Evidence-Based Technical Analysis

1. Technical Analysis Must Evolve into an Observational Science

This book’s central contention is that TA must evolve into a rigorous observational science if it is to deliver on its claims and remain relevant.

From Art to Science. Traditional technical analysis (TA) often relies on subjective interpretations and anecdotal evidence, resembling a faith-based folk art more than a rigorous science. To truly deliver on its claims of forecasting future price movements, TA must embrace the scientific method, grounded in objective observation and statistical inference. This evolution, termed evidence-based technical analysis (EBTA), charts a course between blind faith and relentless skepticism.

The Scientific Method. The scientific method is the only rational way to extract useful knowledge from market data and determine which TA methods have predictive power. It involves formulating testable hypotheses, collecting objective data, and using statistical analysis to evaluate the evidence. While the scientific method doesn't guarantee success, it significantly increases the chances of extracting valuable insights from market behavior.

EBTA's Goal. The goal of EBTA is to create a body of knowledge about market behavior that is as reliable as possible, given the limitations of evidence gathering and the powers of inference. This involves a continual process of testing, refining, and discarding ideas that fail to hold up under scrutiny, leading to a progressively more accurate understanding of market dynamics.

2. Subjective Technical Analysis Lacks Cognitive Content

Much of popular or traditional TA stands where medicine stood before it evolved from a faith-based folk art into a practice based on science.

Meaningless Claims. Subjective TA methods, characterized by their vagueness and reliance on private interpretations, fail to meet the criteria for legitimate knowledge. Because they cannot be objectively tested or refuted, claims of their effectiveness are essentially meaningless. Examples include classical chart pattern analysis, hand-drawn trend lines, and Elliott Wave Principle.

The Programmability Criterion. The acid test for distinguishing an objective from a subjective method is the programmability criterion: A method is objective if and only if it can be implemented as a computer program that produces unambiguous market positions (long, short, or neutral). All methods that cannot be reduced to such a program are, by default, subjective.

Faith-Based Approach. Subjective TA is akin to religion, based on faith rather than evidence. While proponents may offer cherry-picked examples of success, these anecdotes cannot compensate for the lack of objective, statistical validation. Subjective TA is not even wrong. It is worse than wrong. Statements that can be qualified as wrong (untrue) at least convey cognitive content that can be tested. The propositions of subjective TA offer no such thing.

3. Erroneous Knowledge Stems from Cognitive Biases

Although the scientific method is not guaranteed to extract gold from the mountains of market data, an unscientific approach is almost certain to produce fool’s gold.

Systematic Errors. Erroneous knowledge often arises from systematic errors in how we process information, particularly in complex and uncertain situations like financial markets. These biases, unlike random errors, occur repeatedly in similar circumstances, making them predictable and potentially avoidable.

Cognitive Psychology. Cognitive psychology has identified numerous biases and illusions that distort our perceptions and learning processes. These include:

• Overconfidence bias
• Self-attribution bias
• Hindsight bias
• Confirmation bias

Maladaptation. Human intelligence, while powerful, is maladapted to making accurate judgments in uncertain environments. Our brains evolved to find patterns, but not necessarily to distinguish valid from invalid ones. This predisposes us to adopt false beliefs, especially when dealing with complex phenomena like financial markets.

4. Overconfidence and Self-Attribution Distort Reality

In general people are too confident.

The Overconfidence Bias. People tend to overestimate their abilities and knowledge, a phenomenon known as the overconfidence bias. This is especially pronounced in difficult or impossible tasks, such as predicting short-term market trends.

The Self-Attribution Bias. The self-attribution bias further distorts our perception of reality by attributing successes to our skills and failures to external factors. This self-serving interpretation of events reinforces overconfidence and hinders learning from mistakes.

The Knowledge Illusion. The knowledge illusion is a false confidence in what we know—both in terms of quantity and quality. It is based on the false premise that more information should translate into more knowledge.

5. The Hindsight Bias Creates Illusory Validity

The hindsight bias creates the illusion that the prediction of an uncertain event is easier than it really is when the event is viewed in retrospect, after its outcome is known.

Outcome Knowledge. The hindsight bias distorts our perception of past events, making them seem more predictable than they actually were. This creates a false sense of confidence in our ability to make predictions.

Ambiguity Obscured. In subjective TA, the ambiguity inherent in chart patterns and indicators is often obscured by outcome knowledge. After the fact, it's easy to selectively notice features that seem to have predicted the outcome, while downplaying contradictory signals.

Falsifiable Forecasts. To combat the hindsight bias, subjective practitioners should make falsifiable forecasts, clearly specifying the conditions under which their predictions would be considered wrong. This allows for objective evaluation and feedback, reducing the illusion of validity.

6. Narratives Overshadow Objective Facts

A conflict exists between our desire for knowledge and our desire that it be delivered in the form of a good story.

The Power of Stories. Humans are natural storytellers, and narratives have a powerful influence on our beliefs. Compelling stories, with vivid details and emotional appeal, can be more persuasive than objective facts.

Distortion of Truth. The desire for a good story can lead to distortion in secondhand accounts. Storytellers may selectively sharpen some aspects and minimize others to create a more engaging narrative, even if it compromises the truth.

Elliott Wave Principle. The enduring appeal of the Elliott Wave Principle may be attributed to its comprehensive cause-effect story, which promises to decipher the market's past and divine its future. However, its flexibility and loosely defined rules make it difficult to test objectively.

7. Confirmation Bias Reinforces Existing Beliefs

Once a belief forms, we filter information in ways that sustain it.

Selective Perception. The confirmation bias is the tendency to favor evidence that confirms our existing beliefs and dismiss evidence that contradicts them. This bias inhibits learning and reinforces erroneous knowledge.

Vague Evaluation Criteria. Vague evaluation criteria in subjective TA facilitate the confirmation bias. By selectively noticing supportive evidence and downplaying contradictory evidence, practitioners can maintain their beliefs even in the face of poor performance.

The Intelligent Believer. Ironically, more intelligent people may be more prone to the confirmation bias, as they are better able to construct rationales for their beliefs and defend them against challenges.

8. Illusory Correlations Lead to False Predictions

An illusory correlation is the false perception of a relationship between a pair of variables.

Binary Variables. Subjective TA methods can be viewed as asserting a correlation between two binary variables: the presence or absence of a pattern and the occurrence or non-occurrence of a predicted outcome.

The Upper Left Cell. People tend to focus on confirmatory instances, where the pattern occurs and the predicted outcome follows, while neglecting other possibilities. This can lead to the perception of illusory correlations, where a relationship is perceived even when none exists.

Asymmetric Binary Variables. Illusory correlations are especially likely to emerge when the variables involved are asymmetric binary variables.

9. Heuristics Can Lead to Systematic Errors

To simplify, there are basically two types of thought processes: automatic and controlled.

Mental Shortcuts. To cope with the mind's limited processing capacity, we rely on mental shortcuts called judgment heuristics. These rules of thumb are generally helpful, but they can also lead to systematic errors in judgment.

The Representativeness Heuristic. The representativeness heuristic, which involves judging the probability of an event based on its similarity to a stereotype, can lead to the illusion of trends and patterns in random data.

The Availability Heuristic. We rely on the availability heuristic to estimate the likelihood of future events. It is based on the reasonable notion that the more easily we can bring to mind a particular class of events, the more likely it is that such events will occur in the future.

10. The Scientific Method: A Rigorous Path to Knowledge

Of all the kinds of knowledge that the West has given to the world, the most valuable is the scientific method, a set of procedures for acquiring new knowledge.

Objective Reality. Science assumes the existence of an objective reality that can be understood through observation and experimentation. This contrasts with subjective approaches that rely on personal interpretations and intuition.

Explanation and Prediction. The goal of science is to discover rules that predict new observations and theories that explain previous observations. Predictive accuracy and explanatory power are key criteria for evaluating scientific knowledge.

Falsifiability. A scientific hypothesis must be falsifiable, meaning that it can be tested and potentially disproven by empirical evidence. This distinguishes science from pseudoscience, which is often characterized by untestable claims and resistance to empirical challenge.

11. Data Mining Bias Distorts Backtesting Results

Although the scientific method is not guaranteed to extract gold from the mountains of market data, an unscientific approach is almost certain to produce fool’s gold.

The Data-Mining Bias. Data mining, the process of searching for patterns in large datasets, can lead to an upward bias in the observed performance of selected rules. This bias occurs because the winning rule may have benefited from good luck during the back test, which is unlikely to repeat in the future.

Factors Influencing Bias. The magnitude of the data-mining bias is influenced by several factors, including the number of rules tested, the number of observations used to compute performance statistics, and the correlation among rule returns.

Computer-Intensive Methods. Computer-intensive statistical methods, such as White's Reality Check and the Monte Carlo permutation method, can help mitigate the data-mining bias by generating sampling distributions that account for the effects of data mining.

12. A Human-Computer Partnership is the Future of TA

I invite my colleagues to expend their energies adding to legitimate knowledge rather than defending the indefensible.

Synergy. The future of TA lies in a partnership between human experts and computers, leveraging their complementary strengths. Humans excel at proposing new ideas and formulating hypotheses, while computers excel at processing large datasets and identifying complex patterns.

The Expert's Role. The TA expert's role is to propose informative indicators and specify the problem to be solved by data-mining software. This requires domain expertise, creativity, and a deep understanding of market dynamics.

Ethical Responsibility. It is the ethical and legal responsibility of all analysts to make recommendations that have a reasonable basis and not to make unwarranted claims. Objective evidence, obtained through rigorous scientific methods, is the only reasonable basis for asserting that an analysis method has value.

Last updated: February 27, 2025