Rounding top pattern: a slowing advance needs confirmation

Follow the slope change across the crest

The illustration uses invented prices on an arithmetic scale. A prior rise from 85 to 100 precedes the arc. Over equal ten-bar intervals, the sampled closes change by +16, +4, -4 and -15.2. Those increments show the transition from a fast rise to a flat crest and then a steeper decline. They are checkpoints for reviewing the full candles, not enough data to reconstruct every swing.

Fix the neckline, time window and tolerance

For this manual exercise, choose the existing support at 100 before reviewing the right half. The selected left-base low is 100 and the right-base low at bar 40 is 100.2. Keep the line at 100 instead of averaging it upward after seeing the break. The following limits are explicit teaching assumptions, not universal rounding-top settings or fitted parameters.

• These are manually selected checkpoints. A scripted detector also needs an exact rule for selecting support and identifying swings, including when a later candle makes a pivot knowable. It must not backdate a signal to an earlier pivot bar.
• Forty daily bars and forty hourly bars represent different observation periods. Record both the timeframe and the first and last timestamps; do not convert bars into calendar days without the trading calendar.
• The equal left and right spans in this illustration make the geometry easy to inspect. They do not establish that real formations must be perfectly symmetric.

Check volume without letting it replace the price break

Trade Ideas describes volume fading around the crest and increasing on the breakdown. That is supporting context, not proof of who is trading or a guarantee of continuation. For a reproducible example, compare the break bar with the mean volume of the preceding 20 completed bars, excluding the break bar itself.

• The volume numbers and 1.5 threshold are invented for arithmetic review. No evidence here establishes an optimal multiplier or improved accuracy.
• TradingView reports that some symbols have no volume and some provide tick volume, which counts price updates. Check the exact symbol and feed; a tick-volume ratio is not a traded-share ratio.
• If a strategy requires both tests, a price break with insufficient volume is not its entry signal. Save both fields so the price event is not silently removed from the record.

Distinguish a high-level range from a failed break

A plateau near the highs has not completed the descending half. In a counterexample, closes stay between 118 and 120 through bar 40 while support remains 100. There is neither a return to the base nor a neckline break. Calling that range a confirmed top assumes the decline before it happens.

• For this exercise, monitor the five bars after confirmation for a close strictly above 100. The bar-42 reclaim qualifies. A close exactly at 100 does not pass that strict failure test; record it separately.
• A break that remains below 100 for five bars is only a break without a reclaim during that window. It is not automatically a winning trade.
• The geometric downside reference is 100 - (120 - 100) = 80. It does not supply a probability, an expected fill or evidence that price will reach 80.
• A rounding top does not require a handle. An inverted cup and handle adds a separate smaller rebound after the rounded structure; adding that requirement changes the pattern being studied.

Review the formation on TradingView as it develops

Keep the symbol, venue, timeframe and price scale visible. Use standard candles, mark the two base observations and the crest, and save the neckline before advancing through the breakdown. Record when you identified the candidate as well as the dates of its older turning points. TradingView documents that a realtime bar is confirmed only on its closing update.

• Save a screenshot before the candidate resolves, then another at the close used for confirmation. Include the volume pane and its data source when applying the volume filter.
• For the example log, save: neckline 100; height 20; bases 0 and 40; candidate known at 40; break at 41; reclaim at 42; volume ratio 1.6. The failure must remain attached to the original break.
• A close-based pattern event is not an executed short order. A strategy still needs entry timing, risk limits, borrowing or instrument availability, costs and slippage assumptions.

Review the labels with Pineify, then formalize the rules

Upload the labeled screenshot to AI Chart Analysis with the prior rise, full arc, neckline and timeframe visible. Use the analysis to review whether the right side actually descends to support and whether the marked close is below it. Compare the output with the chart values; an image interpretation cannot verify hidden candles or establish recognition accuracy.

• For a script draft, give Pine Script AI Coding Agent the support-selection method, base-touch tolerance, crest width, duration limit, confirmation rule and five-bar failure window. Define candidate selection and overlap handling before requesting automatic scanning.
• Keep price and volume tests separate, specify how missing volume is handled, and require the signal to use only information known at that bar. Review and syntax-check any generated source before comparing its events against labeled examples.