*Published July 23, 2026* --- Sam Zell built a multibillion-dollar fortune in real estate, and his method reduces to a rule simple enough to fit on a napkin: if the downside is small and the upside big, go; if the downside is big and the upside small, run. It gets quoted as a universal law of judgment. Read that way it is almost useless, because the cases that matter are the ones where you cannot yet tell which is which. The rule is not wrong. It is scoped, and the scope is the whole point. How much it helps you depends above all on one thing the napkin leaves out: whether you have an edge. --- Start with the case where you have one. You understand a business, you can estimate what it is worth, and you can tell when the price is well below that. Here the advice is not to scatter small bets and wait for a winner. It is the opposite: you concentrate. This is the Buffett school of value investing, and its record is among the strongest in the field. "Diversification is a protection against ignorance," Buffett says; it makes little sense for someone who actually knows what they are doing. Look at what does the work when you concentrate. It is a forecast. You estimate intrinsic value, then demand a discount to it, a margin of safety, so that even if your estimate is off you do not lose much. This dissolves the false choice the napkin seems to offer. The margin of safety is a capped downside, but you get it by forecasting, not instead of forecasting. The two are the same act. When you have an edge the cap comes for free inside the forecast, and the upside is no mystery: your view is the reason to bet. --- How do you tell which case you are in? Use the test Buffett applies to himself: can you estimate the thing's worth closely enough that a discount to your own number would protect you if you turned out wrong? If you can put a defensible figure on it, you have an edge. If you cannot, if the honest answer is that you are guessing dressed up as analysis, you have none, and wanting one does not manufacture it. Now remove the edge. You are backing a founder in an unproven market, changing careers, starting something new. You cannot predict which shot pays off, and there is no intrinsic value to discount. The strategy inverts, and Zell's rule earns its keep. The only variable you control is the downside, and you do not have to control it by estimating it. You control it by construction: a financial floor, a cut-loss rule set in advance, a stake small enough that losing it changes nothing. Once it is capped, the arithmetic changes. You no longer need to be right about any single bet. Jeff Bezos put it plainly: most business experiments fail, but the returns are long-tailed, so an occasional one can score not four runs but a thousand. A ninety percent failure rate is the correct price of admission, provided no single failure ends the game. Compounding only pays the people still playing. --- Which raises the question the rule assumes away, and it is the one that matters: how do you know your low downside is low? The rule takes it as a given input. In practice it is a claim, and often a wrong one. The bet that looks like it can only cost a little is the one that hides a tail: a correlated exposure, a leverage clause, a way to lose far more than the visible stake. So verify the cap, do not assume it. The cheapest tool is reversibility. A decision you can walk back is capped by construction, because the worst case is that you undo it and lose some time. Amazon runs on this, making its reversible calls fast and its irreversible ones slowly, because an irreversible move is one you cannot walk back and so has to be right the first time. Before you trust a low downside, run the pre-mortem: assume the bet has blown up, and trace the path. If it includes any route to a loss you could not absorb, the downside is not low, and the rule was about to lie to you. --- Two boundaries keep this from hardening into a law of its own. First, "you cannot predict the upside" is only true in a specific place. Tetlock's forecasting tournaments showed trained amateurs beating professional intelligence analysts with classified access by about thirty percent; forecasting is a real, trainable skill, and in ordinary domains it works. The claim belongs only to the true tail, the rare events even superforecasters miss. Between the forecastable business and the black swan lies most of life, where you can and should predict. Second, the many-shots half is not free. Chasing the big payoff can select for a lottery mentality that builds worse things, and a portfolio of small bets only beats concentration when chance rules the outcome. Where skill dominates, spreading thin just dilutes your edge. --- There is a personal version of this. The many-shots strategy assumes spare capacity, that shot-count is cheap. I am the primary daytime caregiver for two young kids, and I have deliberately capped how many things I will carry at once. Shot-count, for me, is the scarcest input I have. When you can take only a few shots, multiplying them is off the table; engineering the downside is not. What replaces the portfolio is sequence: a capped downside means a failed bet does not end the game, so a few shots taken one after another, each survivable, still compound into many draws over a life. You cannot widen the portfolio; you can keep taking the next shot. For a life like mine the thesis concentrates: nearly all of the value sits in capping the downside, almost none in taking more bets. That is the real content of Zell's napkin once you stop reading it as a slogan. It is not "spot the asymmetric bet." It is a claim about where your leverage lives, and the answer depends on your edge and your capacity. Where you have edge, concentrate and forecast; where you have none, stop predicting the winner and build the floor. And the tighter your capacity, the more of good judgment collapses into one discipline: making sure that whatever you get wrong, you can afford to be wrong about it. --- ## Sources - Sam Zell's asymmetric risk/reward rule, in his own words — The Tim Ferriss Show #407 (2020): https://tim.blog/2020/03/08/sam-zell-transcript/ - Warren Buffett, "diversification is a protection against ignorance": https://www.grahamvalue.com/blog/warren-buffett-and-diversification - Jeff Bezos on long-tailed returns ("you can score 1,000 runs") and one-way / two-way door decisions — Amazon 2015 Letter to Shareholders: https://s2.q4cdn.com/299287126/files/doc_financials/annual/2015-Letter-to-Shareholders.PDF - Philip Tetlock, superforecasters beating professionals — *Superforecasting: The Art and Science of Prediction* (2015); overview talk: https://longnow.org/talks/02015-tetlock/ - Portfolio-of-small-bets under non-ergodicity — Ole Peters, ergodicity economics (via Daniel Vassallo) - Lottery-mentality founders build worse companies — Packy McCormick (Not Boring)