# Does Technical Analysis Work? What the Research Actually Says

Web page: https://uptickflow.com/en/learn/does-technical-analysis-work  
Updated: 2026-10-29

Does technical analysis work? What six well-known studies found, why simple chart rules fade out of sample, where TA helps, and how to test your rules first.

Technical analysis works in a limited sense: studies find that some simple trend and momentum rules had real predictive value in past data, but much of it faded out of sample, after costs or once widely known, so no chart rule is a sure thing. Its most dependable use is different: deciding where a trade is wrong, how big it should be and which way the trend runs.

This lesson is about the evidence, not the basics. It covers what six well-known studies actually found, why simple public rules stop working, whether technical analysis works on stocks, whether it is a science, where it is genuinely useful, and how to test your own rules before you trust them. Examples are hypothetical, and trading carries a real risk of loss.

## What does the research say about technical analysis?

Academic work on technical analysis is large and mixed, and each well-known study answers a narrower question than the headlines suggest. If you need the basics first, start with [technical analysis for beginners](https://uptickflow.com/en/learn/technical-analysis-for-beginners). These six come up again and again:

- Brock, Lakonishok and LeBaron (Journal of Finance, 1992) tested simple moving-average and trading-range-break rules on the Dow Jones index from 1897 to 1986 and reported strong support for them: buy signals were followed by higher returns than sell signals, which standard models of random prices could not explain.
- Jegadeesh and Titman (Journal of Finance, 1993) found that buying the stocks that had risen most and selling those that had fallen most earned significant returns over 3- to 12-month holding periods, with part of the gain fading over the following two years. This is momentum, the trend idea in its best-documented form.
- Sullivan, Timmermann and White (Journal of Finance, 1999) widened the test to 7,846 rules and corrected for data snooping. The best rule still beat the benchmark from 1897 to 1986, but over the next ten years, 1987 to 1996, it was not even statistically significant, and on S&P 500 futures, where trading costs were easy to include, no rule outperformed.
- Lo, Mamaysky and Wang (2000) turned chart patterns such as head-and-shoulders and double bottoms into algorithms and ran them on US stocks from 1962 to 1996. Several patterns carried some extra information about returns and may have some practical value; the study measured information, not trading profits after costs.
- Park and Irwin (Journal of Economic Surveys, 2007) reviewed 95 modern studies: 56 found technical strategies profitable, 20 did not and 19 were mixed. They also found that most tests had problems, such as data snooping, rules chosen after the fact and rough estimates of risk and costs, and that the profits held at least until the early 1990s.
- Moskowitz, Ooi and Pedersen (Journal of Financial Economics, 2012) found time-series momentum in every one of the 58 liquid futures they studied, across equity indexes, currencies, commodities and bonds: returns persisted for 1 to 12 months and partly reversed over longer horizons.

- **Data snooping**: The bias that appears when many rules or settings are tried on the same past data and the best one is reported as if it had been the only one tried.
- **Out-of-sample test**: A test on data that was not used to choose or tune the rule, such as the years after the original study ended.
- **Time-series momentum**: The tendency of a market that has risen (or fallen) over the past months to keep moving the same way for a while; the idea behind trend following.

## Why do simple technical rules stop working?

The pattern in that research, strong results in older data and weaker ones later, has several causes. Once a simple rule is published and cheap to run, many traders act on it and the edge is traded away; Sullivan, Timmermann and White suggested that lower costs and more liquidity may have removed short-term patterns. Costs eat small edges, above all for fast rules, which is one of the reasons [most traders lose money](https://uptickflow.com/en/learn/why-traders-lose-money). And the more rules anyone tries, the more likely the best of them looks good by chance alone.

That last point is easy to show. In a simulation we ran on 40 random-walk markets, where by construction there is no pattern to find, the best of 172 moving-average crossover settings over the first six years gained 76% there on average, but only 4.9% over the next two years, less than simply holding (13.1%). Picking the best result from many tries finds noise. [Backtest overfitting](https://uptickflow.com/en/learn/backtest-overfitting) shows how to guard against it.

## Does technical analysis work on stocks?

On stocks, the best evidence is for trend and momentum over months, not for short-term chart signals. The Jegadeesh and Titman result has been studied for decades, and it is the same idea that [momentum trading](https://uptickflow.com/en/learn/momentum-trading) and [trend following](https://uptickflow.com/en/learn/trend-following) rely on. Chart patterns carried some information in the Lo, Mamaysky and Wang tests, but information is not the same as a profit after costs.

Single stocks also gap on earnings and news, which no chart can see coming, so even a sound trend rule needs a position small enough to survive a gap through the stop. Most online debates, including the familiar Reddit threads, talk past each other: one side means “an indicator predicts the next candle”, which the evidence does not support; the other means “trend and price levels help manage risk”, which is a different and more defensible claim.

## Is technical analysis a science?

Not as a whole. It is a collection of methods, some precise enough to test and many too vague to test at all. “Buy when the 50-day average crosses above the 200-day average” can be backtested and checked out of sample; “the chart looks bullish” cannot. The scientific part of technical analysis is a habit: turn a claim into a rule, test it on data it was not built on, include costs, and drop it if it fails.

## Where is technical analysis genuinely useful? A risk map

Even without a predictive edge, a chart answers questions every trade needs: where is the idea wrong, how big can the position be, and is the trend with you or against you? Support, resistance and swing points give a stop a logical place, the stop gives the size, and the next level gives a target and a reward-to-risk ratio. That use does not depend on the next candle being predictable.

Hypothetical: XYZ bounces from support at $48.80, and a long is planned at $50.00. The stop goes at $48.50, just below the support, because a break of it means the bounce failed. With a $20,000 account and 1% risk, $200 ÷ $1.50 = 133 shares. The next resistance, $53.50, is the target: a reward of $3.50 for a risk of $1.50, about 2.3 to 1. The chart did not forecast the move; it set the terms of the trade.

## How to test a technical rule before you trust it

A short protocol, built from what the studies above got right and wrong; [how to backtest a trading strategy](https://uptickflow.com/en/learn/how-to-backtest-a-trading-strategy) walks through the mechanics:

1. Write the rule so precisely that two people would take the same trades: entry, exit, stop, size, market and timeframe.
2. Split the history. Choose and tune the rule on the first part only, then run it unchanged on the rest.
3. Count every version you tried, and judge the winner against the best of the same number of random tries.
4. Include commissions, the spread and slippage; for fast rules, check the cost per trade in R.
5. Compare the result with simply holding the market over the same period, not only with zero.
6. Forward-test 50 to 100 trades on a [paper trading](https://uptickflow.com/en/learn/paper-trading) account before real money, then measure the [trading expectancy](https://uptickflow.com/en/learn/trading-expectancy) after costs.

## Key takeaways

- The research is mixed: simple trend and momentum rules showed predictive power in older data, and time-series momentum appears across many futures markets.
- Much of the edge faded out of sample, after costs or once rules were widely known, and testing many rules on one history inflates the best result.
- Technical analysis is not a science as a whole, but its precise rules can be tested scientifically: out of sample and after costs.
- Its most dependable use is as a risk map: where the idea is wrong, how big to trade and where the next level sits.

## FAQ

**Does technical analysis really work?**

Partly. Some trend and momentum rules showed real predictive value in historical data, but many published results weakened out of sample or after costs. It works best as a framework for risk and trend, not as a forecast of the next move.

**Does technical analysis work on stocks?**

The strongest evidence on stocks is for momentum over 3 to 12 months; short-term chart signals have weaker support once costs are counted. Stocks also gap on news, so any chart rule needs a small size and a plan for gaps.

**Is technical analysis a science?**

Not as a whole: many chart readings are too vague to test. Precise rules can be tested scientifically, by writing them down, testing them on data they were not built on and including costs.

**Why do people on Reddit say technical analysis does not work?**

Because many popular rules fail when tested properly: the best of many tried rules often looks good by chance and fades on new data, and costs remove small edges. Supporters usually mean something narrower, such as following trends or placing stops at logical levels.

**Can you trade with technical analysis alone?**

Many traders do, but a chart method still needs a tested edge after costs, a fixed risk per trade and a plan for news gaps. Without those, the evidence on retail traders suggests most lose, whatever method they use.

---

Educational content, not financial advice. Trading involves risk of loss. https://uptickflow.com/llms.txt