Double Bottom pattern

What is a Double Bottom?

A double bottom is a two-trough reversal pattern that can mark the end of a decline. Price drops to a low, bounces, returns to roughly the same level and finds buyers again. The peak of the bounce between the two lows is the neckline, and the bullish case is confirmed when price breaks above it.

The logic is the reverse of a double top. The first low is where buyers first absorb the selling. The second decline tests whether that demand is still there; when price holds near the same level instead of making a meaningfully lower low, sellers have shown they cannot push further. Short sellers who entered on the way down start to cover, and buyers who missed the first low get a second chance. Clearing the neckline means the market has made a higher high relative to the bounce, the first building block of a new uptrend.

Anatomy of the Double Bottom

• Prior decline: the downtrend the pattern may end.
• First low: a swing low where buying first overwhelms selling.
• Middle peak: the rebound high between the lows. Its level is the neckline, a horizontal resistance line.
• Second low: a return to about the same price. It can sit slightly above the first (a higher low, showing buyers stepping in earlier) or slightly below it (an undercut that runs stops under the first low before reversing).
• Breakout: a close above the middle peak.
• Height: the distance from the lows to the neckline, the basis of the measured-move target.

The rounder and more deliberate the second low, the more it looks like accumulation; a second low made by one sharp spike says less. Time between the lows matters as well: lows a few bars apart are really one choppy low, while lows many months apart may belong to different market phases rather than a single turning point.

How to identify a Double Bottom

A practical checklist. It reflects common practice; PatternX's own limits are listed in the detection section.

1. Confirm that a decline led into the first low.
2. Mark the first low and the rebound high that follows it.
3. Look for a second decline that stops near the first low, neither far below it (a new downtrend leg) nor far above it.
4. Check that the two lows are separated by enough bars to count as separate tests.
5. Confirm that the rebound between the lows was meaningful relative to price, not a small wiggle.
6. Draw a horizontal neckline at the rebound high.
7. Watch how price rallies from the second low: steady progress with rising participation is the classic picture, while a listless drift is not.
8. Wait for a close above the neckline.
9. Calculate the measured move (height added to the breakout price) and decide where the stop goes: under the lows, or inside the pattern.
10. Check for overhead resistance between entry and target, such as an old breakdown level or a falling moving average.

Confirmation

The double bottom is confirmed when price closes above the neckline. Traders look for the breakout to come with rising volume, for the next bars to hold above the old rebound high and, often, for a pullback that retests the neckline from above and holds as support. A break that immediately falls back into the pattern is not confirmation.

Waiting matters because buying the second low is a guess about where the decline stops, and the second low can easily become a third, lower one. The neckline break is the point where the evidence changes from a hopeful bounce to a higher high.

PatternX's trigger is the first close above the neckline on the newest bar, which may not have finished forming, so its signal can arrive before that bar is complete. See breakout for how breakouts are judged more generally.

Invalidation

Before the breakout, a decisive move below both lows invalidates the pattern: the support failed, and the decline continues. After the breakout, a drop back below the neckline is a warning, and a fall deep into the W, toward the lows, signals failure.

Once a double bottom signal fires or a trade opens, PatternX does not look at the pattern again. The pattern's stop is the neckline minus 55% of the height, about halfway down the W, and that becomes the stop-loss unless your own setting is tighter. It is an inside stop: you can be stopped out while the lows themselves still hold, which is the price of a closer stop and a better reward:risk.

Volume

Volume in a textbook double bottom: heavy selling into the first low, noticeably lighter volume on the second decline as sellers run out, then rising volume on the rally from the second low and through the neckline. Bottoms are generally thought to need active buying to complete, so a breakout on thin volume is viewed with more suspicion than a breakdown on thin volume.

PatternX keeps volume simple. It has no double-bottom-specific rule; the only check is universal: the breakout bar needs volume of at least 1.2× its 20-bar average. Because the latest bar may still be forming, the volume counted is whatever has traded so far in that bar. Volume is also one of the 13 factors in the PatternX Score. See relative volume.

Common false signals

Typical failure modes for the double bottom:

• Catching the second low. Buying before the neckline break assumes the low holds; often it does not.
• A W inside a downtrend. In a strong bear trend, a double bottom can complete, push through the neckline briefly and then resume falling. Check the higher timeframe.
• The undercut that keeps going. A second low slightly below the first can be a stop run that reverses, or the start of a new leg; you only know afterwards.
• Thin breakouts. Low-volume pushes above the neckline, especially in illiquid names, are easy to reverse.
• Resistance right above. A moving average or old support-turned-resistance just over the neckline can cap the move.

Reduce them by waiting for the close above the neckline, requiring participation on the break and checking for obstacles between entry and target. A break that fails quickly is a false breakout, not a reason to add.

Market context

A double bottom is most meaningful after a sustained decline into a level that matters on a higher timeframe: a prior breakout level, a long-term moving average or a zone where price has reversed before. It is stronger when the broad market or sector is stabilising too, and when momentum diverges, for example RSI making a higher low while price makes an equal low.

It is less meaningful as a minor pause within a range, or ahead of a known catalyst that can override the chart. No prior-downtrend test exists in PatternX's double bottom detector; rectangles are the only patterns with one. The PatternX Score's trend-alignment and multi-timeframe factors favour agreement with the trade's direction, which a fresh bottom may not have yet.

In crypto markets

Crypto double bottoms commonly form after liquidation cascades: the first low is a forced-selling flush, the second a calmer retest. Wicks below the first low are frequent, and PatternX accepts lows that differ by up to 2.5%, so a modest undercut can still count; a very deep wick will not.

On perpetual futures, deeply negative funding at the second low means shorts are paying to stay short, a crowded position that can accelerate a breakout as they cover. Exchange-specific liquidity matters as well: a thin venue can print a second low that a larger venue never saw, and PatternX evaluates each exchange's chart separately. Crypto signals also need at least $3M of 24-hour volume. See crypto chart pattern recognition and funding rate.

In stocks, ETFs and FX

In stocks, double bottoms often form around earnings: the first low on a disappointing report, the second as the market tests whether the bad news is fully priced. Overnight gaps can carry price straight through the neckline at the open, leaving a worse entry and a larger distance to any stop. For small US margin accounts trading intraday, the pattern day trader rule matters: PatternX's guard, on by default, blocks a fourth same-day round trip within five business days when equity is under $25,000 (live securities trades on US margin brokers).

For ETFs, the bottom reflects the underlying basket; for bitcoin and ether ETFs, it reflects crypto prices that kept moving while the ETF was closed. In FX, double bottoms at round numbers are common, and the London and New York overlap often supplies the breakout.

Timeframes

Short timeframes produce many small W shapes, most of them noise around intraday support; daily double bottoms take weeks and are watched by many more participants, so their necklines tend to matter more.

PatternX scans up to eight timeframes, 1m through 1d. Its double bottom detector reads the latest 90 bars and needs the two lows to be 8 to 50 bars apart, with the second low no more than 25 bars before the breakout. On a 5-minute chart, that is lows 40 minutes to about four hours apart; on a daily crypto chart, 8 to 50 days. A bottom whose lows are three months apart will not be detected, since 1d is the highest timeframe PatternX scans. Swing lows are confirmed three bars after they print.

Example: the logic step by step

For example, a hypothetical ETF on a 1-hour chart:

1. The lows differ by about 0.6%, and the neckline is about 4.3% above their average.
2. Height = 50.10 − 48.05 = 2.05, so the target is 50.30 + 2.05 = 52.35.
3. Pattern stop = 50.10 − 0.55 × 2.05 ≈ 48.97, or 48.87 after the 0.2% buffer.
4. Risk 1.43 to make 2.05: reward:risk ≈ 1.4.

The default 10% take-profit (55.33) and 5% stop (47.79) are both wider, so the pattern levels apply. With a 2% stop-loss setting, the stop would tighten to 49.29 and reward:risk would rise to about 2.0, at the cost of being stopped by smaller dips. Position size is your risk per trade (1% of balance by default) divided by the stop distance.

How PatternX detects the Double Bottom

PatternX detects double bottoms with the same deterministic logic as double tops, mirrored onto lows, over the most recent 90 bars:

1. Swing lows need three bars on each side that do not trade below them, so the detector learns of each low three bars after it prints.
2. A pair of consecutive swing lows (the newest pair is tried first) that sit within 2.5% of each other and 8 to 50 bars apart.
3. The neckline is the highest swing high between the two lows, and it must be at least 2% above the lows' average.
4. The second low must have formed within the last 25 bars.
5. Breakout: the latest close is the first one above the neckline (the previous close was not above it). That bar may still be forming.
6. Levels: a stop 55% of the height below the neckline, and a target one height above the entry, where the height is the neckline minus the lows' average.

Details: pattern detection.

How PatternX Classic evaluates it

In PatternX Classic, on every plan, the double bottom becomes a candidate signal scored by the 13-factor PatternX Score. Pattern geometry is worth 16% (the detector caps it at 85) and a fixed reliability prior for double bottoms 8% (a set input, not a measured result); the remaining factors measure breakout strength, volume, trend alignment, multi-timeframe agreement, liquidity, spread, order book, volatility, funding, open interest and the live win rate PatternX has recorded on that market.

That 0–100 total is the signal's confidence, banded Elite from 80, High from 65, Medium from 50 and Weak below, and it must reach your minimum (65 unless you change it). You get the nearer target and tighter stop of the pattern's levels and your own, nothing below 1:1 reward:risk survives, and in manual mode you have 15 minutes to approve. Before any order, the risk engine checks your limits, open trades and daily loss.

What PatternX AI adds

On Pro and Elite, PatternX AI adds a context check after the deterministic engine has accepted the double bottom; it does not detect, alter or invent patterns. Its fixed formula over eight factors asks whether the market supports a long: does RSI(14) favour it, do the three latest swing highs and lows of the past 60 candles agree, where does EMA12 sit against EMA26, is the latest bar's volume above its 20-bar average, and do these agree (confluence)? On perpetual futures a sentiment factor reads funding; for stocks and FX it stays neutral.

The result can veto the setup or re-weight the confidence shown. It never moves the stop or target or changes position size, and if the evaluation fails the setup is treated as Classic. See the AI context layer.