III · Approach

The best ideas from fifty-eight thousand. Not the best of forty.

Return comes from depth — from actually knowing what a business is worth. Low volatility comes from breadth — from holding enough unrelated positions that no single one decides the year. A fund has always had to pick one, because deep research didn't scale. We don't have to. We underwrite the whole market, keep only the businesses trading far below their worth, and hold them by the hundred.

≈58,000 — the listed universe tens of thousands — full work-ups in flight hundreds — priced far below their worth one book — sized, and held for years

fig. 1 Almost everything that enters the funnel fails the work-up. What survives is the portfolio.

Mandate
We identify mispriced opportunities where standard models fail.
Universe
Every listed company, every exchange, every jurisdiction. Smallness, foreignness, and neglect aren't risks to screen out; they're where mispricing tends to live.
Selection
Every candidate gets the full work-up, and almost none survive it. What's left is a small number of businesses priced far below their worth — cheap enough that we can be wrong about one and still come out ahead.
Portfolio
A few hundred positions, cheap for unrelated reasons, in unrelated places. Held together, the individual swings mostly offset while the discounts remain — which is how markets too small or too far away to be worth a desk's time combine into a portfolio that holds still.†
Horizon
Years. We sell when a thesis is realized or proven wrong, not on a schedule.

fig. 2 A single position swings. Hundreds, cheap for unrelated reasons, mostly cancel one another out — leaving the return, with much less of the movement.

† Diversification here isn't diluted conviction. It's the same conviction, taken many times.

A few hundred cheap, unrelated positions, held at once — the returns of deep value, with much less of its volatility.