Does technical analysis work?
Somewhere in technical analysis there are rules that genuinely clear buy & hold — a real minority do, and finding one is worth everything. But “technical analysis” isn't one thing, and the honest answer depends on which part you mean. We tested the part that can be tested — the famous mechanical rules that act only on a closed bar — the same honest way as everything here: costs on, no hindsight, always measured against buying and holding. Here's what that part shows, and here's the discretionary part no backtest can settle — said plainly, not sold.
As of September 25, 2026 · frozen census release · DOI 10.5281/zenodo.22974116
A frozen, reproducible snapshot — not a live number and not a promise about the future. 25 famous mechanical strategies × 36 famous US stocks & ETFs = 900 strategy-asset combinations, of which 839 had enough trades to judge; each is measured against that asset's own buy & hold, net of modelled costs, over one ~five-year window, with no out-of-sample split, and is not significance-tested.
Technical analysis isn't one thing
“Technical analysis” is a broad practice, not a single method. Under the one name sit very different things: mechanical indicator rules that reduce to an exact instruction a computer can follow — moving-average crossovers, RSI, MACD, Bollinger Bands, the Ichimoku cloud, breakouts; chart patterns and support, resistance and trend lines drawn by hand; candlestick reading; and discretionary, multi-timeframe and intraday judgment. They are not the same claim, and they cannot all be checked the same way — so the honest answer has to say which part it is answering.
RSI, MACD, moving averages, Bollinger, Ichimoku, breakouts — an exact instruction a backtest can run twice. This study measures these.
Support & resistance, order blocks, supply-demand zones, chart patterns — placed by eye, so there is no single rule to test. We decline to score these.
Reading charts by hand, trading inside the day — out of a daily-bar census's reach. We make no claim either way.
What we could measure — the honest result
A rule we can test is one that reduces to a fixed, closed-bar instruction — the same input always gives the same output, so a backtest can run it the same way twice. The famous mechanical indicators are like this, so we backtested them: on daily bars, across famous US stocks and ETFs, with fees and slippage charged on every fill, never acting on a price the rule could not have known, always against simply buying and holding the same asset.
On the part we could test — the famous mechanical rules on famous US stocks and ETFs — 721 of 839 judged backtests (86%) lagged simply holding the same asset. A real minority — 118 (14%) — genuinely cleared it. That is a result for the mechanical, closed-bar part of technical analysis, not a verdict on all of it.
This is the summary, not the scoreboard — see the full honest scoreboard, every winner and loser → or read the per-rule studies: RSI, MACD, Ichimoku and Bollinger.
For what that base rate means for an idea you already have: what beating buy & hold means for your strategy →
What no backtest can settle
The parts that do not reduce to a fixed rule cannot be backtested the same way twice — a zone drawn by eye, an order block marked after the move, a pattern spotted in hindsight will be placed differently by two chartists, so there is no single rule to run. We decline to put a number on those rather than pretend a drawing is a rule, and we say so plainly.
And a skilled discretionary trader's intraday judgment is out of this census's reach entirely: it runs on daily bars and closed-bar rules, so it can say nothing for or against reading charts by hand or trading inside the day. We make no claim there.
On whether people actually make money trading inside the day — a separate question our daily-bar data can't answer — the honest evidence is in the research: can you make money day trading? →
We take the hand-drawn Smart-Money constructs one at a time — and say why we decline each: order blocks, supply & demand zones, liquidity sweeps and break of structure.
So — does it work? The honest read
So — does it work? On the honest evidence, as a repeatable edge over simply holding, the testable part mostly did not: most famous mechanical rules lagged buy & hold once costs were charged, and a beat on this survivor-only set of famous names is flattered by hindsight and by the many variants one could try. A real minority genuinely cleared the bar — which is why the question is worth asking honestly rather than dismissing. The value its defenders point to — framing risk, marking in advance where an idea would be wrong, making a plan repeatable — is a different claim from predicting price, and one a return backtest was never built to settle.
This isn't “technical analysis is useless” — a real minority of these rules did clear buy & hold, and for many traders the practice earns its keep as a way to frame risk, not to predict price.
On crypto the same rules cleared holding far more often, but that is a trend-mirage flattered by overfitting and by crypto's far larger survivorship gap — dead coins and rug-pulled tokens vanish from the free data entirely — so the composition-stable headline here stays the stocks-and-ETFs slice, and the blended figure is context, not the headline.
These are well-known coins that still trade today. Dead coins, rug-pulls and delisted tokens aren't in the set (survivorship bias — larger here than for stocks) — and that history isn't available from the free public data this tool runs on — so this rate isn't a representative base rate for crypto as an asset class.
This is one roughly five-year window on about 36 hand-picked famous US names, survivor-only, with no out-of-sample split and no significance test — an honest slice, not the last word even on the rules it covers. These are famous names that still trade today. Names that delisted, went bankrupt or were wound down aren't in the set (survivorship bias) — and that history isn't available from the free public data this tool runs on — so this rate isn't a representative base rate for every stock or ETF that has ever traded.
Test a rule you've heard of on the asset you care about — free, no sign-up.
These are simulated backtests over roughly a five-year window on a survivorship-biased set of famous names, frozen as a citable release — hypothetical results, not live figures and not advice. Past performance is not a reliable indicator of future results. This is an educational, analytical tool — not investment advice, and not a recommendation to buy or sell anything. You make all decisions and execute on your own broker.