OU blog

Personal Blogs

A picture of Russell Larke

What Do Technical Indicators Measure? RSI, MACD, and the Limits of Computation

Visible to anyone in the world
Edited by Russell Larke, Tuesday 10 November 2026 at 14:19

What Do Technical Indicators Measure? RSI, MACD, and the Limits of Computation

TL;DR: Key Structural Takeaways
  • Indicators are Transformations, Not Predictions: RSI, MACD, and moving averages are mathematical transformations of past price data. They do not possess predictive power; they merely smooth or standardise historical prices.
  • Data Snooping and Overfitting: Academic reviews (Park & Irwin, 2007) find that many profitable indicator strategies are the result of data snooping and ex-post selection, not genuine structural edges.
  • The Efficient Market Challenge: Weak-form market efficiency (Fama, 1970) suggests that historical price data is already reflected in current prices, limiting the structural value of lagging indicators.
  • Volume and Market Microstructure Matter: Indicators that ignore volume and order flow (like RSI) miss the mechanical constraints that actually drive price movement.

Abstract

As established in my previous critiques of why chart analysis is a waste of time, the surface-level patterns of price action are consequences of deeper structural mechanics, not causes of future movement. This article extends that critique to the computational layer of technical analysis: indicators. By analysing the mathematical construction of tools like RSI, MACD, and moving averages, and reviewing the academic evidence on their profitability, we can understand their fundamental limits. They are transformations of historical data, not predictive engines. This analysis forms part of my broader research into market mechanics and participant constraints.

1. The Mathematics of Lag

Every technical indicator is a calculation applied to a time series of past prices. A moving average, for example, simply smooths price data by averaging the closing prices over a specified period. The 50-day MA and 200-day MA are common benchmarks. The Moving Average Crossover (e.g., a Golden Cross when the 50-day crosses above the 200-day, or a Death Cross when it crosses below) is a signal derived from this smoothed data.

The structural problem is that moving averages are inherently lagging. By the time a Golden Cross is confirmed, a significant portion of the price movement has already occurred. As I discussed in my analysis of valuation metrics, the indicator does not tell you what the asset is worth; it only tells you what the average price was over a past period. It is a descriptive statistic, not a predictive one.

2. Oscillators: RSI, MACD, and the Illusion of Overbought

Oscillators like the Relative Strength Index (RSI) and the Moving Average Convergence Divergence (MACD) attempt to measure momentum. The RSI compares the magnitude of recent gains to recent losses to produce a value between 0 and 100. A reading above 70 is considered "overbought," and below 30 is "oversold." The MACD calculates the difference between two moving averages to identify shifts in momentum.

From a systems perspective, these indicators are measuring the derivative of price, not the underlying constraints. An RSI reading of 80 does not mean a stock must fall; it simply means the stock has risen consistently over the lookback period. In a strong trend, RSI can remain overbought for weeks. The indicator is blind to the structural reasons for the trend—such as a short squeeze or a fundamental catalyst—which is why reading market maker inventory is more informative than reading an oscillator.

3. The Academic Verdict: Data Snooping and Overfitting

The academic literature on the profitability of technical analysis is mixed, but it consistently highlights a structural problem: most positive results are artifacts of flawed testing procedures.

In their systematic review, Park and Irwin (2007) examined 95 modern studies on technical trading strategies. They found that while many strategies appeared profitable, the vast majority suffered from data snooping (testing many rules on the same dataset until one works) and ex-post selection (choosing the rules after observing the data). When transaction costs and risk are properly accounted for, the profitability disappears.

Brock, Lakonishok, and LeBaron (1992) provided some of the most famous evidence for simple moving average rules using Dow Jones data from 1897 to 1986. However, subsequent research has shown that these results have weakened significantly in more recent periods, suggesting the "edge" was likely a reflection of changing market microstructure rather than a timeless predictive law. As I have previously argued in my critique of why chart patterns fail, the shadow on the wall is not the cause of the object casting it.

4. The Efficient Market Hypothesis and the Limits of Historical Data

The fundamental structural challenge to technical indicators is the Efficient Market Hypothesis (EMH). In his seminal 1970 review, Eugene Fama defined a market as "informationally efficient" if prices at each moment incorporate all available information.

Weak-form efficiency specifically addresses historical price data. If weak-form efficiency holds, then the past prices used to calculate RSI, MACD, and moving averages are already fully reflected in the current price. There is no structural edge to be extracted from the historical series alone. While behavioural finance has identified anomalies and limits to arbitrage that allow some predictability to persist, the EMH remains a powerful null hypothesis. Indicators that ignore volume, order flow, and market microstructure are structurally blind to the information that actually moves prices.

5. Conclusion

Technical indicators are mathematical transformations of historical data. They can smooth noise, but they cannot predict the future. By understanding their lagging nature, the statistical dangers of data snooping, and the challenge of market efficiency, practitioners can move past the illusion of the "magic indicator." The real structural information lies not in the computed oscillator, but in the mechanical constraints of liquidity, inventory, and forced behaviour.

In the next article, we will begin a new cluster by examining the fundamental layer of market analysis: valuation metrics and the structure of financial statements.

References

  • Brock, W., Lakonishok, J., & LeBaron, B. (1992). Simple technical trading rules and the stochastic properties of stock returns. The Journal of Finance, 47(5), 1731-1764.
  • Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The Journal of Finance, 25(2), 383-417.
  • Park, C. H., & Irwin, S. H. (2007). What do we know about the profitability of technical analysis?. Journal of Economic Surveys, 21(4), 786-826.

About the Author

Russell Larke is a financial market structure researcher and systems thinker, currently completing an MSc in Systems Thinking in Practice at the Open University. He is the founder of Trading Beyond Charts.

About Me | ORCID | Wikidata

Permalink
Share post