Does Technical Analysis Work? What the Evidence Actually Shows

The honest answer: technical analysis works sometimes, but not how you think

Charts genuinely help some traders locate price patterns, support levels (a price where buyers have historically stepped in) and resistance levels (a price where sellers have historically capped rallies) that shape short-term trades. Most academic work, though, finds that the same tools fail to consistently forecast where prices go next, and they tend to underperform a simple buy-and-hold return for retail traders once spreads and commissions are subtracted.

That gap between what technical analysis promises in YouTube tutorials and what it delivers in audited trading accounts is where you lose money. The tool is real, the patterns exist, and skilled short-term traders extract genuine value from them. The complication is that the pattern-spotting brain that lets you see a head-and-shoulders formation is the same brain that sees faces in clouds, and without a way to tell those apart you end up trading noise and paying spreads to do it.

Where you trade shapes your results almost as much as how you trade; see the best forex brokers and their conditions.

What technical analysis actually claims to do

Technical analysis assumes that historical price movements, chart patterns and trading volume (the number of contracts or shares traded in a period) contain useful information about where price is heading next. Traders who use it look at charts to find entry and exit points, and they ignore company fundamentals, macro data and news except where price already reflects them.

The underlying claim is stronger than it might first sound. It says price alone contains enough information, because price already reflects every buyer's and seller's opinion, budget and fear. From that premise flow the familiar tools you have almost certainly come across:

• Moving averages, meaning an average of recent closing prices.
• The relative strength index or RSI, a momentum gauge from 0 to 100.
• MACD, which compares two moving averages to flag shifts in momentum.
• Fibonacci retracements, used to project possible pullback levels.
• Candlestick patterns, which read short sequences of open, high, low and close.
• Trendlines drawn to link successive highs or lows.

Each tool is essentially a rule for turning past prices into a signal, and whether those signals carry information or simply repackage the same random walk into a shape your brain recognises is the empirical question.

Why the research is mixed and what it actually proves

Academic studies on technical analysis produce genuinely conflicting results, and that inconsistency itself is the finding. Some papers document statistically significant returns from moving-average and momentum rules in specific markets and periods, particularly in foreign exchange and emerging markets during the 1980s and 1990s. Others, using stricter transaction-cost assumptions and out-of-sample tests, find that the same rules stop working, sometimes decades before they were published.

The pattern that emerges from the literature is one of fragility rather than fakery. Rules that worked in one decade, one market and one liquidity regime often stop working in the next, and when you subtract realistic spreads, commissions and slippage (the difference between the price you expected and the price you got), whatever edge remained frequently disappears altogether.

The honest summary runs roughly like this: in efficient, deep markets such as large-cap US equities, published tests generally find that mechanical technical rules fail to beat a passive index net of costs over long horizons. In less efficient corners, certain rules have shown some persistence, though nothing seems to survive publication forever. If you build your strategy on a paper you found online, it is safest to assume the edge is already gone or was never robust to begin with. The strongest regulator-published data on retail trading outcomes, from the FCA in the UK and ESMA in the EU, consistently shows that a large majority of retail CFD accounts lose money, whichever method those traders say they use.

The role of self-fulfilling prophecy and market psychology

One reason technical analysis appears to work in the short term has little to do with the future being encoded in past prices, and much more to do with the fact that millions of traders are watching the same lines. When enough participants place stop losses just below an obvious support level, price piercing that level triggers a cascade of selling, and the level becomes real precisely because everyone acted as if it already were.

The 200-day moving average is the clearest example of this dynamic. No law of physics makes 200 days special, yet because so many institutional and retail traders react to a break of that line, the break itself generates the follow-through it was supposed to predict. Round numbers behave in a similar way, as you can see with prices such as gold at $2,000, the S&P 500 at 5,000 or EUR/USD at 1.10.

This is a real effect and one you can trade, though it remains an unstable one. Self-fulfilling patterns keep working until the crowd changes its focus, until high-frequency algorithms front-run the retail reaction, or until a large news event overwhelms the technical setup entirely. Treating a self-fulfilling pattern as though it were a physical law is how traders get destroyed when the pattern eventually fails, because they sized the trade on the assumption that the signal was reliable.

Common mistakes that make technical analysis appear to work when it does not

The reason so many traders believe technical analysis is a settled winning method is that human cognition is built to see winners and forget losers. Three specific biases do most of the damage, and recognising them in your own trading is more valuable than any indicator.

Survivorship bias means you see the trader on social media who called a top with a trendline, not the ten thousand who drew the same line and were wrong. Confirmation bias means you notice the flag pattern that broke out as predicted and forget the three that failed last week. Curve-fitting, sometimes called overfitting, is the technical form: you tune a strategy on historical data until it looks perfect on the past, then discover it makes nothing on new data because you fitted noise, not signal.

Distinguishing a genuine pattern from random noise requires out-of-sample testing (running your rules on data you did not use to design them) and a large enough sample of trades to be statistically meaningful. A strategy with twelve trades is a story, not evidence.

How to use technical analysis without fooling yourself

If you choose to use technical analysis, treat it as one input into a decision, not a prediction. The traders who extract value from charts do it inside a framework that assumes any single signal is wrong roughly half the time.

Start with a written plan. Define the setup in words: the pattern, the timeframe, the confirmation, the invalidation. Define the position size as a percentage of account equity, not a number of contracts. A common rule is to risk no more than 1% of your account on a single trade, so a $10,000 account risks $100 per trade, and the stop-loss distance in pips determines the position size, not the other way around. A pip is the smallest standard price move in a currency pair, typically 0.0001 for most majors.

Backtest on a long enough period to include a bear market and a range-bound market, and reserve at least a third of the data for out-of-sample testing. Track every trade in a journal with the entry, the stop, the target, the outcome and, honestly, the reason you took it. After 100 trades, compare your net return with what a passive buy-and-hold would have delivered over the same period. If you are behind, the indicators are not the problem; the strategy is.

Finally, respect leverage caps. As of the FCA's product intervention rules, UK retail clients can access maximum leverage of 1:30 on major forex pairs, 1:20 on major indices and 1:5 on individual equities, and CFDs on cryptoassets are prohibited for UK retail. Those caps exist because higher leverage magnifies both signal and noise.

Technical analysis versus fundamental analysis: which one matters more

Fundamental analysis examines company earnings, cash flow, balance sheets, industry position and macroeconomic data to estimate what an asset is worth. Technical analysis ignores all of that and studies price and volume. The two approaches answer different questions, and treating them as competitors is a category error.

Fundamentals dominate over long horizons. If a company grows earnings for a decade, the share price follows, regardless of chart patterns. Technicals dominate over short horizons because in a five-minute window, no fundamentals have changed, and price is moved by order flow, sentiment and positioning. A long-term investor who ignores fundamentals is guessing; a day trader who ignores technicals is trading without seeing the current price. Neither method guarantees profit, and neither is a substitute for position sizing and a stop-loss.

The bottom line: does technical analysis work for you

Technical analysis works as a tool for structuring entries, exits and risk on short-term trades. It does not work as a crystal ball, and no combination of indicators turns a losing trader into a winning one. The traders who profit from charts do so because they are disciplined about position size, honest about their track record and sceptical of their own conviction.

If you want to use technical analysis, learn it properly, test it out of sample, size your trades small and benchmark yourself against buy-and-hold. Your edge is behaviour, not the moving average.