Does technical analysis work?
Last updated
In-sample, not a promise. Every figure on this page comes from a backtest on past prices.
Short answer: in our tests, no. We tested 1,342 chart-based trading strategies. 265 of them beat a simple benchmark on the raw number. After a check for luck, 0 still stood.
What technical analysis means
Technical analysis is trading from price charts. You use patterns, moving averages and indicators such as RSI and MACD to decide when to buy or sell, instead of looking at a company's profits.
What we tested
We wrote each popular rule down as exact code, ran it on past prices of stocks, ETFs and crypto, and charged it modelled trading costs. That gave 1,342 strategies in 77 families, from moving-average crossovers to the golden cross. The yardstick was a benchmark that holds the same assets and pays no costs.
What came out
- 1,342 strategies tested.
- 265 finished ahead of the benchmark on the raw number.
- 39 of those also had a p-value under 0.05 (a common measure of how unlikely luck is) before any correction.
- 0 of 1,774 test runs passed the luck check (Benjamini-Hochberg, the standard fix for trying many things at once).
- The typical strategy trailed the benchmark by 8.5 percentage points a year.
Why 265 ahead still means 0 passed
Toss a thousand coins ten times each and a few will land heads nine times out of ten. Nothing special about those coins. A strategy tested on past prices works the same way: try enough of them and some look brilliant by chance. The luck check asks how many would look this good with no real edge at all. For ours, luck could explain every one.
What this does not mean
It does not prove that no rule ever works. We tested rules that are published or popular, on daily prices, over one stretch of history. A rule we did not test, or a different market, could behave differently. It does mean that a rule being popular, or looking good on a chart, is not evidence.
Questions
Does technical analysis work?
Not in our tests. Of 1,342 strategies, 265 finished ahead of the benchmark on the raw number, and 0 of 1,774 test runs passed once we corrected for luck. That is in-sample and does not prove no rule ever works.
Why do so many traders say it works?
Test enough rules on past prices and some will look good by chance. Without a correction for how many rules were tried, a lucky one looks like skill.
Can I still use charts?
Yes, as a way to see what the price has done. Our free score sums up ten common indicators into one number out of 10. It describes the chart today. It is not a prediction.
See for yourself
Every test, the failed ones too, is on the evidence page. For one rule at a time, read the golden cross, the RSI strategy or the moving-average crossover. To see the charts side, get a free score.
General information only. It doesn't consider your objectives, finances or needs. Tickfloor holds no financial services licence and never places trades.