Technical Analysis Explained Honestly: Charts, Indicators and the Evidence

Technical analysis is the practice of studying past prices and trading volume, usually on charts, to try to predict future price moves. The academic evidence is mixed: some simple rules showed real predictive power in older data, but most of that edge shrinks or disappears once you account for trading costs, the many rules tested, and periods after the research was published.

In short: technical analysts believe that patterns in price and volume reflect the behaviour of buyers and sellers and can repeat. Common tools include support and resistance levels, trend lines, moving averages, and momentum indicators. Peer-reviewed studies have found that some of these signals carried information, especially in older data and in some markets, but there is little reliable evidence that ordinary investors can use them to beat a simple buy-and-hold approach after costs.

How it differs from fundamental analysis

Fundamental analysis asks what a business is worth, using its financial statements and prospects (see how to value a stock). Technical analysis largely ignores the business and looks at the market’s behaviour: where the price has been, how fast it moved, and how much stock changed hands. The underlying idea is that prices reflect collective psychology, and that psychology leaves patterns.

The basic toolkit

Price charts

Line charts connect closing prices.
Bar and candlestick charts show each period’s open, high, low and close. A candlestick’s “body” spans the open and close, and the thin “wicks” show the high and low.
Volume bars, usually beneath the price chart, show how many shares traded.

Support and resistance

Support is a price level where a falling stock has repeatedly stopped falling, on the theory that buyers step in there. Resistance is a level where a rising stock has repeatedly stalled, on the theory that sellers appear. When a price moves decisively through one of these levels, technical analysts call it a breakout (or breakdown), and the old resistance is often expected to become new support.

There is a plausible mechanism: many people remember round numbers and recent highs, and limit orders do cluster at some prices, as described in how stock prices are set. But whether a level is “real” is usually only clear in hindsight.

Moving averages

A simple moving average (SMA) is the average closing price over the last n days, recalculated each day.

A worked example with five closing prices for a hypothetical stock: $50, $51, $53, $52, $54.

5-day SMA = ($50 + $51 + $53 + $52 + $54) ÷ 5 = $260 ÷ 5 = $52.00

If the next close is $56, drop the oldest price ($50) and add the new one:

New 5-day SMA = ($51 + $53 + $52 + $54 + $56) ÷ 5 = $266 ÷ 5 = $53.20

Traders often watch the 50-day and 200-day averages. A common rule says to be “in” the stock when the price is above its moving average and “out” when it falls below. Another watches for crossovers: when a shorter average rises above a longer one, sometimes called a “golden cross,” or falls below it, a “death cross.” An exponential moving average (EMA) works the same way but gives more weight to recent prices.

Momentum and oscillators

Relative Strength Index (RSI) compares the size of recent up days with recent down days on a 0 to 100 scale. Readings above 70 are often called “overbought” and below 30 “oversold.”
MACD (moving average convergence divergence) tracks the gap between two exponential moving averages.

These indicators are all transformations of the same price history. They can make patterns easier to see, but they do not add new information beyond past prices and volume.

What the evidence says

This is where honesty matters, because the research does not support either “it never works” or “it works.”

Some rules looked strong in older data

In a 1992 paper in the Journal of Finance, William Brock, Josef Lakonishok and Blake LeBaron tested 26 simple moving-average and trading-range-breakout rules on the Dow Jones Industrial Average from 1897 to 1986. They found that buy signals were followed by higher returns than sell signals, a result that was hard to explain with standard models of returns. It was one of the most cited pieces of evidence in technical analysis’s favour.

But they did not hold up out of sample

In 1999, also in the Journal of Finance, Ryan Sullivan, Allan Timmermann and Halbert White revisited those rules and expanded the set to 7,846 variations. They used a statistical method designed to correct for data snooping: the danger that, if you test enough rules, some will look good by luck alone. Their findings:

Within the original 1897-1986 sample, the best rules’ performance survived the correction.
In the ten years after the original study, 1987 to 1996, it did not. The best rule as of the end of 1986 produced results that were not statistically significant, and the authors found “scant evidence that technical trading rules were of any economic value” in that period.
Using S&P 500 index futures from 1984 to 1996, where trading costs were low and short selling was easy, they found no evidence that the rules outperformed.

Some patterns contain information, which is not the same as profit

In 2000, Andrew Lo, Harry Mamaysky and Jiang Wang, writing in the Journal of Finance, used an automated pattern-recognition method on many US stocks from 1962 to 1996. They concluded that several technical patterns, such as head-and-shoulders and double bottoms, “do provide incremental information and may have some practical value.” They were careful to add that being informative does not by itself mean a pattern can be used to earn excess trading profits.

The broad survey: positive results with serious caveats

Cheol-Ho Park and Scott Irwin reviewed the literature in the Journal of Economic Surveys in 2007. Of 95 modern studies, 56 found positive results for technical trading strategies, 20 found negative results and 19 were mixed. Early studies tended to find profits in foreign exchange and futures markets but not in stock markets, and modern studies found profits in various markets at least until the early 1990s. But the authors also stressed that most studies suffered from problems in their testing, including data snooping, choosing rules after the fact, and difficulty estimating risk and transaction costs.

The most robust evidence in this area comes from momentum: the tendency for stocks that have done well over the past several months to keep doing relatively well for a while, and losers to keep lagging. Narasimhan Jegadeesh and Sheridan Titman documented it for US stocks in the Journal of Finance in 1993, and it has been studied extensively since. Momentum is a statistical pattern across large portfolios of stocks, rebalanced systematically. It is not the same as reading a chart of a single stock, and it comes with sharp, occasional crashes.

What happens to real traders

The most direct evidence is what people who trade frequently actually earn. A study of day traders in Taiwan by Brad Barber, Yi-Tsung Lee, Yu-Jane Liu, Terrance Odean and Ke Zhang, published in the Review of Asset Pricing Studies in 2020, found that day traders’ aggregate performance was negative and that 97% of day traders were likely to lose money in future day trading. The study drew on detailed trade records from the Taiwan Stock Exchange covering many years and a very large population of traders, not a handful of unlucky individuals.

A worked example: what costs do to a signal

Suppose a moving-average rule generates 20 round trips (a buy and a later sell) a year. Each round trip costs you the bid-ask spread once, plus any commission. Say that adds up to 0.15% per round trip on a liquid stock.

Annual trading cost: 20 × 0.15% = 3.0% a year

If the rule’s average edge before costs is 2% a year over buy-and-hold, it loses 1% a year after costs. On a less liquid stock with a 0.5% round-trip cost, the drag becomes 20 × 0.5% = 10% a year. In a taxable account, frequent selling also realises gains sooner than buy-and-hold would. In the US, gains on positions held a year or less are taxed at higher short-term rates. In Canada, very frequent trading can lead the CRA to treat gains as business income rather than capital gains. Small edges are easily consumed by costs that a backtest may understate.

What can go wrong

Seeing patterns in noise

Random price series produce trends, “head and shoulders” and “double bottoms” by chance. People are very good at seeing patterns and much less good at judging whether they are meaningful. Looking at a chart after the fact makes signals look far clearer than they were in real time.

Backtest overfitting

With modern software, you can test thousands of rule variations in minutes. The one that looks best on past data is very likely to be partly luck, which is exactly the data-snooping problem Sullivan, Timmermann and White measured.

Underestimating costs

Spreads, commissions, slippage on fast-moving prices, taxes, and the time spent all reduce returns. Backtests often assume you trade at the closing price with no market impact.

Leverage magnifies mistakes

Technical trading is often combined with margin or options, which turns small losses into large ones. See margin trading explained for how margin calls work.

Behavioural traps

Frequent trading tends to amplify overconfidence and the urge to act. Our article on investor behavior mistakes covers the research on how these habits affect returns.

The short version

Technical analysis uses past prices and volume to try to predict future moves. Some simple rules did carry information in historical data, and momentum is a well-documented effect across large portfolios. But the stronger the test, with out-of-sample periods, corrections for data snooping, and realistic costs, the weaker the evidence becomes. The research on actual frequent traders is the least encouraging of all. If you use charts, treat them as a description of what has happened, not a reliable map of what will.