Price Action

Price action is the family of trading approaches that derive their decisions primarily from the price series itself — the open, high, low and close of each bar, and the structure those bars build over time — rather than from indicators computed on top of it. The argument is not that indicators are useless, but that each is a lagging summary of the same data; reading the bars directly stays closer to the moment supply and demand shift.

Most price action strategies are not a single pattern. They stack three layers, and the quality of an implementation usually comes down to how strictly each layer is defined.

**Structure** establishes context: where the swing highs and lows sit, whether the market is making higher highs and higher lows, whether a prior swing point has been broken, and whether price is trending or rotating inside a range.

**Location** narrows the chart to the few prices worth acting at: prior day high and low, session opens and ranges, untested swing points, round numbers, or the origin of an earlier impulsive move.

**Trigger** times the entry: a rejection wick, an engulfing bar, an inside-bar break, a failed break that returns into range, or simply a close beyond a level.

A rule set with all three layers is testable. One with only a trigger — “buy the bullish engulfing bar” — is a pattern rather than a strategy, because the same candle carries different information at the exhaustion of an extended move than at a level the market has already respected several times.

## Main variants

Implementations cluster into a few families. Trend-continuation systems wait for a break of structure and trade the pullback or retest in the direction of the break. Range and level-to-level systems fade the edges of a defined range and target the next level rather than a fixed distance. Breakout systems trade the move out of a compression — an inside bar, a narrow overnight range, an opening range — often requiring a retest before committing. Session-based variants anchor everything to a clock: the Globex range, the first hour of a cash session, the close of a specific hourly candle. Liquidity-based or “smart money” variants formalise the location layer into named objects: sweeps of prior highs and lows, imbalances, order blocks. Hybrid versions keep price action as the trigger but delegate the bias to an indicator such as a moving-average ribbon. The same skeleton appears from one-minute scalping to daily swing trading, and across indices, futures, forex and crypto.

## What separates one implementation from another

Two strategies described in the same words can behave very differently, and the differences that matter are mechanical: how a level is defined (a single price, a zone, or something drawn by eye); whether a break requires a close beyond the level or only a touch; whether entry is a market order on the signal close, a limit at a retest, or a stop beyond the bar; where the stop sits (behind structure, at a volatility multiple, at a fixed distance); how the target is chosen (a multiple of risk, the next level, a trail); which higher-timeframe or time-of-day filter can veto a signal; and what happens after a stop-out. Those choices, not the pattern name, shape the trade distribution.

## Common mistakes

The most frequent is trading the trigger without the location — patterns evaluated anywhere on the chart lose most of their meaning. Close behind is retrospective level drawing: levels that look obvious once the reaction is visible are rarely the ones that were drawable beforehand. Rules that cannot be written as unambiguous conditions leave the outcome to whoever trades them, which makes results non-reproducible. On short timeframes, spread and commission frequently exceed the edge being measured, and patterns are often assumed to transfer across instruments and sessions without retesting.

## Evaluating and backtesting a version

Start by rewriting the strategy as pseudocode with no undefined terms. Anything you cannot express — “strong candle”, “clean level”, “obvious trend” — is where the discretion, and later the disagreement about results, will live.

Then check the mechanics of the test. When a bar's range contains both the stop and the target, an hourly or daily backtest cannot know which was hit first; use lower-timeframe data for fills, or assume the worse outcome. Apply realistic spread, commission and slippage, then re-run with them doubled to see how much survives. Keep a held-out period the rules were never inspected on, and prefer walk-forward when parameters are involved.

Finally, look past the aggregate. Examine the distribution of trades, the dependence on a handful of outliers, and how results change at neighbouring parameter values — a robust rule set degrades smoothly rather than collapsing one setting away. Compare against a stripped-down baseline: the same levels without the candle trigger, or the same trigger without the session filter. If the baseline behaves similarly, the extra component is decoration.

The versions collected on this page decode how individual traders specified those choices. Reading several side by side is usually more informative than studying one: the recurring elements tend to be structural, and the differences show where judgement was applied.

Strategies in this concept (160)

Frequently asked questions

Is price action trading the same as trading without indicators?

No. Price action describes where the decision information comes from, not a ban on tools. Many versions use an indicator as a directional bias filter or a volatility measure while keeping entry, stop and target rules based on structure, levels and bar behaviour. What defines the category is that the signal itself is read from the price series; a version whose entry condition is an oscillator crossing is an indicator strategy regardless of how the chart is drawn.

Which timeframe is price action best suited to?

The logic — structure, location, trigger — is timeframe-agnostic, but the economics are not. On one- and five-minute charts the average move is small relative to spread, commission and slippage, so a version that looks viable in a costless test can be unprofitable in practice. Higher timeframes reduce that pressure but produce fewer trades, so a sample large enough to judge takes longer to accumulate. Choose the timeframe together with the cost assumptions, not before them.

How do “smart money” concepts relate to price action?

They are a vocabulary layered on the same raw data. Liquidity sweeps, order blocks and fair-value gaps are named ways of specifying the location layer — where a level is and why it might matter — rather than a different information source. The naming makes some setups easier to define precisely, but it does not make them automatically more testable. The same questions apply: is the zone drawable in advance, and does the rule behave differently from a plain prior-high / prior-low version?

Can a price action strategy be backtested automatically?

Yes, provided every term is defined. Swing points, break of structure, engulfing bars and range boundaries all have codable definitions, though several defensible ones exist for each — a swing high confirmed by two bars on either side is not the same as one confirmed by five. The practical test is whether two people coding from the same description would produce the same trade list. If not, the description needs tightening before any result means anything.

Why do two traders following the same price action rules get different results?

Usually because part of the rule set was never written down. Level selection, what counts as a valid trigger, and whether a marginal setup is taken are judgement calls unless explicitly constrained. Differences in instrument, session, broker spread and stop placement add further variance. When comparing versions, look first at how much of the method is mechanical and how much depends on the reader's interpretation.

All strategy concepts · Explore all strategies