Bear flag pattern: breakdown rules and squeeze risk

Recognize the falling pole and rebound channel

A bearish flag chart pattern needs a preceding decline. An upward channel by itself is not evidence of bearish continuation. TradingView describes the flag as a parallel correction after the main move; a pennant instead narrows toward an intersection. Decide how much rebound and how many bars your definition permits before inspecting the outcome.

A hypothetical channel with an explicit breakdown rule

This worked example uses a pole from 120 at bar 0 to 100 at bar 8. Its height is 20 price units. Assume the rebound has lower anchors at bar 10, price 101, and bar 14, price 103, plus an upper anchor at bar 12, price 106. These are illustrative inputs, not observed prices, recommended settings or a complete automatic detector.

• Assume all flag closes from bar 10 through bar 16 stayed between L(t) - 0.2 and U(t) + 0.2, and no high exceeded 110. Otherwise this example never becomes eligible.
• Before a breakdown, invalidate on a high above 110 or a close above U(t) + 0.2. Check invalidation first when both conditions occur in one bar. Expire an unconfirmed candidate after bar 22.
• Freeze the anchors at creation and consume the candidate after its first breakdown. The 0.2-unit buffer belongs only to this hypothetical price scale.
• A pole-height projection from the bar 17 lower boundary is 104.5 - 20 = 84.5. This geometric reference is neither an expected return nor proof that price will reach it.

A failed breakdown can turn into a sharp upward move

Use the same saved channel to compare the paths below. For this example, an early failed break means a close above L(t) + 0.2 within five bars after confirmation. That custom definition is distinct from a trade stop and from TradingView's built-in pattern status. The original breakdown stays in the log even when the failure occurs.

• A short squeeze involves short sellers buying back positions as prices rise, which can add buying pressure. Schwab explains this mechanism; the bear flag shape does not establish that a squeeze is happening.
• For a hypothetical short filled at 104, covering at 111 loses 7 per unit before costs. A planned stop at 109 does not prove a 109 fill if price gaps through it or available liquidity is insufficient.
• A close-based failure exit is evaluated only at the close. It can leave the position exposed during the bar; an intrabar stop is a separate order rule with its own fill assumptions.

Hypothetical paths after candidate creation at bar 16

A. Wick only, bar 17:
   low 104.0, close 104.6; L(17) = 104.5
   No confirmed breakdown below 104.3.

B. Failed breakdown:
   bar 17 close 104.0 -> bar 18 close 105.5
   L(18) = 105.0; 105.5 exceeds reclaim level 105.2.
   Record a breakdown followed by an early failure.

C. Upward acceleration after path B:
   bar 19 high 111.0, close 110.5; U(19) = 109.5
   Price has also moved above the flag and the rebound cap.
   A short still open is exposed to the rise.

TradingView confirmation arrives after the swing

TradingView's Bearish Flag indicator uses 5/5 pivots and fixes a breakout using the close. In Progress can include emerging endpoints that are not confirmed pivots. Its Pattern Breakout alert registers the target-direction break, so an absent breakout alert does not prove that an upward invalidation was harmless.

• A five-right-bar pivot at bar 14 cannot be confirmed until bar 19. The example above assumes its anchors are known at bar 16 and therefore does not reproduce the built-in detector. A custom two-right-bar rule would have a different confirmation delay.
• Keep the anchor bar, anchor confirmation bar, candidate creation bar and breakdown bar separate. Back-plotting an anchor on an earlier candle must not move the trade signal into the past.
• Check signals one completed bar at a time. Waiting for close prevents reliance on a temporary intrabar condition, but it does not establish that all other sources of repainting are absent.

Stock and Bitcoin shorts require different assumptions

A bear flag trading pattern describes price structure, not account eligibility. The SEC explains that stock short sales generally involve borrowed shares and can have theoretically unlimited losses. A backtest order is a broker-emulator event; it does not reserve shares, check an account's permissions or prove that the historical borrow cost was affordable.

• For a Bitcoin bear flag chart pattern, record the exchange symbol, quote currency, contract type and timeframe. Do not mix a spot signal series with perpetual execution results without specifying how the two are connected.
• Coinbase International Exchange documents periodic funding payments and margin-based liquidation. Those are venue rules, not a universal funding rate or liquidation threshold for every Bitcoin instrument.
• Model borrow charges or funding separately when the test does not include them. Mark unavailable historical execution data as a limitation rather than silently assigning zero cost.

Turn the definition into a testable strategy specification

Write the detector rules before choosing entries and exits. The channel example begins with supplied anchors; an automatic detector additionally needs a repeatable rule for selecting the pole, alternating swings, ties and overlapping candidates. Different choices can produce different samples even when each chart looks like a bearish flag pattern.

• Specify order timing, size, stop, target and maximum holding period. Under typical default TradingView strategy behavior, an order created after a bar closes can first fill at the next bar open, not automatically at the signal close.
• Keep immediate breakdown entries and retest entries as separate experiments. A retest condition needs its own price band and expiry; some declines never return.
• Give Pineify AI Coding Agent the full specification to draft the Pine Script. Review the code, check its syntax and verify signal timestamps against the written rules before using its trade report.
• Expose supported numeric or Boolean choices as strategy inputs for Strategy Optimizer. Compare a limited, declared parameter set on a research period, then examine the selected rules on a separate period.
• Retain invalidated candidates, expired candidates, failed breakdowns and all executed trades. Report costs and adverse moves as well as net results. Parameter selection cannot establish borrow availability or future profitability.