Bull flag pattern: the shape, its failures and how to count them
A bull flag describes a shape, not an outcome. Define the pole, the flag and the break so they can be counted, then look at the break that continues, the one that fails and the one that goes sideways, after costs.
Start in plain English
The bull flag is one of the most searched chart patterns among day traders. This lesson treats it the way the course treats every setup: as a situation with several outcomes, defined precisely enough to count. For a small-cap scalper the pattern appears many times a day on 1-minute charts, so the difference between a remembered flag and a counted one is large.
The shape has two parts:
- The pole: a fast move up over a few bars.
- The flag: a short, shallow drift sideways or down, usually on lighter volume, that stays in the upper part of the pole.
The common story says the next move continues up. Sometimes it does. Sometimes the break fails at once, and often price goes sideways until a time stop ends the trade. A bull flag describes what already happened on the chart. What happens next is a distribution you can only see by counting.
How to read candles belongs to the How to read candlestick charts lesson in US Day Trading Fundamentals.
Two paths from the same start
All prices, sizes and costs in this example are Illustrative. Marlpit Aerospace is fictional.
The setup at 10:08 ET. Marlpit gapped up on news and traded quietly from 09:50, averaging about 150,000 shares a minute. From 10:00 to 10:03 it rose from $2.40 to $2.68, four bars with higher closes on 300,000 to 510,000 shares each: the pole. From 10:04 to 10:08 it drifted between $2.57 and $2.65 on 150,000 to 240,000 shares: the flag. Quote at 10:08:59: $2.62 × $2.63.
The plan, written before 10:08. Trigger: a trade above the flag's high of $2.65. Entry with a marketable limit (ceiling $2.68), filled at $2.67. Stop at $2.55, two cents under the flag low. Target $2.91 (2R). Time stop 10:19. Planned risk $0.12 per share; $50 risk = 416 shares. Costs $0.005 per share round trip (Illustrative), plus any slippage.
Path AThe flag breaks and continues
10:09 trades $2.66 and the buy fills at $2.67. Price clears the pole high, prints $2.80 at 10:12 and $2.92 at 10:15. The $2.91 limit sell fills.
Path BThe break fails within 2 minutes
10:09 trades $2.66 and the buy fills at $2.67, one cent under the pole high. Sellers at the pole high fill the bids; at 10:10 price drops through the flag low, the $2.55 stop triggers and fills at $2.53.
Path CThe flag keeps flagging
The buy fills at $2.67; price chops between $2.60 and $2.72 for ten minutes. At 10:19 the time stop sells at the bid, $2.65.
Review note: "Bull flag per definition v1; plan followed; outcome A, B or C." The chart up to 10:08 was identical in all three. A screenshot of Path A would show a textbook flag; screenshots of B and C rarely get posted. The rest of this lesson defines the pattern so it can be counted, looks at volume and measured-move claims, and sets out what you need to measure.
Bull flag defined so you can count it
Pattern books describe flags loosely, and loose definitions let a trader count the flags that worked and call the others "not real flags." Here is one countable definition. It is a choice, not a standard; change it before you look at outcomes, not after.
| Part | Definition used in this lesson |
|---|---|
| Pole | At least 3 consecutive 1-minute bars with higher closes, rising at least 8% from the first bar's open, each on volume above the prior 10-bar average |
| Flag | 3 to 8 bars after the pole high, with no high above the pole high and no close below the pole's 50% retracement |
| Flag volume | Average flag volume below average pole volume |
| Break (trigger) | First trade above the highest high of the flag bars |
| Invalidation | 2 cents under the flag low |
| Hold (Path A) | Target before stop or time stop |
| Fail (Path B) | Stop before target; "fast fail" if within 2 minutes |
| Chop (Path C) | Neither by the 10-minute time stop |
Marlpit's pole rose 11.7% ($2.40 to $2.68); the 50% retracement sits at $2.54, and the flag's lowest close was $2.59. Ambiguities remain: whether the 10:04 bar belongs to the pole or the flag, and whether a flag that dips below 50% for one bar still counts. Write the rule down; the count only means something if the rule stays fixed.
Volume: what it shows and what it does not
The usual description is heavy volume in the pole and light volume in the flag. That is an observation about the bars, and the definition above uses it. What it does not show is why: light flag volume fits buyers waiting, sellers waiting, or simply fewer participants. Heavy volume on the break fits both a continuation and sellers meeting buyers at the pole high, as in Path B. Record volume ratios as conditions; do not read intent into them.
Measured-move claims and why they are not targets
A common rule adds the pole's height to the breakout price to get a "measured move." For Marlpit: $0.28 pole + $2.65 break = $2.93. In Path A the high was $2.92, one cent short. A trader who used $2.93 as the target would have recorded Path C, not Path A, on the same chart. A measured move is a projection someone chose, not a level the market knows about. If you use one, treat it as a target choice and count its outcomes like any other.
Observe vs. infer
Back at 10:09 in Marlpit Aerospace (fictional; all figures Illustrative):
| What the screen shows (observed) | What a trader might infer | Why it is not established |
|---|---|---|
| Four pole bars with higher closes on 300,000–510,000 shares | "Strong buyers in control" | Volume and closes do not show who traded or why |
| Flag volume averaged about 180,000 shares, below the pole's | "Sellers are drying up" | Light volume fits waiting buyers, waiting sellers or a quiet tape |
| The flag held above the 50% retracement | "A healthy flag" | "Healthy" is judged by what happens next |
| Pole height $0.28 | "Target $2.93" | A projection, not a level the market knows about |
Odds box
Under these assumptions the plan breaks even at a Path A share of about 30% (computed from the assumed split, Illustrative). Two or three points in either direction change the sign. That is why a published bull flag percentage without a sample size, a definition and costs says little about your plan.
What shifts the odds
There is no FennEdge sample for flags, so the items below are hypotheses to record, not findings:
- Pole size and speed. 8% in 4 bars vs. 25% in 2 bars.
- Flag depth. A flag that holds the upper third of the pole vs. one near 50%.
- Flag length. 3 bars vs. 8 bars.
- Distance from the day's high and from the premarket high.
- Volume ratio of the break bar to the flag average.
- Spread against the target. See Is it liquid enough to scalp?.
- Time of day and recent halts. A pole that included an LULD pause is a different situation.
Cost of being wrong
Marlpit Aerospace plan, 416 shares (Illustrative):
| Outcome | Exit | R after costs | Dollars after costs |
|---|---|---|---|
| Path A: target | $2.91 | +1.96R | +$97.76 |
| Path B: fast fail, 2-cent slippage | $2.53 | −1.21R | −$60.32 |
| Path C: time stop | $2.65 | −0.21R | −$10.40 |
| LULD pause on the drop, reopening at $2.35 | $2.35 | −2.71R | −$135.20 |
The halt row is real for small caps: a fast drop can trigger a limit up-limit down pause, and a stop cannot execute during it.
How many trades before this means anything
Suppose you log 25 flags and 8 reach the target (32%). A 95% interval runs from about 17% to 52% (Wilson interval, computed on these hypothetical counts), which includes the 30% break-even and values well on either side. Several flags in the same stock on the same morning share one news event; count them, but weigh them less. How many trades before your numbers mean anything explains why.
What this lesson does not claim
- It does not claim bull flags continue, fail or carry any edge. The odds box is Illustrative.
- The definition is one workable choice, not an industry standard.
- Measured-move projections are target choices, not market levels.
- Marlpit Aerospace and its bars are fictional.
Go deeper
- FennEdge first.
- In-Play Days shows 1-minute charts of past active days. Without Max, including as a guest, you can open each ticker's most recent active day in full; earlier days and the two-year history need Max (as of 2026-10-05). Use it to find flags by your definition and record all of them, not only the clear ones.
- Top Gappers Calendar lists past gappers by day (guests, Free and Plus: latest session at default filters; full history Pro and Max).
- Bear flags. The mirror image: a fast drop, then a shallow drift up. Trading it short adds borrow costs, locates, short-sale restrictions and open-ended upside risk; US Day Trading Fundamentals covers short-selling basics, and FennEdge's SSR article covers the rule.
- Pennants. A flag whose highs and lows converge. Count it under its own definition rather than merging it into flags.
- Flags inside halts. A pole that ends in a volatility pause reopens through an auction; the flag after it starts from a different book.
- Counting patterns by computer. Lo, Mamaysky and Wang (2000) showed how to define chart patterns precisely enough for an algorithm to find them, testing them on NYSE/AMEX and Nasdaq stocks from 1962 to 1996, sampled at random from every size quintile, smallest to largest. Flags were not among their patterns, but the method is the same as this lesson's: define first, then count.
- Related lessons: A trading plan as a hypothesis, VWAP reclaims and rejections, Where a stop goes, Slippage, Marketable limit orders. The Advanced course continues with the first pullback and high-of-day breaks.
Exercises
Seven exercises follow this lesson: three concept checks, two replays of the static 1-minute sequence above (separate observation from inference; then pick every outcome consistent with the information at the decision time before Path B is revealed), one scenario on counting rules, and one practice task. Exercises, answers and progress are available after signing in.
Practice task: count 20 flags before you judge them. On replayed charts or in a simulator, record 20 flags that meet your written definition, with every column of the FennEdge practice log: download the practice log (CSV), with a column guide. Fix the definition, stop, target and time stop before the first entry, apply a slippage penalty to every simulated fill, and record Path A, B or C with R after costs. Log flags you skipped as well, and log flags that broke without you, so the count is not filtered by outcome. Keep the log separate from the Trading Journal, which holds only real broker-synced trades. Then review the block as in Reviewing scalps.
Primary sources
- Lo, A. W., Mamaysky, H. and Wang, J. (2000). Foundations of Technical Analysis: Computational Algorithms, Statistical Inference, and Empirical Implementation. The Journal of Finance 55(4).
- LULD Plan. Limit Up-Limit Down Plan.
- FINRA. Frequent Intraday Trading: Understanding the Basics.
Sources checked 2026-10-05. Marlpit Aerospace and all prices, volumes, outcome splits and costs in this lesson are fictional and Illustrative. This lesson is education, not a recommendation to buy or sell any security. Past patterns do not indicate future results.