Everyone quotes the same number about search funds: 35% IRR, 4.5x MOIC. It comes from Stanford's 2024 study, and it's real. It's just not what an LP takes home. So I rebuilt it — 100,000 Monte Carlo trials on Stanford's own outcome data, from the seat of the investor who actually writes the checks. Here's the bridge from the headline to reality: 4.5x — Stanford's pooled, all-in headline 3.2x — strip out the top 5 funds (Stanford's own number) 2.75x — count the 37% of searchers who never buy a company 2.55x — diversify across 10 searchers (median, not mean) 2.14x — after the searcher's carried interest Two things every LP should sit with: → The headline is dollar-weighted and outlier-driven. A handful of grand slams carry the whole average. → Your real protection isn't picking winners — it's portfolio size. A single searcher returns less than your capital more often than not. Build a book of ten and that risk drops to roughly 4%. A 2.1x net median is still a genuinely good business. Just know which number you're actually underwriting.redacted