Reversal patterns — double tops, double bottoms, head and shoulders
Why this lesson exists
Most retail content treats reversal patterns — double tops, double bottoms, head and shoulders — as reliable signals. They are not. They fail roughly half the time on first attempt, and the failures are often more expensive than the successes because traders who "saw the pattern" refuse to exit when it stops working.
This lesson covers the classic reversal patterns, the volume and context cues that separate real ones from noise, where the stops and targets go, and when each pattern fails. The Pattern Lab serves these with roughly 40% failures by design — the realistic mix in live markets. Selection matters more here than in any other strategy in this module.
Double tops and double bottoms
A double top is two roughly equal highs separated by a pullback, followed by a break of the pullback low (the "neckline"). The pattern signals exhaustion: the second test of the high couldn't push further, and the subsequent neckline break confirms supply is taking over. Double bottoms are the mirror image — two roughly equal lows with a neckline break above.
The 2021 cycle top. BTC printed an all-time high near $69,000 on 2021-11-10, retraced to $42,000, and made a lower second high at $48,000 on 2022-03-28. The structure — a failed retest after a months-long pullback — is the same family of signal. The neckline at roughly $33,000 broke in May 2022 alongside the LUNA collapse, and BTC went to $17,000.
The 2024-03 mini double-top failure. BTC printed $73,100 on 2024-03-13 and $73,777 on 2024-03-14 — a classic double-top setup that retail shorts piled into. BTC then made a new high above $74,000 later that month and again in October. The pattern failed because the macro regime (ETF inflows) hadn't changed.
High-conviction double-top recipe: two highs within ~1% of each other, a 3–5%+ pullback between them, lower volume on the second peak, a confirming neckline break on volume, and higher-timeframe context that supports exhaustion (long prior uptrend, distribution signs, momentum divergence). Miss any of the last three cues and the hit rate falls steeply.
Head and shoulders (and inverse)
A head-and-shoulders pattern is three peaks: a left shoulder, a higher centre (the head), a lower right shoulder. The neckline connects the two intervening lows. A close below the neckline confirms the pattern. Inverse H&S is the same structure flipped — three troughs with the middle one the lowest, neckline above.
The inverse H&S at the FTX bottom. BTC printed a left shoulder at $18,500 (September 2022), a head at $15,500 (the 2022-11-21 FTX bankruptcy bottom), and a right shoulder at $16,500 (December 2022). The neckline at $22,000 broke on 2023-01-12. Measured-move target: $22K + ($22K − $15.5K) = $28,500. Hit on 2023-04-14 within 1% of projection.
High-conviction H&S recipe: head clearly above both shoulders (not a triple top), right shoulder visibly lower than head and similar to left, volume declining from left shoulder through head to right shoulder, neckline break on volume with 1–3 candle follow-through, measured-move projection landing at a real structural level (not floating in empty space). The neckline can tilt slightly; steep tilts above 5–10° reduce reliability.
Where stops and targets go
Double top — short. Entry on close below neckline. Stop above the second peak with cushion. Target the measured-move (peaks-to-neckline distance projected down); second target at the next significant prior support. For 2024-03: short below $73,000, stop $73,800 — risk ~₹66,000 per BTC. The pattern failed, but the structural stop made the loss cost 1% of risk, not 10%.
Double bottom — long. Mirror: entry above neckline, stop below second low, measured-move up.
Head and shoulders — short. Entry on neckline break. Stop above the right shoulder (not the head — that's too wide and ruins R:R). Target the measured-move down.
Inverse H&S — long. Mirror.
The 2023 FTX inverse H&S: entry above $22,000, stop below $19,500. Measured-move target $28,500. R:R 2.6:1. Played out over three months.
When it fails
Reversal patterns fail in three dominant modes, and each has a specific signature.
Regime hasn't actually changed. The 2024-03 double top is the canonical case: pattern complete, neckline broken, short fills, price goes higher anyway. ETF inflows were the real signal; the chart pattern was noise on top. Before taking a reversal trade, identify what would have to change for the trend to actually end. If you can't name a structural cause (macro shift, distribution evidence, prior cycle exhaustion), the pattern is probably going to fail.
Premature entry. A trader takes the short on the second peak itself, anticipating the neckline break. Sometimes price never breaks the neckline — it rallies from the second peak to a higher third peak, invalidating the pattern. The pattern isn't a pattern until the neckline breaks.
Stop placement too tight. A reversal trade entered on the neckline break with a stop just above the break level (instead of above the right shoulder or second peak) gets shaken out by normal noise. The structural invalidation level is the peak/shoulder, not the neckline. Wide stops mean smaller position sizes (see position sizing) — but the alternative is being right about the pattern and still losing the trade.
Risk and sizing
Wider stops mean smaller position sizes. On a ₹5L Standard xtree account at 1% risk: ₹50,000 stop distance → 0.1 BTC; ₹2,00,000 stop distance (macro-scale reversal on weekly chart) → 0.025 BTC. Per-trade rupee risk is unchanged because the sizing formula is leverage-neutral.
R:R expectations: high-conviction reversal patterns typically offer 2:1 to 4:1 with a 50–60% hit rate. The expectancy is positive but thin, which is why selection matters. A trader who takes every visible double top hits closer to 30–35% — at 2:1 R:R, that's negative expectancy. The math is in risk-reward and expectancy.
The consistency rule matters specifically for reversal trades because wins are clustered. A textbook H&S that prints ₹40,000 in a day when other winning days are ₹3–5,000 creates a 40%-best-day breach. The defence: scale out aggressively on outsized winners. Don't optimise the single trade at the cost of the evaluation.
On xtree, multi-day reversal trades incur compounding funding costs. A short held through three 8-hour windows at 0.05% on ₹10L notional is ₹15,000 of funding. A setup with apparent 2:1 R:R can become 1.5:1 after funding and 1.2:1 after funding plus the 10 bps round-trip fee. Build holding costs into the R:R math before entry.
Worked example
ETH on the daily has run from ₹1,40,000 to ₹2,40,000 over four months. Price prints a high at ₹2,40,000 on 2024-04-15, pulls back to ₹2,05,000 over three weeks, then makes a second high at ₹2,38,500 on 2024-05-20 — within 1% of the first peak. Volume on the second peak is ~65% of the first. The pullback low at ₹2,05,000 is the neckline.
Recipe check: two highs within 1% — yes. Pullback ~15% — yes. Lower volume on second peak — yes. Macro context: ETF chatter cooling, no clear continuation catalyst. Need neckline break to confirm.
On 2024-06-05 a daily candle closes at ₹2,02,000 — below the neckline. Volume on the break is 1.8× the prior week's average. Confirmation hit.
- Entry: ₹2,02,000. Stop: ₹2,41,000 (above the higher peak). Risk per ETH: ₹39,000.
- Measured-move target: ₹1,70,000 area. Reward: ₹32,000. R:R: 0.82:1.
That R:R is poor. If the chart shows structural support at ₹1,80,000 between entry and projection, the realistic target is ₹1,80,000 — reward ₹22,000, R:R 0.56:1. Worse. The disciplined response: skip the trade. This is the pattern most retail traders take anyway because "it's a textbook double top." The expectancy math says don't. Compare with the FTX inverse H&S where the measured-move was 2.6:1: that's a take, this is a skip. Same pattern family, different math.
Common misunderstanding
"The pattern is the trade." The pattern is the setup; the trade is the pattern plus the regime context plus the R:R math. The 2024-03 BTC double top was technically complete and failed because the trade ignored the regime. The same pattern in May 2022 worked because the regime was breaking.
Second misunderstanding: that reversal trading is more "sophisticated" than trend trading and therefore better. It isn't. Worse hit rates, more context judgement required, higher failure costs. Most professionals take more trend trades than reversal trades because trends are easier to identify than tops. Reversal trades belong in the toolkit but should be the exception, not the default.
Recap
- Double tops, double bottoms, and H&S are setup candidates, not signals. The neckline break with volume confirmation is the trade trigger.
- Stop above the right shoulder (H&S) or second peak (double tops). Wider stop is structural, not optional.
- Targets are measured-moves cross-checked against actual chart structure. Floating targets rarely get hit.
- Reversal patterns fail when the regime hasn't shifted, entry is premature, or the stop is too tight. 2024-03 BTC is the canonical case study.
- Selection is 70% of the edge. Pass on patterns where the R:R math doesn't work.
Test yourself
Next up: Module 4 covers the xtree platform itself — placing orders, managing margin modes, and reading your equity curve. Module 3's strategy vocabulary becomes actionable on the actual interface.