Collected Heuristics and Rules of Thumb

Actionable rules distilled from More Than You Know by Michael J. Mauboussin. Each rule states the heuristic, a brief "because" rationale, and a chapter citation.

Decision Process

Judge decisions by process, not outcome -- because in probabilistic domains, any single result is dominated by noise; only process quality is within your control (Ch. 1).
Use the process-vs-outcome matrix before updating beliefs -- because good process can produce bad outcomes (bad break) and bad process can produce good outcomes (dumb luck); a single data point cannot distinguish skill from luck (Ch. 1).
Never abandon a good process after a bad outcome -- because the failure mode is asymmetric: abandoning good process after losses is common and costly, while maintaining bad process after wins is equally common but invisible until catastrophe (Ch. 1).
Target expected value, not win rate -- because the magnitude of correctness matters more than its frequency; a portfolio with a 25% hit rate can still generate strong total returns (Ch. 3).
When the most probable outcome and the best bet diverge, follow expected value -- because asymmetric payoffs mean you can be bearish on probability yet correctly positioned on magnitude (Ch. 3).
Pre-commit to decision rules before stress arrives -- because structural behavioral constraints (Ulysses strategies) beat willpower under emotional pressure (Ch. 10-11).
Categorize opportunities by circumstance, not static attributes -- because "low P/E" is not a strategy; the situation determines whether a metric signals value or a trap (Ch. 4).
Treat investment philosophy as a diet, not a trade -- because it only works if sensible over the long haul and you stick with it; temperament beats raw intelligence as the binding constraint (Ch. 1).

Portfolio Construction

Extend your evaluation period to at least one year -- because loss aversion (~2.5x pain-to-gain ratio) combined with frequent evaluation produces negative experienced utility at any horizon shorter than twelve months (Ch. 8).
Check portfolio prices less frequently -- because hourly returns are positive only ~50.4% of the time vs. ~72.6% annually; less frequent observation mechanically increases perceived reward and reduces panic selling (Ch. 8).
Favor low-turnover strategies -- because low-turnover funds outperform high-turnover funds across 3-, 5-, 10-, and 15-year horizons; only one-third of the drag is transaction costs, the rest is behavioral (Ch. 8).
Frame bets as part of a portfolio, not in isolation -- because narrow framing (evaluating each position alone) triggers loss aversion on every line item; broad framing lets diversification absorb individual losses (Ch. 8).
Stay invested through volatility clusters rather than timing exits -- because the 50 best and 50 worst trading days bunch together; you cannot avoid bad days without missing adjacent good days (Ch. 5).
Never trust a risk model that assumes normal distributions -- because actual market returns have fatter tails and higher peaks; events beyond three standard deviations occur far more often than Gaussian models predict (Ch. 5, 31).
Distinguish experience from exposure -- because standard risk measures describe what has happened, not what could happen; leveraged portfolios are acutely vulnerable to this confusion (Ch. 31).

Valuation

Assess economic returns before growth -- because growth only creates value when incremental capital earns above its cost; companies can and do grow their way to bankruptcy (Ch. 25).
Apply the three sequential questions: (1) returns above cost of capital? (2) how long can excess returns persist? (3) if below, what is the turnaround probability? -- because duration of excess returns is the primary driver of valuation multiples, not the current level of returns (Ch. 25).
Anchor turnaround expectations to base rates: only ~29% of downturns produce sustained recovery -- because this rate is remarkably stable across tech and retail over decades; the classic value trap is buying a cheap company that deserves to be cheap (Ch. 25).
Respect mean reversion in corporate returns -- because competition and capital flows drive above-average returns back toward cost of capital; the gap between top and bottom quartile CFROI collapses from 3,000 bp to 300 bp within ten years (Ch. 25).
Watch for survivorship bias in bottom-quartile "improvement" -- because 40% of bottom-quartile companies exit within five years; apparent recovery often reflects removal of the worst, not genuine operational turnaround (Ch. 25).
Only 11-14% of companies sustain above-cost-of-capital returns for their full history -- because competitive entry erodes most moats; paying a persistence premium requires evidence of increasing returns to scale (Ch. 25).
Discount growth projections against base rates -- because average projected three-year earnings growth is ~13.4% while average realized ten-year sales CAGR is ~6.2%; consensus systematically excludes negative outcomes (Ch. 27).
Feel the exponential: a jump from 15% to 20% CAGR over 20 years more than doubles terminal value -- because compounding intuition failure causes investors to treat small growth-rate differences as trivial when they are decisive (Ch. 27).
Large-company growth bets are bets against statistical structure -- because growth-rate variance compresses with scale (power-law/Zipf pattern); extreme growth from a large base is improbable in either direction (Ch. 27).
Only 13% of large companies ($500M+) met all three value-creation hurdles in the 1990s -- because even in an unusually favorable decade, sustained profitable growth is rare; two-thirds of companies with double-digit plans failed to deliver (Ch. 27).

Behavioral Defense

Loss aversion is fixed; evaluation frequency is a policy choice -- because you cannot reduce the 2.5x pain multiplier, but you can choose how often you look, and that choice determines whether equities feel rewarding or punishing (Ch. 8).
Beware frequency-based evaluation of managers -- because a manager with the worst stock-picking percentage can still be the best total performer; frequency metrics systematically eliminate magnitude-driven outperformers (Ch. 3).
Recognize that high-frequency strategies feel good but underperform -- because loss aversion drives preference for many small wins, which are not necessarily high-expected-value strategies; this is hardwiring, not a knowledge gap (Ch. 3).
Resist the urge to find proportional causes for large market moves -- because in self-organized critical systems, small inputs can produce large outputs; the trigger is incidental, the fragility is structural (Ch. 31).
Distinguish risk from uncertainty -- because risk has a known distribution while uncertainty has an unknown one; applying frequency tools (VaR, Black-Scholes) to genuine uncertainty systematically underestimates tail events (Ch. 5).
Treat most market commentary as degrees-of-belief probability, the weakest form -- because it presents itself with frequency-level confidence while lacking valid reference classes or known distributions (Ch. 5).
Do not trust expertise for forecasting in complex adaptive systems -- because expertise excels in linear, rule-based domains but markets lack stable cause-effect relationships; research shows little advantage to domain expertise in forecasting accuracy (Ch. 6).
Watch for lollapalooza effects when multiple biases align -- because the compounding potency of several biases acting simultaneously far exceeds the sum of their individual effects (Ch. 11).
Valuation depends on your time horizon -- because two rational investors can disagree on the value of the same asset purely due to differing evaluation periods; long-term investors are willing to pay more for identical risky assets (Ch. 8).

Competitive Analysis

Monitor investor diversity, not individual rationality -- because market efficiency depends on heterogeneity of decision rules, not quality of individual judgment; correlated errors, not errors per se, cause mispricings (Ch. 14).
Track three forces that erode diversity: asymmetric information copying, agency-cost herding, and conformity preference -- because each is individually rational, making the drift toward homogeneity invisible until the tipping point (Ch. 13-14).
When diversity collapses, the system is fragile regardless of apparent stability -- because self-organized criticality means the crash trigger is incidental; the true cause is internal loss of heterogeneity (Ch. 14, 31).
Agency-cost imitation is self-reinforcing -- because the more managers herd, the more deviation becomes career-threatening, narrowing diversity further and creating the mispricing opportunities contrarians exploit (Ch. 13).
Herding has measurable directional price impact -- because herded-into stocks outperform herded-out-of stocks by 4% over six months, especially in small-cap and growth segments; sequential imitation in analyst recommendations is empirically documented (Ch. 14).
A market requires time-horizon diversity to function -- because if all investors adopted long horizons, the risk premium they exploit would vanish; efficiency depends on participants across all evaluation frequencies (Ch. 8).
Use behavioral finance to monitor diversity, not to catalogue individual biases -- because the practical edge lies in detecting collective homogeneity, not in assuming mispricings exist everywhere (Ch. 14).
Recovery from herding is slow and individual -- because crowds go mad in herds but recover their senses one by one; the asymmetry between formation and dissolution of consensus is structural (Ch. 13).

Risk Management

Prepare for fat tails explicitly -- because extreme events are far more frequent than normal distributions predict; October 19, 1987 would remain "unlikely" even repeating the universe's lifetime one billion times under Gaussian assumptions (Ch. 31).
Standard risk-reward models (CAPM) assume linearity that does not exist -- because nonlinearity is endogenous to markets as complex adaptive systems; if you regularly observe five-sigma events, your sigma is wrong (Ch. 31).
Measure expectations embedded in current price, then contemplate value ranges with explicit tail-event probabilities -- because heuristic scenario analysis outperforms formulaic precision built on false distributional assumptions (Ch. 31).
Markets are prediction-dependent, not prediction-independent -- because correct insight leads to buying, which raises price, which compresses prospective return; expected value is endogenous to the crowd's beliefs, unlike roulette (Ch. 5).
Do not confuse the trigger with the cause -- because in critical systems, searching for proportional explanations of large moves is structurally misguided; fragility builds internally through diversity loss (Ch. 31).
Size positions for the world where your model is wrong -- because the gap between experience and exposure is where spectacular failures concentrate, particularly in leveraged strategies (Ch. 31).
Removing the 50 best days (1978-2007) drops annualized S&P returns from 9.5% to 0.6% -- because return-generating events are temporally clustered and inseparable from drawdowns; the mechanical structure of returns makes timing a losing strategy (Ch. 5).
If you regularly observe five-sigma events, your sigma is wrong -- because the frequency of "impossible" outcomes is itself evidence that the distributional assumption is false; update the model, not the surprise threshold (Ch. 31).