Dragon's Den InvestingThe framework.
Seven ideas that drive every position. They are written down so they can be checked — against the market, and against each other.
01Concentrated, high-conviction
About six stocks — "safe" compounders (Meta, Halozyme) paired with volatile high-upside names (Cloudflare, Natera, Unity, Celsius). Concentration is the point: a twentieth position is a hobby, not a thesis.
Discipline cuts both ways. The first-to-sell candidates are named explicitly, up front — Unity and Microsoft were flagged as such long before Microsoft was actually sold in January 2026. No retroactive storytelling.
02Distribution and data win
The core AI framework: moated distribution beats foundation models. Whoever already reaches billions of people — or holds the decisive dataset — captures the value, even if someone else builds the model.
That means skepticism of the hardware-permabull trade and the silicon-shortage-forever narrative, and a real bet that biotech is where AI makes actual progress: Natera was called "the largest beneficiary of AI in the market," and both biotech picks beat the genomics ETF by a wide margin in 2025.
03Geopolitics first
Five years living in China left a permanent lens: tariffs, trade truces, sanctions, and supply chains move sectors before earnings do. The "TACO trade" — Trump Always Chickens Out — shaped positioning into the midterms; the Middle East and Iran shaped the (since exited) oil thesis.
Macro is a lens, not a prediction machine — the 2025 geopolitics grades came back D+, and that grade is published on the performance page.
04Entry discipline
Two different doors, from the Drunkenmiller playbook. For fast-moving short-term tech trends (the NVDA and memory trades that got away): buy first, analyze later — speed is the edge. For medium- and long-term theses (oil, Intel): wait for maximum market disillusion, keep a watch list, and enter only at panic prices.
And the corollary both rules share: judge the thesis, not the price. "Has the thesis changed because of new, structurally changing information?" If not, the red ink is noise.
05Stay invested — ~1% cash
The portfolio used to hold 20–30% cash waiting for crashes. For over a year now it has run at about 1% — "undoubtedly left money on the table" in the old regime, but the math changed: inflation erodes idle cash, rate cuts inflate asset prices, and the dollar's decline does the rest.
The rule is clean: keep 3–6 months of expenses in cash outside the portfolio. Inside it, stay invested — "I cannot perfectly predict when markets will bottom, and if I do it's probably as much luck as it is skill."
06The Munger influence
Charlie Munger is the patron saint of this portfolio: invert, stay in your circle of competence, and let temperament do the compounding. The original oil thesis was Munger-inspired — and when the thesis broke, the position was sold and the mistake was published.
The clearest Munger rule in practice: no chart-based technical analysis. Price action enters only through sell ranges and targets (Unity $50–$55, Halozyme $100+) — decisions, not lines.
07Radical transparency
Every quarter: performance vs the S&P 500 with dollar tables and cost bases. Every year: predictions graded A through F. Every mistake: admitted in print — selling OXY at $42 before it ran 54%, selling AMD and Intel before they ran, missing gold's 62% year and the memory boom.
The benchmark philosophy is the whole game: performance is portfolio minus index. If the index would have made $100K and active management made $250K, the time yielded $150K — not $250K. Everything on this site is built to make that number checkable.