Diversification Through Outlier Capture
Hold 5-10 names so winners can be sold and recycled while the rest mature.
The 'I only invest in the one company I believe in most' meme is mathematically inferior to a 5-10 stock portfolio even when your conviction pick wins. The reason is timing: the stock market may not agree with you for a decade, and during that decade your capital compounds nothing. With multiple holdings, the probability that at least one peaks early enough to redeploy is dramatically higher.
Sasha treats portfolio construction like YouTube content strategy — you can't reliably predict which video or stock breaks out, but having multiple shots and aggressively redistributing the winners into laggards generates outlier upside that single-name conviction cannot match. The redistribution is the engine: selling at fair value and recycling capital into discounted names compounds returns far beyond simple holding.
A tight portfolio (5-10 positions) keeps the analyst load tractable while still spreading the timing risk across enough independent paths to capture outliers.
- You cannot predict which of your companies will peak first or when.
- Concentration in your highest-conviction name forfeits the option value of multiple paths.
- Selling winners and redistributing into laggards is the compounding engine.
- Time is the binding constraint — too many holdings means none get analyzed properly.
- Diversification is not a hedge against being wrong; it is a way to capture timing outliers.
- Cap the portfolio at what you can modelSet a hard cap (Sasha runs ~7 major positions plus smaller ones, max 10). Beyond that, you can't read every filing, every competitor, and every earnings call without skimping.WarningMore than 10 names and you become the underperforming fund manager who 'has analysts' but doesn't actually do the homework.
- Pick names with independent driversChoose companies with distinct catalysts, sectors, or geographies so they don't all peak or trough together. The point is timing diversification, not just sector exposure.
- Size positions by conviction and modeled upsideLarger positions for higher modeled upside and tighter ranges; smaller speculative positions for high-variance bets like Sasha's early Lucid stake.
- Sell when any one position hits targetDon't wait for the whole portfolio to ripen. Whichever name reaches fair value first triggers a partial or full sale.
- Redistribute proceeds into the largest current discountsRecycle capital into the names with the biggest gap between market price and modeled fair value. This compounds the whole portfolio.Pro tipKeep a ranked watchlist of upside-by-name so redeployment is fast when a sale happens.
When markets ran 'absolutely bananas' in 2021, multiple Sasha holdings hit target simultaneously. He sold heavily across the portfolio rather than holding for further upside.
Sasha frames the same logic via content: making one 'best video ever' is dominated by making 100 videos where five randomly drive 50% of channel growth. You can't predict the breakouts, but you can ensure you have enough shots.
Sasha noticed that even great companies can trade sideways for years while bad companies rip — sentiment and timing are not under your control. By running 5-10 modeled positions, he turned timing luck into a structural advantage: the redistribution effect from whichever name peaks first is what compounds the portfolio.