Edges
Chart Reversal Patterns: Every Way Price Turns at a Zone
Chart reversal patterns are usually taught as a picture book — memorise the shapes, spot them, trade them. That is why so many traders can name a double top and still lose money at one. The shape is the result. What you actually have to read is the behaviour that produced it, and at any given zone price only has two ways of turning: it rejects immediately, or it stalls first and turns later.
Everything below is a catalogue of those two behaviours — the candle signatures of an instant rejection, the structures that form during a stall, the win rates each carries at each timeframe, and the sequence that separates a reversal from a pause.
The two behaviours, before any pattern name
This is the whole map. A zone does not promise a reversal — it promises a decision, and the decision arrives in one of those two shapes. Which one you get changes what you are allowed to do: behaviour 1 gives you force to read and almost no time; behaviour 2 gives you time and a structure to wait for, and punishes you for entering before it resolves.
Behaviour 1: instant rejection at the zone
Price reaches the area and the opposing side answers inside one or two bars. No pause, no range, no time to deliberate. What you see:
- The wick appears first. Price pierces into the zone and is thrown back out, leaving a long tail. The wick is the whole message: an attempt was made and refused.
- A pin bar — small body, dominant wick — forms right at the level.
- An opposing candle engulfs the previous body entirely. Not hesitation, a takeover.
- The reversal begins before the zone is reached. Buyers or sellers step in early because they are positioned at an area, not a price.
The single-candle version is fast and often unconfirmable — the follow-through you want may never print separately, because the rejection is the event. That is the trade-off you accept: you are reading force in real time rather than waiting for structure to form.
Behaviour 2: consolidation before the reversal
The second behaviour is slower and far more common. Price arrives, and instead of turning, it stops being able to continue. The tells are consistent:
- candle bodies shrink; each bar covers less ground than the last
- wicks appear on both sides — no side can hold its extreme
- price trades in a narrow band, repeating highs and lows in the same area
- attempts to break the band fail and return inside
- the liquidity resting above the range high or below the range low gets swept
- then a decisive break out of the band, with force, in one direction
That is a consolidation — a sideways range where buying and selling pressure are temporarily balanced. It is the market taking time to decide, and the decision is made at the edges of the range, not in the middle.
Sometimes it forms a named pattern, sometimes it doesn't
When the swings inside a consolidation are clean enough, you get the classical shapes: double top, double bottom, triple top and bottom, head and shoulders, triangles, wedges, three-swing structures. These are not a separate category of thing. They are consolidation behaviour that happened to be structured enough to be seen as a picture.
More often it is formless — a cluster of green and red bars with wicks in both directions, overlapping in the same area. That is a perfectly valid consolidation and it carries the same information as a textbook double top. Do not force a name onto it.
**Price arrived → it could not continue → it consolidated → pressure built → which side won?**
That sequence is what you actually trade. "What pattern is this?" is a question about vocabulary; the sequence is a question about who is in control.
The full catalogue: 23 reversal formations, with win rates by timeframe
Filter by behaviour and direction, and switch the timeframe selector to see how quickly the edge erodes as you speed up.
Three things in that gallery deserve to be said out loud rather than left in a card.
Win rates fall as the timeframe falls, and it is not a gentle slope. Head and shoulders is 83% on the daily and roughly 53% at five minutes. A pin bar goes 75% to 50%. At the fastest timeframes many of these are coin flips before you have paid a spread — and the round-trip cost of trading an option runs 0.06 to 0.27 R, which is far more than a two-point edge over random can pay for.
Multi-bar structures need bars. A head and shoulders needs 30–40 candles to form. On a one-minute chart that is most of an hour, and the "pattern" you think you see in six bars is not one. This is why the chart-pattern column reads N/A at the fastest speeds rather than showing a small number.
Location changes everything. The same hammer is a different trade at a tested demand zone than in the middle of an established downtrend, where it is just a pause in the selling. Our breakdown of how reliable chart pattern win rates really are goes further into how quickly these figures decay once conditions differ from the study that produced them.
| Formation | Behaviour | Daily | 1h | 15m | 5m |
|---|---|---|---|---|---|
| Head & Shoulders | consolidation | 83% | 70% | 60% | 53% |
| Morning / Evening Star | instant | 80% | 65% | 58% | 52% |
| Double Top / Bottom | consolidation | 80% | 65% | 52% | 52% |
| Triple Top / Bottom | consolidation | 78% | 66% | 58% | n/a |
| Ascending / Descending Triangle | consolidation | 78% | 70% | 65% | 58% |
| Engulfing | instant | 75% | 66% | 62% | 56% |
| Pin Bar | instant | 75% | 65% | 55% | 50% |
| Rising / Falling Wedge | consolidation | 73% | 62% | 57% | 52% |
| Hammer | instant | 72% | 62% | 58% | 53% |
| Harami | instant | 72% | 63% | 55% | 49% |
| Piercing / Dark Cloud | instant | 68% | 60% | 55% | 50% |
| Shooting Star | instant | 65% | 58% | 52% | 48% |
| Broadening / three-swing | consolidation | 62% | 54% | 50% | n/a |
Where do these numbers come from, and how much should I trust them?
The daily, 4-hour and 1-hour figures are historically observed rates measured to the next-bar close. The 15-minute and 5-minute columns are derived rather than measured: no large-scale published study provides per-pattern win rates at those speeds, so they are extrapolated from the documented timeframe-degradation gradient plus category-specific decay rates — candlestick singles decay around 5% per step, multi-bar formations around 6%, volume-based patterns around 3%. Treat the fast columns as benchmarks to calibrate against your own instrument, not as facts. And treat every number as conditional: a win rate measured across all market conditions tells you very little about this pattern, in this zone, in this regime.
The trap: sideways resolves both ways
This is the single most expensive misunderstanding in reversal trading. A consolidation at a zone is not evidence of a reversal. It is evidence of balance, and balance breaks in whichever direction has more behind it.
The exact same cluster of overlapping candles can be:
- distribution — the opposing side absorbing every attempt to continue, before a reversal; or
- accumulation — a pause where the trending side reloads, before continuing.
The candles inside the range cannot tell you which. They look identical, because they are identical — the difference is not inside the range, it is at the edges. That is why the answer lives in what happens to the liquidity sitting above and below it: our piece on liquidity sweeps and stop hunts covers why the range extremes are where the information is.
The read sequence: what actually confirms a reversal
For behaviour 2, four things in order. Each one is weak alone; agreement between them is what the word confirmation is supposed to mean.
- Liquidity. The range high or low is taken and immediately reclaimed. Stops resting there have been collected — the fuel for the move in the opposite direction.
- Structure. Price breaks the prior swing against the trend that brought it here. This is where a change of character separates itself from an ordinary pullback, and where market structure hierarchy matters: a break on your entry timeframe means nothing if the higher timeframe structure is untouched.
- Price action. A force bar out of the range, an engulfing candle, a clean rejection — the candle-level evidence that one side has actually taken control rather than merely poked at it.
- Entry. Only when several elements point the same way, with the stop beyond the rejection extreme and the target decided before the click.
Where are you in that sequence right now? Tick what price is doing and find out:
Putting the two behaviours side by side
| Behaviour 1 — touch and turn | Behaviour 2 — stall and turn | |
|---|---|---|
| Speed | fast, often one or two bars | slow, can take dozens of bars |
| What you see | rejection wick, pin bar, engulfing | shrinking bodies, overlap, a range or a pattern |
| Confirmation | may never print separately | sweep, structure break, force bar |
| Main risk | entering on arrival, not on rejection | assuming sideways means reversal |
| Requires | reading force in real time | patience and an edge event |
Neither is better. They are different market conditions, and the mistake is applying the method of one to the other — waiting for a structural confirmation that a fast rejection will never give you, or trading a slow consolidation with the urgency of a rejection candle.
The takeaway
Your zone has exactly one job: tell you where to start watching. Price behaviour has the other job: tell you whether it is time to decide.
Before your next reversal entry, answer three things in order:
- Which behaviour is this? Instant rejection, or a consolidation that has not resolved?
- If it is a consolidation, what has actually happened at the edges? A sweep, a structure break, a force bar — or nothing yet?
- Do several elements agree? One is a story. Several pointing the same way is a setup.