Double Top (M Pattern) — Recognition, Confirmation and Target
The double top — the M pattern — is, alongside head and shoulders, the most commonly watched-for uptrend-reversal pattern. The intuition is simple: the market tries to break the same level twice, fails twice, so the bulls are weakening. The trouble is, intuition without measurement can be expensive. Thomas Bulkowski's measurements reveal two things most guides stay quiet about: first, the average decline from a double top is moderate (16-19%, depending on the variant); second — and this is the real gem — a busted double top later rises an average of 38%, a bigger move than the pattern delivers when it "works."
In other words, the most interesting thing about the M pattern is often what happens when it fails. Let's go through it.
How to Recognize a Double Top
The pattern consists of two peaks at a similar level, separated by a clear trough, after an uptrend.
Validity conditions:
- An uptrend before the pattern. The pattern reverses a rise — without a rise, there's nothing to reverse.
- Two peaks at a similar level. The average difference in Bulkowski's database is about 1%; the peaks should look like they stopped at the same barrier.
- A clear trough between the peaks. The correction between the peaks should reach at least around 10% (on stocks; on crypto, scale it proportionally to volatility). Two peaks separated by a shallow dip is just consolidation.
- Time gap. The peaks are usually separated by a few weeks to a few months — setups with two peaks a day apart are noise, not a pattern.
- Confirmation: a close below the trough between the peaks. This is a hard requirement. Bulkowski calls unconfirmed setups "squiggles" — scribbles on the chart. A significant share of peak pairs never confirm at all, and price just goes higher.
Bulkowski splits peaks into Adam (narrow, spiky, often a one-day spike) and Eve (wide, rounded) types — giving four pattern variants: Adam&Adam, Adam&Eve, Eve&Adam, Eve&Eve. Differences between the variants exist but are secondary to the confirmation question. Worth knowing they exist, since you'll see them listed separately in statistics tables.
Volume is usually higher on the first peak than on the second — fading demand on the second attempt is the pattern's classic signature.
[Chart coming soon: Diagram of the M pattern — uptrend, first peak A, a correction of about 10% to trough C, second peak B at the level of A, a horizontal confirmation line at the height of trough C, a downward arrow at the candle close below the line labeled "confirmation," a marked pullback to the line from below and the measure-rule target below]
What the Numbers Say (Bulkowski, Encyclopedia of Chart Patterns)
First, the terms:
- Break-even failure rate — how often price, after confirmation, didn't even drop 5%. The pattern "fired" but the move died on arrival.
- Average decline — the average move from the breakout price to the lowest trough before a bounce of at least 20% (the ultimate low), calculated on perfect trades. Use it to compare patterns against each other, not as a forecast of your own P&L.
- Target % — the share of cases where price reached the measure-rule target.
- Pullback — price returning to the breakout level (the pattern's trough) within 30 days of a downside breakout. Common; it gives a second entry chance, but patterns with a pullback statistically end up with worse results.
The key numbers for double tops:
| Metric | Value |
|---|---|
| Average decline after confirmation | 16-19% (depending on the Adam/Eve variant) |
| Busted double top — average rise after failure | 38% |
| Single bust — average rise | 54% |
| Share of double tops that bust | 36% |
Two takeaways from this table.
First: the double top is, at best, an average pattern. A decline of 16-19% looks weak next to H&S (~22%, and 29% in a bear market). On top of that, over a third of confirmed patterns break upward. The M pattern is popular because it's easy to spot, not because it's an outstanding performer.
Second: the asymmetry between a "successful" pattern and a busted one is striking. A successful double top delivers a decline of a few tens of percent, on average. A busted one delivers an average 38% rise — 54% in the single-bust case. The market pays more for a bear trap than for the bearish pattern itself.
The usual caveat: data from US stocks, daily timeframe, mostly a bull market (which inflates upward-move statistics and softens the downward ones). On BTC/ETH the mechanics are the same; the numbers haven't been measured on a comparable sample.
How to Trade a Double Top
Entry. Short on a candle close below the trough between the peaks. A variant with a better R:R: enter on the pullback — price returns to the broken trough from below and gets rejected. Don't enter right at the second peak "because it'll probably fail" — that's betting against the statistic, which says a large share of these setups never confirm.
Stop loss. Classically above the second peak. If that's too far away (the pattern is very tall), it's better to skip the setup than to place a stop "in mid-air," in the middle of the pattern, where ordinary noise will take it out.
Target — the measure rule. Pattern height = the highest peak minus the trough between the peaks. Subtract it from the breakout level (the trough). Bulkowski recommends being realistic: multiply the height by the target-achievement rate for the given variant, since only some patterns reach the full range. Treat any support along the way as the first target.
Management. Given the 36% risk of the pattern busting, quick partial profit-taking and trailing the stop make sense — a double top isn't a pattern you "hold blindly to target."
A numerical example. A double top on BTC: two peaks at $110,000, a trough between them at $100,000, confirmation on a close below $100,000. Height = $10,000, full target $90,000; a realistic variant (say, × ~70% hit rate) gives about $93,000. Stop above the second peak, $111,000 — risk of about $11,000 against a potential $7,000-10,000. A sub-1:1 R:R is a common problem with tall M patterns: either find a tighter technical stop location (the local high of the pullback), or the setup simply doesn't deserve a position. This is an illustration of the mechanics, not a signal.
Volume as a filter. The pattern's classic signature is weaker volume on the second peak — demand didn't deliver on the second attempt. If the second peak forms on HIGHER volume and price stalls just under resistance, you're more often looking at accumulation ahead of an upside breakout than distribution. These are exactly the cases that later end up in the busted statistics.
Illustrative crypto example: BTC's peaks from April and November 2021 are sometimes described as a giant double top — the second peak marginally higher, confirmation on a close below the July trough. It played out by the book, but remember: that's one case, and the statistics above come from thousands of cases on a different market.
Busted Double Top — When the Trap Becomes the Setup
This is the most important section of this piece. Bulkowski's definition: a double top is busted when, after confirmation, price drops less than 10%, reverses, and closes above the pattern's highest peak.
The numbers once more, because they're worth it: 36% of double tops bust. Of those busted, 67% bust only once (single bust) — and that's the gold: the average rise after a single bust is 54%. The average for all busted patterns is +38%. The rest are double busts (5%) and multiple busts (28%) — price whipsaws through the pattern both ways, and it's usually best to walk away.
Where does this strength come from? The trap mechanics: below the pattern's trough sat long stops and breakout shorts. When the decline dies before -10% and price returns above the peaks, the shorts have to buy back, while the broken resistance is left behind. Fuel from two directions.
How to trade it: a buy stop just above the pattern's highest peak (Bulkowski's own example — CONN — describes it as "a penny above the top"). Stop loss below the pattern's trough, or, tighter, below the trough that price reversed from. From Bulkowski's pattern-pairs research: entries work best when the breakout happens above the 200-day moving average and when the trend leading into the pattern was short (up to 3 months).
A myth to bust before we close: "a failed pattern means a failed trader." No. A failed pattern is information — and in the case of a double top, statistically more valuable information than the pattern itself. If your short got stopped out because price closed above the peaks, you just watched a long signal with a bigger average reach than the move you were originally hunting.
FAQ
When is a double top confirmed? When price closes below the trough between the peaks. Without that, two peaks are just scribbles on the chart — a large share of such setups end with an upside breakout instead.
What is a busted double top? A pattern that, after confirmation, dropped less than 10%, then reversed and closed above the peaks. The average move up after a bust: 38%, and 54% for a single bust. It's traded with a buy stop above the pattern's peak.
Do both peaks need to be at an identical level? No — the average difference is about 1%, and deviations are normal. What matters more is a clear trough between the peaks (about 10% of correction) and confirmation on a close below it.
FAQ
When is a double top confirmed?
What is a busted double top?
Do both peaks have to sit at the exact same level?
Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.
🎁 Grab Strefa’s free TradingView indicators
Drop your email — we’ll send you links to our free TradingView indicators plus a no-fluff starter kit. Zero spam.
You’re joining the Strefa Tradingu list. Unsubscribe with one click, anytime.Check your inbox (and the Spam/Promotions folders) and add us to your contacts.