Is a pennant pattern bullish or bearish?

It can be either. A bullish pennant follows an upward pole and seeks an upper-boundary break. A bearish pennant follows a downward pole and seeks a lower-boundary break. The triangle shape alone does not decide direction.

What is the difference between a pennant and a triangle?

A pennant adds a sharp preceding impulse and a brief pause to the converging shape. An ordinary triangle need not meet that flagpole requirement. If the prior move is weak or the consolidation outlasts your chosen limit, keep it outside the pennant sample.

Is a flag pennant pattern one formation?

Flag and pennant patterns are related continuation formations with different boundaries. Flags use roughly parallel lines; pennants converge. They can share an impulse requirement, but the same boundary test should not label both shapes indiscriminately.

Does a pennant breakout have to reach the pole-length target?

No. Adding or subtracting the pole length produces a reference level, not an obligation for price to reach it. A confirmed break can reverse before that level, and a gap can change the entry price and planned risk.

What is the best timeframe for pennant trading?

No best timeframe is established here. Fix the instrument, session, price source, duration rule and costs before comparing timeframes. The hypothetical examples demonstrate classification and timing, not profitable settings.

TradingView chart patterns

Pennant pattern: a flagpole, a pause and a breakout

A pennant pattern is a short, converging consolidation after a sharp price move called the flagpole. A bullish pennant follows an advance; a bearish pennant follows a decline. The continuation setup needs a breakout in the direction of that move. A small triangle without the preceding impulse is not enough, and even a confirmed breakout can fail.

See the decision guide

Bull pennant, bear pennant, flag or triangle?

The pennant chart pattern combines context with shape. StockCharts describes a brief triangle after an impulse, while a flag has approximately parallel boundaries. TradingView treats its ordinary Triangle pattern as able to break either way. Calling every converging range a pennant removes the flagpole condition that distinguishes the continuation hypothesis.

Bullish pennant pattern and bull pennant pattern name the same setup. Bearish pennant pattern and bear pennant pattern are also synonyms.
A bullish candidate breaking downward fails its continuation condition. Record that event before considering a separate bearish setup.

Formation	Before the pause	Boundary and confirmation
Bull pennant pattern	Sharp advance	Converging range; bullish confirmation requires an upper-boundary break
Bear pennant pattern	Sharp decline	Converging range; bearish confirmation requires a lower-boundary break
Flag	Sharp advance or decline	Roughly parallel channel, usually tilted against the impulse
Ordinary symmetrical triangle	No flagpole prerequisite in this comparison	Converging boundaries; breakout direction remains unresolved

Make the flagpole and short pause measurable

The following review rules are illustrative choices, not universal settings or a copy of the TradingView detector. Mark the pole start S and endpoint E first. Use positive prices, ordinary candles and one fixed timeframe. Let H be the absolute price difference between S and E.

Choose the swing-selection method before automating. Record which confirmed highs and lows anchor each line, and freeze those anchors for the candidate. Two visually convenient points picked after a breakout introduce hindsight.
Measure the upper and lower line at each bar. Their prices change with time even when the anchors stay fixed. Expire the candidate at the apex if it arrives before the maximum duration.
A 15-bar limit on an hourly chart is different from 15 daily bars. StockCharts discusses duration in weeks; that convention does not establish a best intraday setting.
Volume can be an additional filter. Specify the comparison window and what to do when volume is unavailable; do not silently count missing observations as confirmation.

Decision	Illustrative requirement	Reason to reject
Pole strength and speed	H / S is at least 8%, with 3 to 10 bars from S to E	The move is too small or too slow for this chosen impulse rule
Pause duration	Allow a breakout from the fifth through fifteenth bar after E	An earlier break lacks the required pause; a later one is expired
Retracement	Before breakout, movement against the pole is no more than 40% of H	The consolidation gives back too much of the selected impulse
Convergence	Upper boundary U slopes down, lower boundary L slopes up, and U stays above L	Parallel lines describe a flag; crossed lines mean the apex has passed
Confirmation	After the boundaries are known, require a completed close outside them in the pole direction	A wick outside followed by a close inside does not qualify

A hypothetical bullish breakout and its bearish mirror

Assume a pole runs from 100 to 112 in six bars. Its height is 12 and its percentage move is 12%, so it passes the illustrative pole rule. Let t count bars after the endpoint. The saved lines are U(t) = 112 - 0.25t and L(t) = 108 + 0.15t. Assume their anchors are confirmed by t = 8, all earlier consolidation closes stayed inside, and the deepest pullback was to 108.

For a bearish mirror, let the pole fall from 112 to 100. Use U(t) = 104 - 0.15t and L(t) = 100 + 0.25t. At t = 9, a close of 101.8 is below L = 102.25 and confirms the downward example under the same timing assumptions.
A pole-length projection from the boundary at the breakout bar gives 109.75 + 12 = 121.75 for the bull example, or 102.25 - 12 = 90.25 for the bear example. These are geometric reference levels, not forecasts or assured exit prices.
Define failure separately. For this review, a bull signal fails if any of the next three closes is below the saved breakout-boundary price of 109.75; the bear signal fails if a close exceeds 102.25. This custom label is not the built-in TradingView Failed status or a protective stop.

Observation	Calculation	Interpretation
Pullback depth	(112 - 108) / 12 = 33.3%	Within the illustrative 40% limit
At t = 8	U = 110, L = 109.2, close = 109.6	Candidate remains inside a narrowing range
At t = 9	U = 109.75, L = 109.35, close = 110.2	First eligible close above U confirms this example
Alternative at t = 9	High = 110.2, but close = 109.5	Wick-only break; no confirmation
At t = 10 without a prior breakout	U = L = 109.5	Expire at the apex rather than extend an inverted triangle

TradingView confirmation has a timestamp

TradingView documents 5/5 pivots for its bullish and bearish Pennant indicators. Five bars to the right must exist before such a swing can be confirmed. In Progress can include unconfirmed final points. The documented Pattern Breakout alert uses a completed bar outside the pennant in the target direction; an opposite break does not trigger that alert.

Keep the pivot bar, pivot-confirmation bar and breakout bar distinct. A marker drawn on an earlier swing does not mean a tradable signal existed then.
In a custom detector, decide what happens when price breaks before the last required pivot is confirmed. Reject it, or label it as a separate provisional event; never backdate an entry once later candles reveal the shape.
Closing-bar checks reduce intrabar signal changes, but they do not guarantee that every source of repainting disappears. Source-data revisions and historical plotting choices still need review.
Compare the event log with bar-by-bar replay. Include candidates that expired or broke the wrong way so the chart does not retain only successful-looking pennants.

Test the cost of false continuations

A pennant trading pattern becomes a strategy only after entry timing, exits, position size and costs are defined. Distinguish a shape that fails your written rules from a correctly detected pattern whose trade loses. Optimizing profit cannot tell you which of those two problems caused a result.

Use a fixed review sample to count rule violations, expired candidates and confirmed breakouts. Record the false-continuation definition before measuring it, such as the three-close test above.
Keep the trade model fixed while comparing pole-strength, retracement and duration inputs. Otherwise an apparent improvement may come from changed exits or position size.
A signal at the close is not proof of a fill at that close. TradingView strategies normally fill a newly created market order at the next available tick, often the next historical bar open. Include commissions, slippage and any relevant short-position costs.
Give Pineify the full selection and timing rules to draft the code, then request syntax validation and inspect the events. For an implemented strategy, Strategy Optimizer can compare input combinations and backtest results. Check a separate period before interpreting the selected settings.

Where Pineify fits

Pine Script AI Coding Agent

Draft a pennant detector from your pole, convergence, duration and confirmation rules. Inspect the code and request a syntax check before testing. This page explains the pattern; it does not scan live markets.

Also useful: Strategy Optimizer. Compare input combinations after adding explicit trade rules to your strategy. Review detection errors separately; a better historical result does not prove better pattern recognition.

Draft pennant rules

Related resources

TradingView chart patterns Double top: a reversal with a neckline Bar-close alert frequency Backtesting accuracy and execution assumptions

Frequently asked questions

Educational information, not investment advice. All prices, equations and thresholds are hypothetical. No success rate or trading performance is reported. Pattern confirmation and geometric targets do not guarantee continuation, a fill price or profit.

Sources and verification

Last verified: 2026-10-16

Published by Pineify. Pineify is an independent product and is not affiliated with TradingView. Sources support the technical explanations; examples do not establish trading performance.