Howard Marks imagines someone building a ski resort in Miami. Three months later, a freak blizzard drops twelve feet of snow, and the resort enjoys a profitable first season. Profit has rewarded a decision that made no sense when it was made. The test is what an informed person could reasonably have expected beforehand. Investment records require the same discipline. Investigate how the result arose, what risks went untested and whether the method could plausibly work again.
Drawing on Nassim Nicholas Taleb, Marks calls the other plausible outcomes alternative histories. Only one becomes visible. A profitable portfolio might have been vulnerable to several disasters that never happened; a prudent one might meet an unusual loss. Even a gain for an ordinary reason can mislead. If you buy a stock expecting a corporate development that never occurs, a rising market may still reward you. The profit has not validated the original thesis.
Selection adds another difficulty. Among a large population of investors taking chances, some will accumulate remarkable winning streaks through luck. Looking only at those survivors conceals the unsuccessful attempts that made a spectacular winner likely somewhere. This does not make all success random. It makes many observations over many years, including unfavorable conditions, necessary evidence before attributing success to repeatable ability.
Subtract the market's contribution
A low-cost index portfolio supplies a useful baseline. It obtains the selected market's exposure without requiring superior security selection. An active manager must add something to justify departing from that baseline and incurring additional costs. When evidence of an advantage is weak, indexing remains a practical alternative to paying for ordinary market participation.
Marks illustrates the distinction with beta, a portfolio's sensitivity to market movements. Suppose the market gains fifteen percent and the portfolio has a beta of one point two. Multiplying fifteen by one point two gives eighteen. An eighteen-percent return could therefore be explained entirely by greater exposure. The extra three percentage points need not show superior judgment. That sensitivity can magnify declines too. In the chapter's simplified model, alpha represents the contribution of skill beyond the return explained by market exposure. Calculating a favorable difference alone cannot establish repeatable skill.
Beta also has limits. Responsiveness to market prices does not measure every way an investor might permanently lose capital. Marks favors judgment about the risks taken. Earning the same return with less risk can demonstrate ability, just as earning more at equal risk can. And, as Seth Klarman cautions, losing less than a collapsing index may still leave the investor with a devastating absolute loss. Relative success does not pay that loss back.
Test the method under pressure
Investment style sets another expectation. An aggressive investor should gain more in rising markets; a defensive one should lose less in falling markets. Test each when conditions oppose its usual approach. After allowing for style, does the manager capture a larger share of gains than of losses? That improvement on the normal tradeoff is the pattern Marks seeks. Oaktree aims to keep pace in good markets and lose less in bad ones. Sustaining that pattern would be evidence of skill; announcing the goal proves nothing.
The method must explain the pattern. Klarman asks whether it makes sense, can be repeated and exploits an opportunity that competition has not removed. Joel Greenblatt stresses examining the actual investment process, including risk control. One random setback need not discredit sound reasoning. Repeated failures, however, require reconsideration. Crediting every gain to skill and every loss to luck prevents the record from ever correcting the investor.
Greenblatt also warns that success can attract more money than a method can use. Suitable opportunities may be plentiful for a small portfolio and insufficient for a much larger one. Managers then have to invest differently, so the old record may describe a strategy they can no longer follow. Returning excess capital can preserve the fit between the method and its opportunities. Intelligence alone cannot remove that capacity limit.
Even steadiness needs an explanation
Exceptional returns need identifiable sources: unusually cheap purchases, scarce skill, substantial risk, borrowing or luck. None is available without limits. Even skill cannot reliably force high returns from an unpromising market. Unusual consistency deserves scrutiny too.
Bernard Madoff's Ponzi scheme illustrates why. Marks describes fraudulent claims of returns around ten percent a year, with remarkably few losing months. The annual number looked familiar; the smoothness did not resemble the variable stock-market returns behind a similar historical average. Marks found neither a convincing explanation nor sufficient market capacity for the stated method. A claimed record must be achievable through its described mechanics. When someone offers allegedly assured profits, also ask why they are sharing the opportunity instead of keeping the gains.
These tests require conscientious implementation, not merely an attractive philosophy. Valuation, risk control and behavior must support one another in actual decisions. Marks accepts that other disciplined approaches can work. Learn by connecting decisions with their consequences; Paul Johnson recommends recording forecasts so unsuccessful calls remain visible. Avoid chasing recent performance, buying after gains and leaving after losses, which can undermine even a good investment. Trust grows from a method that survives examination and remains usable. No past record removes the need to keep examining it.
Three things to keep
Judge decisions using what was knowable before the outcome arrived.
Market exposure and favorable conditions can imitate investment skill.
Trust requires a repeatable process that still fits its opportunities and capital.