Rules-Based Investment Discipline
Replace emotional investment decisions with consistent numerical rules
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 96%
Rules-Based Investment Discipline converts an investment thesis into measurable criteria that can be tested and applied consistently. The investor begins with relevant historical and fundamental data, identifies factors associated with future returns, and defines objective rules for buying and selling. Those rules remain in force when markets become stressful, preventing a persuasive executive, recent price movement, or emotional attachment from overriding the process. The mechanism is behavioral as much as statistical: human judgment is retained when designing and reviewing the model, but the model governs recurring portfolio decisions. Performance must be evaluated across varied environments rather than isolated successes. Rules should change only when substantial evidence reveals a weakness, not because an uncomfortable position or temporary drawdown creates pressure to intervene.
Origin
O'Shaughnessy traced the approach to teenage research comparing Dow companies by valuation. He later expanded the work with Compustat data, leading to What Works on Wall Street and systematic portfolios.
Core principles
- 01Use measurable evidence instead of persuasive narratives
- 02Define buying and selling rules before stress arrives
- 03Apply the same criteria consistently
- 04Treat emotional judgment as a source of investment error
- 05Study market history to test proposed rules
How to run it
- 1
Frame a measurable question
Translate the investment idea into a relationship that can be observed, such as whether lower valuations predict stronger subsequent returns.
Pro tip Start with a small, comprehensible universe before expanding the test.
Watch out Do not begin with a favored company and reverse-engineer rules to justify owning it.
- 2
Assemble historical evidence
Collect consistent data across companies, periods, and market environments. Include failures and difficult markets rather than studying only surviving winners.
Pro tip Clean definitions and point-in-time data before drawing conclusions.
Watch out Biased or revised data can make a weak rule look predictive.
- 3
Test candidate factors
Compare future outcomes across different levels of valuation, quality, momentum, or other measurable variables. Look for durable patterns rather than one striking episode.
Pro tip Test combinations only after understanding each component independently.
Watch out Complexity can disguise overfitting.
- 4
Write explicit portfolio rules
Define what qualifies for purchase, what triggers exclusion or sale, and how positions are sized and rebalanced.
Pro tip Make each rule clear enough that another person could reproduce the decision.
Watch out Vague exceptions reintroduce the emotions the system is meant to control.
- 5
Execute consistently
Apply the established rules during both calm and stressful markets. Separate discomfort from evidence that the process is broken.
Pro tip Automate recurring screens and rebalancing checks where practical.
Watch out Abandoning a tested strategy during its expected weak period can lock in underperformance.
- 6
Review with evidence
Evaluate whether results match the framework's expected behavior over an appropriate horizon. Amend the model only when new data demonstrates a persistent flaw.
Pro tip Document every proposed change and the evidence supporting it.
Watch out Do not optimize the model around the most recent market regime.
In the wild
O'Shaughnessy described declining to own Enron because its numbers were poor. A fundamental manager instead trusted personal assurances from company executives and bought more shares. The quantitative screen prevented personal rapport and default trust from overriding deteriorating evidence.
→ The systematic process avoided an emotionally reinforced investment in a collapsing company.
O'Shaughnessy's team identified the largest differences between favored managers' factor exposures and the broader market, then converted those differences into portfolio screens. The resulting rules preserved the managers' repeatable characteristics while avoiding some of their emotionally driven selling decisions.
→ In the reported comparisons, the systematic clones outperformed the managers they modeled.
Common mistakes
Trusting the story over the numbers
A confident executive or compelling narrative can activate the human default to trust even when measurable evidence is deteriorating.
Overriding rules under stress
Changing a decision because of panic, attachment, or recent losses defeats the behavioral advantage of a systematic process.
Treating a backtest as certainty
Historical relationships are evidence, not guarantees, and must be tested for robustness across periods and assumptions.
Is it for you?
Best for
It is best for investors and portfolio managers who can access reliable historical and fundamental data.
Not ideal for
It is not ideal for situations where outcomes depend primarily on unique qualitative information that cannot be measured reliably.
From the transcript
“models don't get stressed, you know,”
“They just keep applying those criteria. You're in, you're out.”
“Very very high valuations tended to crash and burn, Very very cheap valuations tended to do well.”
From the episode
A Timeless Collection of Infinite Wisdom with Jim O'Shaughnessy