fiscal capacity all affect the real value of the bond. Government borrowing is, in essence, spending future resources today; buying long-dated sovereign debt is a view on the future path of real rates and inflation. We are still working through this.

For commodities, property, long-dated sovereign bonds, and other areas where we have not reasoned deeply enough, we would rather mark them as "still building our understanding." Admitting boundaries is not a weakness. Pretending there are no boundaries is dangerous.

03 · How Research Becomes A Judgement

Research is not the collection of opinions, and it is not a pile of data arranged into a conclusion. Its purpose is to turn a view into a reasoning chain that can later be reviewed.

For every important judgement, we want to write down the core question, the evidence supporting it, the reasoning chain, the strongest opposing view, and the conditions that would prove us wrong. The last part matters most. A view without falsification conditions easily becomes belief. In investing, the danger is often not being wrong at the start; it is finding reasons to stay wrong afterwards.

One example we have worked through carefully is an AI compute leader. The real question was not "is this a good company?" but "can demand for AI compute continue compounding?" The strongest counter-case was not "will the share price fall?" but "if inference demand becomes commoditised, or migrates to specialised chips, does the premium at this layer disappear?" The falsification conditions therefore sit around the demand curve, competitive structure, and value capture, not short-term share price movement.

When credible sources disagree, we do not rush to pick a side. Research reports, management teams, media, experts, and historical data are inputs, not answers. If the disagreement is about interpretation, we go back to the underlying mechanism and ask which explanation fits the facts better. If the data itself conflicts, we seek more sources and compare definitions and weight. If it still cannot be reconciled, we leave that data aside. Forcing data together so a conclusion looks complete is a subtle form of self-deception.

04 · Avoiding False Precision

We are cautious around things that look very precise. Many valuation models produce target prices with one or two decimal places, but those numbers often rest on a chain of assumptions: growth, margins, discount rates, terminal value, exit multiples. If one important assumption moves, the final precision disappears.

This does not mean we reject models. The value of a model is to expose assumptions, compare scenarios, and understand the relationship between variables. It should not create the illusion that the future has been calculated. We would rather be roughly right than precisely wrong.

Where precision is real, we insist on it: fees, taxes, liquidity limits, legal terms, redemption mechanics, position limits, and execution rules. These can be verified and they materially affect outcomes. For future returns, macro variables, and market timing, we have to acknowledge uncertainty. Professional discipline here is not pretending to know a confident number; it is knowing what can be calculated, and what must be handled through ranges, scenarios, and margin of safety.

05 · Risk And Benchmark

We do not define risk as volatility. Volatility is uncomfortable, but it does not necessarily cause permanent harm. The real danger is refusing to admit that the logic has broken, or preserving nominal capital while quietly falling behind the growth of the world's wealth.

To distinguish temporary drawdown from permanent loss, we do not ask only how far the price has fallen. We ask whether the reason for buying still holds. If the logic is intact, a large fall can be endured. If the logic is broken, we should adjust whether the position is up or down. The rule has to be written before purchase; when the market is actually falling, emotion is not a reliable decision-maker.

Our benchmark is not an attractive absolute return number. For family capital, the more real question is whether the family's relative position in the distribution of social wealth can hold up over time, rather than quietly slipping while nominal numbers rise. That means returns after taxes, costs, and necessary spending should try to keep pace with society's overall wealth growth. Because family wealth is held across equities, property, businesses, and other assets, we look at the long-run growth of major asset categories, build a rough range, and try to keep the portfolio near the upper edge of that range.

This explains our attitude to returns. We do not seek spectacular results, and we do not define success as beating everyone. We want to come in a little above the market, steadily enough to cover taxes and friction costs, and to avoid falling behind the world.

06 · Facing Uncertainty: Time And Diversification

We have no confidence in the short term, and we do not pretend to. Short-term markets are too affected by emotion, liquidity, narrative, and positioning. Few people can forecast them consistently. Our conviction belongs in only two places: whether an asset class's long-term logic holds, and whether certain global trends are large, slow, and hard enough to reverse.

When the long-term logic is clear, we do not try to call the exact bottom. We enter in tranches so that timing risk is spread across time. But tranche investing is not vague delay. Once the rule is set, it should be executed like a machine: the date, condition, and proportion are written down in advance, and not changed by mood in the moment. This does not guarantee a low entry price. It reduces the most expensive errors: abandoning a sound long-term view in fear, or rewriting discipline in excitement.

We diversify for two reasons. First, we do not know where the true future winners will be; too much concentration in a single name can cause us to miss the world. Second, our goal is not extreme return, but to stay a little above the market and not fall behind social wealth growth. But diversification has a limit. The more deeply we understand something, the more we can concentrate. Buying a long list of things we cannot explain is not risk control; it is handing judgement over to the appearance of variety. If a portfolio shows no view at all, it is closer to closet indexing than active judgement.

07 · Where A Small Edge Comes From

We do not attribute that little bit to intelligence, privileged information, or mysterious access. A more honest explanation is that it comes from bearing discomfort that others are unwilling to bear for long.

The first source is sustained exposure to slow variables. Ageing populations, a more multipolar world, AI, clean energy, and industrial restructuring are not hidden. Many investors can see them. The difficulty is that they move slowly, and short-term market fashions often drown them out. We are willing to tilt the portfolio toward these directions gradually and deliberately when they are not fashionable, rather than chasing them only after everyone is talking about them.

The second source is understanding the role of both public and private markets. We do not believe private markets are inherently superior, nor that illiquidity automatically means higher return. But a complete long-term risk exposure often requires seeing both public and non-public markets. Future growth may occur in listed companies, before listing, in acquisitions, or in earlier business structures. Family capital's time horizon allows us to bear some illiquidity and long lock-ups, but only when price, terms, and underlying logic all make sense.

The third source is behavioural discipline. Many investment mistakes happen not because people do not know the principle, but because the principle is not executed under pressure. We want the research phase to be thorough: write the counter-case, falsification conditions, and execution rules in advance. In execution, reduce improvisation. Do not let fear, excitement, regret, or crowd pressure rewrite the rule. This will not make results look good every day, but it may be the most stable small edge over time.

08 · Our Discipline

Knowing what we will not do matters as much as knowing what we will. Clear boundaries make it harder to be carried away when markets are noisy.

We will be wrong, so the method needs a path for admitting it. Whether a view is truly broken or merely down is not decided by the current gain or loss. It is decided by whether the underlying logic still holds. If the logic remains, we endure volatility. If the logic is gone, we acknowledge the mistake and adjust. Whether we are up or down at the time, stubbornness and emotion cannot replace judgement.