Descending triangle pattern: support and lower highs

Recognize horizontal support and descending resistance

The descending triangle chart pattern narrows because each rebound reaches a lower high while the lows stay near the same level. TradingView describes this geometry in its Triangle drawing tool and explicitly allows an upside break. A hand-drawn outline identifies a candidate; it does not establish a trade entry or a probability of success.

A hypothetical five-touch specification

This example uses three support visits and two rebound highs, in the order L1, H1, L2, H2, L3. It is a custom specification, not a TradingView default or a recommended trading setup. Use standard price candles. Confirm each swing only after three bars on either side have strictly lower highs or strictly higher lows, respectively; reject ties.

• Require every close from L1 through candidate creation to remain between S - 0.3 and R(t) + 0.3. Freeze the selected swings, S and m when the candidate becomes available.
• Here L3 becomes confirmed at the close of bar 33. Assume the containment rule still holds then; evaluate the first breakout on bars 34 through 39. A move that happened before creation is not backdated into a signal.
• If consecutive detected pivots have the same type, retain the more extreme one before selecting the five-swing sequence. Once a candidate is active, keep its anchors unchanged until its first breakout or expiry.
• The 0.3-unit band is meaningful only for this hypothetical price scale. Percentage or volatility-based tolerances are alternative models that require separate tests.

Compare a breakdown, a false breakdown and an upside break

All paths below start from the same saved candidate and branch after bar 33. A downside event is the first completed close below S - 0.3; an upside event is the first completed close above R(t) + 0.3. Equality does not qualify. Consume the candidate after its first event so one structure cannot generate repeated entries.

• For this experiment, label an early false breakdown when a close above S + 0.3 occurs within the next five bars after the downside event. A later recovery is logged separately. This definition is not TradingView's built-in Failed status.
• Apply a corresponding rule to upside events: a close below R(t) - 0.3 within the next five bars is an early failed upside break. Use the saved line for this follow-up, even if the original candidate has expired.
• An upside break before a downside event is not a failed short trade when no short entry existed. Keep candidate invalidation, post-signal failure and trade profit or loss in separate columns.

Measure volume without using it to choose the answer

Volume describes participation around a move; it cannot establish that support will break or that a breakout will persist. Schwab discusses volume as evidence used alongside price trends. For this example, compare the same volume rule for both directions and retain the unfiltered sample as a baseline.

• Record formation volume over fixed windows, such as bars 10 through 19 versus bars 20 through 29. Report whether the average rose or fell; do not require contraction after seeing the eventual outcome.
• An illustrative breakout filter is current completed-bar volume divided by the average volume of the previous 20 completed bars, excluding the trigger bar, at least 1.5. A ratio of 1.8 passes; 1.2 does not. These thresholds have no measured advantage here.
• Log a missing or zero denominator as unavailable. Do not silently count it as a pass, and document whether the feed represents traded volume or tick activity before comparing instruments.
• On intraday charts, a simple rolling average can mix quiet and active parts of the session. Comparing equivalent times of day is another model; keep that change separate from the pattern rules.
• Higher volume on a failed breakdown remains a failed breakdown. A volume filter should not remove losing events from the unfiltered record.

Separate the swing date from the signal date

TradingView's automatic Triangle detector documents 5/5 pivots and close-based breakouts. Its In Progress option can include emerging points that are not confirmed pivots. This differs from the three-right-bar example above. The drawing tool, the automatic indicator and a custom detector need not select the same triangle.

• A pivot at bar 30 with three required bars on its right becomes known at bar 33, not bar 30. Store the pivot bar, confirmation bar and breakout bar separately.
• Plotting a confirmed pivot back on its original candle can make a historical chart appear earlier than the information was available. Replay the candidate one completed bar at a time when checking the implementation.
• Waiting for a completed bar addresses an intrabar condition that may disappear before close. It does not prove that every other source of repainting has been removed.
• Do not reconstruct the historical sample from only the patterns still visible on the chart. Save events when they occur, including candidates that expire and breakouts that later fail.

Test both directions with fixed trade assumptions

Pattern recognition supplies events, not a complete strategy. Before comparing outcomes, specify order timing, position size, stops, targets, maximum holding time, commissions and slippage. Use the same candidate population for the downside-only and two-direction versions, then explain which events each version trades.

• An illustrative height projection uses H1 - S = 10. A downside reference is S - 10 = 90; an upside reference from R(34) is 102.4 + 10 = 112.4. This is a chosen geometric convention, not a forecast or guaranteed fill.
• Keep entry-at-breakout and entry-after-retest as separate experiments. Waiting for a retest can miss a move that never returns, and a touched level does not by itself specify an executable order.
• Ask Pineify AI Coding Agent to draft the saved-anchor, confirmation and event-log rules. Review the generated code and its syntax, then check that signals appear on the correct bars in TradingView.
• After those rules are implemented as a strategy, expose tolerances and confirmation choices as supported inputs for Strategy Optimizer. Compare parameter combinations on a research period, then evaluate the chosen settings on a separate period.
• Report expired candidates, both breakout directions, early failures and net trade outcomes. A profitable selected example or the best parameter combination cannot establish general reliability.