Bull Trap: Anatomy of a False Breakout, and How to Measure One

A bull trap is a false breakout. A stock that has been ranging below — or rallying into — a resistance level briefly trades above it, convincing buyers that a new up-leg has begun. They buy the breakout; short sellers cover. Then, instead of following through, price reverses back below the level. The breakout buyers are now underwater on a position they entered at the highs, and the stops they place just under the breakout become fuel for the decline. The trap is the reversal that converts a bullish signal into a loss.

Put in one line, the bull trap meaning is this: a breakout that fails and reverses fast enough to punish the traders it recruited. It is the bullish mirror image of a bear trap — a false breakdown below support that reverses upward and squeezes the sellers it recruited. Bull trap versus bear trap is a question of which side gets recruited and then reversed on: the trap always springs on the crowd that acted on the level.

The pattern is costly precisely because breakouts are among the most widely traded signals in the market. The more participants act on a level, the more concentrated the resting stops just beyond it, and the more violent the unwind when the level fails to hold. A bull trap is therefore a statement about the quality of demand, not the direction of price: the tape can be higher and the move can still be low-grade, carried by momentum and covering flow that has no reason to defend the level once it stops working.

The setup usually pairs a level a lot of orders sit on with a burst of attention that pulls in fresh chasers. The level is a prior high, a round number, or the top of a multi-week base — a place where both resting orders and trader eyes cluster. The attention is a catalyst (an earnings beat, a clinical readout, an upgrade) or simply a strong momentum run that recruits trend-followers and fear-of-missing-out.

As price pushes through the level, three flows fire at once: breakout traders buy the move, short sellers are squeezed into covering, and systematic momentum strategies add exposure. That cluster of demand can carry price well above the level in a hurry. The problem is that all three are, in effect, the same trade with the same trigger, and none of them is a patient holder accumulating because the asset is cheap. When the trigger is spent, there is no natural bid underneath — the buyers who would defend value are not the ones who just transacted at the highs.

The behavioral engine is over-extrapolation: investors take a genuine piece of good news and project it too far, too fast, the long-documented tendency of markets to overreact to salient information (De Bondt and Thaler, 1985). Short-horizon return reversal — the empirical regularity that individual-stock returns exhibit negative serial correlation over days to weeks (Jegadeesh, 1990; Lehmann, 1990; Lo and MacKinlay, 1990) — is the statistical footprint of those overreactions correcting.

Decompose the demand at the breakout and the failure becomes mechanical rather than mysterious. Breakout entrants buy at market above the level, willing to pay up because the signal is the reason for the trade. Short sellers cover, and covering is inelastic demand — a squeezed short buys to stop a loss, not because the price is attractive; the response to price-contingent stop flow is empirically larger and more one-directional than to ordinary limit demand (Osler, 2005). Momentum and volatility-scaling systems re-gross into rising price. The common feature is that every one of these buyers is triggered by the move itself, so their demand is exhausted at roughly the moment the move completes.

What is missing is a valuation bid. Long-horizon holders who buy because the discounted cash flows justify the price do not chase a vertical; they sit below. So once the trigger-driven flow clears, the order book above the old resistance is thin, and the marginal seller — a breakout buyer now stopped out, or a systematic book de-grossing as realized volatility rises — meets little resistance on the way down. Uninformed, trigger-driven order flow moves price through temporary impact that decays once the flow stops (Kyle, 1985); the reversal is that decay made visible.

The stop placement itself is the accelerant. Breakout buyers cluster their protective stops just under the breakout level, which is exactly where a modest pullback will trip them. When price re-enters the range, those stops fire as market sells into an already-thin book, and the selling begets more selling — the positive-feedback cascade that stop clustering produces (Osler, 2005). This is why a bull trap so often reverses faster than it rose: the flow that built it was one-directional demand, and the flow that unwinds it is one-directional supply landing on the same clustered level.

The single most useful discriminator is volume. A genuine breakout is an expansion event — price clears the level on rising volume as new demand overwhelms the supply parked at resistance. A bull trap typically makes its high on the same or lower volume than the preceding advance, because it is running on the fumes of covering and momentum rather than fresh accumulation. A new price high on falling volume is a non-confirmation, and it is the classic fingerprint of a trap. The volume-price relation is not folklore: trading volume and the magnitude of price change are positively related across a large literature (Karpoff, 1987), so a high made on shrinking volume is genuinely anomalous, not merely aesthetically weak.

The second tell is momentum divergence. When price prints a higher high but a momentum oscillator prints a lower high, the rate of change is decaying beneath the surface even as the headline price rises. Divergence does not time the reversal; it flags that the move is being driven by fewer and weaker hands.

The third is candle structure and where price closes relative to the breakout: wide-range climax bars, long upper wicks, and — most importantly — a failure to hold above the level on subsequent sessions all argue the level was tested and rejected rather than reclaimed. The operational definition of the trap springing is a decisive close back below the breakout shelf. Bull trap versus breakout is decided here, in the confirmation window, not at the instant price first clears the line.

The tells are only useful if they are specified precisely enough to compute the same way every time. Volume non-confirmation: form the ratio of breakout-session volume to the average volume of the advance that produced it (say the trailing five sessions). A ratio below roughly 1.0 on the new-high bar — the high made on average or lighter volume than the run into it — is the non-confirmation flag. In the Merck case the decisive 25 November high traded ~22.2M shares against 28.3M on 24 November and 23.2M on 21 November: a ratio well under 1, a lower high in participation against a higher high in price.

Momentum divergence: RSI is the Wilder relative-strength index, 100 - 100/(1 + RS), where RS is the ratio of average up-closes to average down-closes over a lookback (14 sessions by convention). MACD is the difference between the 12- and 26-period exponential moving averages, read against its 9-period signal line. The testable condition is a bearish divergence — price posts a higher swing high while the oscillator posts a lower swing high — evaluated on the same two swing points, not eyeballed.

Range and retest: measure range expansion against Average True Range, the mean over 14 sessions of the greater of the high-low range and the gaps to the prior close. A blow-off bar often spans two or more ATRs, the signature of a climax rather than accumulation. The retest rule then converts the pattern into a falsifiable line: mark the breakout level, require price to hold above it on a close-basis retest to confirm, and treat a decisive close back below it as the trap confirmed. For Merck the relevant shelf was the $100 round number over the ~$95-97 base; the 5 December close of $99.72 lost it and confirmed the reversal, en route to the 9 December trough of $96.89.

The two are indistinguishable at the instant of the breakout; they separate in the confirmation. A real breakout (i) expands on volume, (ii) holds above the level on a retest, and (iii) follows through within a few sessions. A bull trap (i) makes its high on flat-to-declining volume, (ii) fails to hold the level, often alongside momentum divergence, and (iii) reverses back through the breakout and triggers the stops parked beneath it.

This is why disciplined breakout traders wait for confirmation — a volume-backed close above the level, or a retest that holds — rather than buying the first print above resistance. The cost of that patience is missing the very start of a real move; the benefit is sidestepping the traps, which on a weakly-confirmed breakout into an over-extended, news-driven run are the base-rate outcome over the following one to two weeks.

Two literatures bear on a bull-trap fade. The first is short-horizon reversal. Individual-stock returns display negative serial correlation at daily-to-weekly horizons, and a meaningful share of the resulting contrarian profit reflects genuine overreaction rather than microstructure artifact or cross-autocorrelation (Jegadeesh, 1990; Lehmann, 1990; Lo and MacKinlay, 1990; De Bondt and Thaler, 1985). The effect is horizon-specific — it flips to positive autocorrelation, momentum, at three-to-twelve months (which is exactly why a trap fade is set to days, not months). A vertical, news-extrapolating advance that clears a level on non-confirming volume is the configuration the reversal literature says is most likely to partially retrace once the originating flow is spent.

The second is the evidence on technical levels themselves. Support-and-resistance and trading-range-break rules are among the specific families that survive formal testing: Brock, Lakonishok, and LeBaron (1992) found that moving-average and range-break signals on the Dow carried return patterns inconsistent with random-walk, AR(1), and GARCH null models, and Lo, Mamaysky, and Wang (2000), automating pattern recognition with kernel smoothing, found several chart formations carry incremental information over the 1962-1996 cross-section. The evidence is real but conditional. The most careful survey — Park and Irwin (2007), reviewing 95 modern studies, of which 56 were positive, 20 negative, and 19 mixed — concludes that documented profitability is regime- and cost-dependent and has attenuated over time. A level-and-volume signal has edge in the right configuration; it is not a mechanical printing press.

Why the edge is not immediately arbitraged away is itself expected. Arbitrage is bounded by capital and risk limits, so transparent dislocations can persist for non-trivial intervals before converging (Shleifer and Vishny, 1997). A trap fade monetizes that bounded convergence over a defined window; it does not assert the dislocation was impossible.

Through early November 2025, Merck was basing in the mid-$80s. Then it ran, hard. From an $86.75 close on 10 November the stock advanced to a $105.66 close on 25 November — roughly +22% in two weeks — clearing the prior-high shelf around $95-97 and then the $100 round number.

The catalyst was real and substantive. Merck's enlicitide decanoate — an investigational oral PCSK9 inhibitor for LDL-cholesterol lowering — had met its endpoints in the CORALreef Phase 3 program, with detailed cardiovascular data presented at the American Heart Association meeting in mid-November and an investor event highlighting the results; the FDA subsequently granted a priority-review voucher that could compress its assessment timeline. Strategically, enlicitide is what the bull case wanted: a large, non-oncology growth driver to help offset the looming 2028 loss of exclusivity on Keytruda.

All of that is legitimate. What made the late-November tape a bull trap was not the news but the way price discounted it: a 2028-and-beyond revenue option was extrapolated into an immediate, near-vertical re-rate, and the breakout failed its confirmation. The decisive new-high session on 25 November printed on about 22.2 million shares — below the 28.3 million of 24 November and the 23.2 million of 21 November. A new high on fewer buyers: non-confirmation, and by the measurement rule above, a breakout-to-advance volume ratio under 1.

The output is a retrospective SELL classification of the 25 November close of $105.66, written in June 2026, with an estimated return band of -4.2% to -7.6% over roughly six to twelve trading sessions. Each parameter follows from the thesis and from the chart structure at the 25 November close. The entry is the non-confirming new-high close itself — the point at which the volume and divergence tells were both present and the over-extension was maximal. The horizon is short by construction: the reversal signal the trade harvests is a days-scale, negative-autocorrelation phenomenon that decays and eventually inverts to momentum, so holding period is a parameter of the thesis, not an afterthought.

The band is anchored to structure, not to a round guess. The natural downside objective for a failed breakout is a round-trip back through the levels it cleared — here the $100 shelf and the ~$95-97 base — which from a $105.66 entry maps to a mid-single-digit-percent decline, consistent with the historical dispersion of short-horizon reversals after comparable over-extensions. The band's lower bound sits at the base; its upper bound at the first shelf.

The invalidation level this classification sets is falsifiable: a volume-confirmed reclaim of the $105.84 intraday high — which would signal that information-based demand, not covering flow, had taken control — or a fresh, discrete fundamental catalyst that would justify a step-change in fair value. Either would falsify the premise of an information-thin, mean-reverting fade and downgrade the classification to monitoring. Neither occurred inside the window.

The trap sprang inside the classification's six-to-twelve-session horizon. Price stalled for two sessions, then lost the $100 shelf — closing at $99.72 on 5 December — and continued to a window trough of $96.89 on 9 December, down 8.3% from the classification close at roughly session ten. The breakout buyers from the $100-105 zone were, by definition, underwater, and the round-number break added stop-driven supply into a thin book. The realized decline passed through the classification's -4.2% to -7.6% band and ran beyond its lower bound.

The sequence matched the mechanism section point for point: a level cleared on a genuine catalyst, a trigger-driven burst of breakout and covering demand, a new high on non-confirming volume, and a reversion back through the level as the burst exhausted, accelerated by clustered stops firing at the $100 line — with no valuation bid appearing to defend the highs.

The invalidation is the position's spine: it converts a directional opinion into a falsifiable line. Because the classification defines the exit in price, time, and evidence alongside the entry, the trade cannot silently mutate from a short-horizon statistical fade into an open-ended fundamental short. Position sizing follows from the invalidation distance — the gap from the $105.66 entry to the $105.84-reclaim line is small, so a stop there is tight and the size that risks a fixed fraction of capital is correspondingly larger; risk is budgeted to the distance to being wrong, not to conviction.

What would have flipped this from a trap into a real breakout is the same test in reverse. A volume-confirmed hold above $105.84 — a higher high in price matched by a higher high in participation, ideally with a successful close-basis retest of the breakout — would have said the flow was information-based and the up-leg genuine. That is the falsifier, and none of it printed inside the window; the high was made on shrinking volume and the level did not hold.

The final discipline is the time-stop. Held past the window in which negative autocorrelation operates, a reversal signal decays toward noise and the dominant driver becomes the next scheduled catalyst — on which this classification takes no differentiated view. Consistent with that, Merck reclaimed the high-$90s and pushed back toward its highs later in December as the broader uptrend resumed. The classification does not claim that move: it falls outside the six-to-twelve-session horizon and reflects the trend reasserting after the over-extension had been worked off. The edge was in the over-extension, not in a view on Merck's pipeline; good news and a bad entry are not mutually exclusive, and the classification's time-stop, which ends the fade within six to twelve sessions and ahead of the trend's resumption, exists for exactly that reason.

Bull trap — a false breakout above resistance that reverses, trapping the buyers who chased it.

Bear trap — the mirror image: a false breakdown below support that reverses upward, trapping short sellers.

Resistance / breakout level — a price (prior high, round number, base top) where supply has historically capped advances; clearing it is a breakout.

Volume non-confirmation — a new price high made on flat-to-declining volume; here, a breakout-bar volume below the average of the advance that produced it.

Momentum divergence — price makes a higher swing high while a momentum oscillator (RSI, MACD) makes a lower swing high; the move is weakening internally.

RSI — Wilder's relative-strength index, 100 - 100/(1 + RS), RS = average up-close / average down-close over a lookback (14 by convention).

MACD — the 12-period minus 26-period exponential moving average, read against a 9-period signal line.

ATR — Average True Range; the mean over 14 sessions of the greater of the session range and the gaps to the prior close; a scale for range expansion.

Over-extrapolation / overreaction — projecting a piece of news too far into price; empirically prone to subsequent reversal.

Short-horizon reversal — the tendency of individual-stock returns to exhibit negative serial correlation over days to weeks.

Stop cascade — clustered protective stops firing in sequence, each sale pushing price into the next, a positive-feedback source of accelerated reversal.