Process
Selection processes tend to expand until they are mostly ritual. Ours is deliberately narrow. We ask four questions, we try very hard to answer them without flattering ourselves, and we decline when any one of them comes back unclear.
One: where does the return come from?
A manager should be able to explain their edge in a way that identifies who is on the other side of the trade and why that counterparty is willing to lose. Answers that survive scrutiny usually sound structural: a forced seller created by a mandate constraint, a premium paid by someone hedging a real-world risk, a market too small for large allocators to bother with, a complexity that deters generalists.
Answers that fail scrutiny sound like description rather than explanation — "we are disciplined value investors", "we trade momentum across liquid futures" — without any account of why the opportunity persists. We are not hostile to systematic or quantitative strategies; we simply want the same clarity about why the pattern has not been arbitraged away.
Two: does the record support the story?
We attribute returns before we admire them. That means decomposing performance into market exposure, factor exposure, and genuine idiosyncratic return, then asking whether the residual is large enough and stable enough to be worth paying for. A strategy that turns out to be leveraged beta with a good narrative is not uninvestable — but it should be priced as beta, not as alpha.
We pay particular attention to the worst quarters. How did the drawdown behave relative to what the manager said it would do? Did risk come down, and did it come down by decision or by stop-out? Did investors redeem, and how was that handled? A manager's conduct in their worst six months tells us more than five good years.
Three: can the business survive?
Investment skill housed in a fragile business is a wasting asset. We look at the firm's breakeven point against current fee income, the stability of the existing investor base, key-person concentration, and whether the non-investment functions are staffed by people with real authority. A firm where the founder is also the risk manager, the head of operations and the final word on valuation is carrying a risk that no investment process can offset.
Four: do the terms and the structure fit?
Liquidity terms should match the underlying assets. When a fund offers monthly liquidity on positions that would take six months to exit in stress, someone is being subsidised, and in a crisis it will not be us. We look for honest alignment between dealing terms and asset liquidity, gates and side-pocket provisions we could live with, fee structures that reward genuine outperformance, and a transparency level sufficient for us to monitor drift.
How the answers combine
We do not score managers and rank them. A high conviction answer to question one can carry a merely adequate answer to question three, provided the position is sized for that fragility and we have structural protection. But a weak answer to question one is fatal regardless of how strong the rest looks, because without a credible source of return we are simply paying fees for volatility.
The final test is a portfolio test rather than a manager test. An excellent strategy that behaves exactly like three managers we already hold adds cost and correlation, not diversification. We have passed on genuinely good funds for that reason alone, and told them so.