Mohnish Pabrai, an Indian-American investor who ran Pabrai Funds, wrote The Dhandho Investor (2007) and made a second career of openly cloning Buffett and Munger. He and Guy Spier famously bid ~$650,000 for a charity lunch with Buffett. He's a self-described "shameless cloner" of other people's analysis, so don't look to him for a new valuation technique. What he offers is a crystal-clear framework for the shape of a good bet, and the courage to size it.
What is Dhandho?
Dhandho (Gujarati, roughly "endeavors that create wealth") is Pabrai's name for the low-risk, high-return bet. His favorite parable is the Patel motel owners, immigrants who bought distressed motels with little of their own money, ran them cheaply, and captured enormous upside with tiny downside. One slogan holds the entire philosophy: "Heads, I win; tails, I don't lose much." Look for situations with capped, survivable downside and large, uncapped upside. That's favorable asymmetry, and it's Graham's margin of safety and Klarman's downside-first thinking re-expressed as a payoff shape.
His principles boil down to four. Buy simple, understandable businesses. Prefer distressed businesses in distressed industries, where fear has already capped the price and the downside. Demand a durable moat and a margin of safety on every bet. And the temperamental core is "few bets, big bets, infrequent bets." You wait, mostly doing nothing, for the rare heavily asymmetric opportunity. When it comes, you bet big, because an edge you refuse to size does you no good.
Advertisement
What does cloning mean in Dhandho investing?
Pabrai's most provocative teaching is that originality is overrated. You can build a superb record by cloning, copying the best investors' best ideas. Read their 13F filings, the quarterly US holdings disclosures. (In India, the closest thing is the exchanges' bulk/block-deal and shareholding disclosures.) Work out why they own what they own, and buy the ones that fit your circle. His case is that the value world is small, the great investors' reasoning is often visible, and pride is a needless handicap.
The slogan hides a demand, though. Clone the reasoning, and not just the ticker. If you buy a name just because a famous investor owns it, without rebuilding the thesis, you'll have no idea when the thesis has broken and no ability to hold through a drawdown. You'll sell at the bottom, because it was never your conviction.
Like the Magic Formula or Lynch's mall, cloning is a sourcing method. It feeds your own analysis and can't stand in for it.
How big should a Dhandho bet be?
"Few bets, big bets" raises a question that the decision-science and position-sizing lessons answer properly: how big? Pabrai draws on the Kelly criterion, a formula from information theory, carried to the blackjack table by Ed Thorp, that gives the bet size maximizing long-run compound growth for your edge and odds. For a simple binary bet, f* = (b·p − q) / b, where p is your probability of winning, q = 1 − p your probability of losing, and b the net odds (the win amount per unit staked). Pabrai lives by the intuition behind it. Size up when the edge is large, and refuse the bet when it isn't, which is the opposite of the diversifier's equal weighting.
The later lessons make the caveats rigorous, but you need them now. First, full Kelly is far too aggressive for real investing. It assumes you know the true odds, which you don't. It also tolerates gut-wrenching drawdowns. So practitioners use fractional Kelly, a half or a quarter of the formula's bet. Second, Kelly's whole logic presumes you avoid ruin, because the compounding it maximizes only happens if you survive every round. That's the ergodicity point.
Pabrai's "big bets" are conviction sizing inspired by Kelly. They're deliberately fractional and hedged by an obsession with capped downside. So "tails I don't lose much" is more than a slogan. It's the precondition that makes big bets survivable.
Worked example: the asymmetry, computed
Worked application (the asymmetry, computed). Pabrai hunts in both India and the US. He has long invested in Indian names (check the current holdings), and the shape he looks for never changes. Take a synthetic distressed-but-surviving situation, with the stock at ₹100, judged over ~3 years:
| Outcome | Probability | Value | Return |
|---|---|---|---|
| Recovery to fair value | 60% | ₹250 | +150% |
| Muddle through | 30% | ₹120 | +20% |
| Permanent impairment | 10% | ₹40 | −60% |
Expected value = 0.60 × 250 + 0.30 × 120 + 0.10 × 40 = 150 + 36 + 4 = ₹190, a +90% expected return. The worst realistic case is −60%, which would hurt but wouldn't wipe you out, and there's only a 10% chance of even that. That's "heads I win (a lot), tails I don't lose much." The expected value dwarfs the price, and the downside is survivable. Compare a coin flip that pays +100% or −100% at even odds. It has the same headline "upside," but zero expected value and a ruinous tail. Dhandho means pulling the trigger only on the first shape and passing on everything else, however exciting. And because the asymmetry here is large, it means sizing the bet meaningfully: fractional-Kelly big, and nowhere near index-fund small.
Worked Example 3 carries this into the sizing math and the ruin caveat.
When does Dhandho work, and when does it fail?
Dhandho works when you can actually find capped-downside, uncapped-upside situations. They usually turn up where fear has done the capping for you: in distress, cyclicality, complexity. It also needs the temperament to bet big rarely and sit still the rest of the time.
It fails in three situations. The first is when the "capped downside" was an illusion. Leverage, fraud, and terminal decline turn "tails I don't lose much" into "tails I'm wiped out," and Pabrai's own worst mistakes involved leverage and commodity or cyclical names where the floor gave way. The second is when cloning stands in for understanding, because you can't hold what you don't own intellectually. The third is when big bets meet a temperament that can't survive the drawdowns concentration guarantees.
Everything rests on the downside actually being bounded. Verify the floor before you believe the asymmetry.