Price action trading means making decisions from the raw price chart — candlesticks, swing highs and lows, trend, and horizontal levels — instead of from lagging indicators layered on top of it. The idea is that price already contains every input a moving average or oscillator is built from, so you read price directly.
That is a reasonable starting philosophy. It is also where a lot of confident storytelling begins, so this piece separates what price action is from what the evidence supports.
What price action trading actually is
Price action trading is a family of methods that read supply and demand straight off the candles rather than from derived signals.
A moving average, an RSI, or a MACD is a mathematical transform of past prices. It smooths, so it lags. A price action trader argues that if the information is already in the open, high, low, and close, you should read those four numbers and the shapes they form, not a delayed average of them.
In practice, price action work leans on three things:
- Candlestick patterns — the shape of one or a few candles (engulfing, pin bar, doji).
- Market structure — the sequence of swing highs and lows that defines a trend.
- Support and resistance — horizontal levels where price has previously stalled or reversed.
None of these are secret. They are the oldest ideas in charting, which is exactly why they need scrutiny rather than reverence.
Candlestick patterns, and the shapes traders name
Candlestick patterns are named formations that traders read as a hint about who is winning — buyers or sellers — over one to three bars.
The vocabulary is small and worth defining once:
- Bullish/bearish engulfing — a candle whose body fully covers the previous candle's body in the opposite direction, read as a momentum shift.
- Pin bar / hammer — a candle with a small body and one long wick, read as a rejection of the extreme the wick reached.
- Doji — a candle that opens and closes at nearly the same price, read as indecision or a possible turning point.
These read cleanly in hindsight. The harder question is whether they predict anything forward — which is where the research matters.
What the evidence actually says
Across markets, the academic evidence on candlestick-pattern profitability is mixed to weak, and edges that survive in-sample often vanish after costs.
The most cited stress test is Marshall, Young and Rose, who evaluated a large set of common candlestick patterns on Dow Jones stocks from 1992 to 2002 using a bootstrap that generates random OHLC paths. They found the patterns produced no value beyond chance (Marshall, Young & Rose, 2006, Journal of Banking & Finance). A later out-of-sample-style study on the Stock Exchange of Thailand reached a similar verdict: for most patterns the mean returns were not statistically different from zero, and adding oscillator filters barely helped (Tharavanij, Siraprapasiri & Rajchamaha, 2017, SAGE Open).
It is not unanimous, and honesty cuts both ways. In the less-efficient Taiwan market, Lu, Shiu and Liu reported that a few bullish reversal patterns were profitable, with the effect strongest in smaller, lower-priced stocks (Lu, Shiu & Liu, 2012, Review of Financial Economics). And support/resistance — a price action staple — has a genuine microstructure story: Carol Osler showed that support and resistance levels help predict intraday trend interruptions in FX, because stop and take-profit orders cluster at round numbers (Osler, 2000, FRBNY Economic Policy Review).
| Price action claim | What the research suggests |
|---|---|
| Candlestick patterns predict reversals in liquid US stocks | Little to no edge beyond chance (Marshall, Young & Rose) |
| Pattern edges hold after transaction costs | Most patterns' returns not different from zero once tested (Tharavanij et al.) |
| Patterns can work in some markets | A few bullish patterns paid in less-efficient Taiwan stocks (Lu, Shiu & Liu) |
| Support/resistance levels have real predictive value | Yes, intraday in FX, tied to order clustering (Osler) |
Market structure and the appeal of fewer indicators
Market structure describes a trend as a sequence: an uptrend is higher highs and higher lows, a downtrend is lower highs and lower lows, and a break of that sequence is read as a possible change of trend.
uptrend -> HH HL HH HL (higher highs, higher lows)
downtrend -> LH LL LH LL (lower highs, lower lows)
structure break -> the pattern of highs/lows stops repeatingThis is the intuitive core of price action, and it explains the appeal. It uses no external parameters, so it looks like it should transfer cleanly across every asset and timeframe — a doji is a doji whether you trade EURUSD or gold. Structure reading is also the backbone of Smart Money Concepts, which packages structure, liquidity and order blocks into a single narrative.
But "no parameters" is partly an illusion. Deciding what counts as a swing high, how large a body must be to "engulf," or which wick is long enough to be a pin bar are all parameters — you have just moved them from a settings box into your own judgment.
The subjectivity problem
The biggest weakness in price action is not the tools; it is that the same chart supports several stories, and confirmation bias picks the flattering one.
Because a pin bar or a structure break is defined loosely, a trader can find one almost anywhere after the fact — you remember the ones that preceded a reversal and quietly forget the ones that did not. A discretionary "it worked in review" feeling is not evidence; it is a highlight reel.
The fix is to make the rules objective enough to be counted by a machine. If "bullish engulfing" means the close is above the prior open and the current body is at least X% larger than the prior body at a level defined by a specific rule, then you can test it. If you cannot write it down that precisely, you cannot backtest it — and then you cannot know its win rate, its drawdown, or whether it survives spread and commission. That is the same trap covered in why your backtest lies: a method that resists precise definition tends to resist honest measurement too.
Frequently asked
Do candlestick patterns actually work?
The evidence is mixed to weak. Rigorous tests on liquid US stocks found no edge beyond chance, and most patterns in other markets show returns close to zero after costs. A few patterns have shown profitability in less-efficient markets, but that is market-specific and needs its own out-of-sample proof.
Is price action better than using indicators?
Neither is inherently better. Indicators lag because they are transforms of price; price action is more direct but more subjective. What matters is whether the rules — from either camp — are defined precisely enough to test, and whether the tested edge survives real trading costs.
Can price action strategies be backtested?
Yes, but only once every rule is objective: an exact definition of the pattern, the level, and the entry and exit. A discretionary "I know it when I see it" version cannot be backtested, so its performance is unknown rather than good.
Why do support and resistance levels seem to work?
There is a real microstructure reason in FX: stop-loss and take-profit orders cluster at round numbers, so those levels genuinely influence intraday price. That is different from claiming any hand-drawn line is predictive.
The stubborn takeaway
Reading raw price is a legitimate lens, and a couple of its ideas — clustered levels especially — have real support in the data. But a candle shape that looks obvious in review is the easiest thing in trading to fool yourself with. If your price action rules are precise enough to survive an objective backtest across real costs, keep them. If they only work when you are the one deciding, in hindsight, what counts as a signal, you do not have an edge — you have a story. Test it, don't trust it.