--- title: "What Tool Can I Use to Find Double-Bottom Chart Patterns on IDX Stocks?" description: "An actual Trade Inc chat walkthrough: scan scope, POLA’s confirmed pattern, two developing candidates, and the checks that keep a chart label from becoming a trade claim." author: "Trade Inc Research" published_at: "2026-09-30T08:08:38Z" updated_at: "2026-10-03T09:09:02Z" canonical: "https://tradeinc.app/en/blog/ai-chat-double-bottom-confirmation" tickers: ["POLA","BBMD","RIGS"] citations: ["https://tradeinc.app/en","https://www.interactivebrokers.com/campus/glossary-terms/double-bottom/"] --- # What Tool Can I Use to Find Double-Bottom Chart Patterns on IDX Stocks? An actual Trade Inc chat walkthrough: scan scope, POLA’s confirmed pattern, two developing candidates, and the checks that keep a chart label from becoming a trade claim. ## Summary You can use Trade Inc chat to search the available IDX stock universe for double-bottom candidates, ask for their confirmation status, and open a returned ticker in the chart workspace. In our authenticated demonstration on 3 October 2026, the broader scan reported 992 symbols scanned and one excluded. Its ten returned rows included POLA as confirmed, BBMD and RIGS as active, and seven invalid patterns—not ten valid opportunities. This is a walkthrough of an actual product session, not a ranking of competing tools or a test of investment returns. Trade Inc Research is affiliated with the platform. Updated 3 October 2026; the market observations below refer to stored completed daily candles through 2 October. ## Key takeaways - State the universe. Zero matches in five named stocks does not mean zero across IDX. - Separate active, confirmed and invalid results. A displayed pattern is not necessarily a current candidate. - Inspect the underlying troughs and neckline, then check volume separately. - A pattern reconstructed today may not have been identifiable at its historical breakout. - The screenshots are real product views. Indicator buy/sell labels are not this article’s recommendations. ## What are you asking the tool to find? A double bottom is a possible reversal structure after a decline: two broadly similar troughs separated by a rebound. The intervening peak is the neckline. A move above that level and increased volume are commonly used confirmation evidence; neither guarantees a lasting reversal. [IBKR’s double-bottom definition](https://www.interactivebrokers.com/campus/glossary-terms/double-bottom/) The useful product question is not simply whether AI can recognize a W. It is whether you can move from a search to a dated, inspectable candidate. A single-candle pattern, such as a hammer, is a different task from a multi-swing double bottom. ## Step 1: specify the universe before sending the prompt Open [Trade Inc chat](https://tradeinc.app/chat), enter the research request and press the upward-arrow Send button. We initially restricted the request to TLKM, ASII, INDF, ANTM and MEDC. That returned no qualifying matches. Expanding the request to the available IDX universe produced a different result without requiring a saved Screener or a backtest. A practical starting prompt is: > Search the available IDX universe for double-bottom chart patterns using completed daily candles. State the evaluation date, number scanned, exclusions and output cap. Separate active, confirmed and invalid patterns. Show both trough dates and prices, the neckline and latest close. Do not create a saved Screener, run a backtest or place an order. Flag missing evidence instead of filling it in. Ask for scope first, then inspect the answer. “Full universe” remains limited to the tool’s available coverage; it is not proof that every listed security has complete history. ![Trade Inc chat with a follow-up summary prompt entered beside the POLA chart.](https://cdn.sanity.io/images/eybaj6wm/production/a2746896b123e893ccd4bd83b5c9f04e017658d1-1280x720.jpg) *Actual follow-up input, captured 3 October: after the scan, we asked for a compact summary of POLA, BBMD and RIGS. The upward-arrow button submits the request. This screenshot is the summarization step, not the original scan input.* ## Step 2: read statuses, not just ticker names The scan reported 992 scanned and one excluded, accounting for 993 symbols in that run. The response said coverage was capped at 1,000 symbols and the output at ten rows. We therefore do not describe ten as an uncapped total or calculate a market-wide hit rate. | Ticker | Returned status | First trough | Second trough | Neckline | Latest close | | --- | --- | --- | --- | ---: | ---: | | POLA | Confirmed | 30 Jun 2026: Rp50 | 27 Aug 2026: Rp50 | Rp68 | Rp83 | | BBMD | Active, not confirmed | 29 Jul 2026: Rp1,900 | 18 Sep 2026: Rp1,900 | Rp2,060 | Rp2,060 | | RIGS | Active, not confirmed | 9 Jun 2026: Rp575 | 18 Sep 2026: Rp580 | Rp790 | Rp640 | These are the three non-invalid rows among the ten returned, not picks ranked by expected return. The other seven were labelled invalid and are excluded from this candidate table. Latest closes are dated 2 October 2026. Trough prices and latest closes were checked against a separate read-only retrieval of Trade Inc’s stored candles; that is an internal consistency check, not an independent market-data audit. BBMD illustrates why equality matters: Rp2,060 equals its reported neckline, rather than closing strictly above it. RIGS remained below Rp790. Neither should be casually described as a confirmed breakout. ![Actual Trade Inc result table for POLA, BBMD and RIGS alongside the POLA chart.](https://cdn.sanity.io/images/eybaj6wm/production/2f4a490bdb832c1a2f8ba118553cf59411eccb79-1280x720.jpg) *The follow-up table preserves the different statuses. A no-match result from the initial five-stock request was scope-specific; it was not a market-wide conclusion.* ## Step 3: open the candidate and inspect its chart In the returned results, clicking POLA opened its chart beside chat. The workspace link then opened the ticker page; “Buka grafik trade” opened the larger trading chart. Keep the daily interval selected when checking a daily scan. ![POLA’s actual daily trading-chart workspace with price candles and volume.](https://cdn.sanity.io/images/eybaj6wm/production/0db5642f8efa4061fafa0db1c698de5b94d857f7-1280x720.jpg) *Full POLA workspace, captured 3 October. Existing trend and structure overlays are separate indicators, not annotations of the two troughs in our table. This image demonstrates chart navigation; the dated calculations below come from the retrieved candle rows, not measurements from screenshot pixels.* For POLA, the stored series placed both selected lows at Rp50. The maximum intervening high was Rp68 on 8 July. After the second trough on 27 August, the first stored daily close above Rp68 was Rp70 on 8 September. Its 2 October close was Rp83. This verifies a retrospective price sequence. It does not establish that a user received a timely signal on 8 September, that a trade could be filled at Rp70, or that the detector has a particular success rate. ## Check volume—and avoid the hindsight trap Our separate calculation compares POLA’s 8 September volume of 202,937,300 shares with the mean of the preceding 20 stored sessions, excluding 8 September: 7,507,450 shares. The ratio is approximately 27.03 times. That is a large relative-volume observation, not proof of buyer identity or future returns. The pattern scan itself did not apply a volume-confirmation filter. Swing recognition can require later candles. A low’s date and the date it becomes recognizable are therefore different. In this demonstration, the detector described a look-forward confirmation requirement; we did not establish a contemporaneous detection timestamp. Do not label the historical breakout an executable signal or infer backtest performance from it. ## Product fit, limits and what would change the assessment This workflow suits a researcher who wants to narrow a universe, question an AI answer and inspect the chart in one workspace. It was tested in an authenticated Pro demo account; it does not establish access for every account, current pricing, or every plan entitlement. Check the [product overview](https://tradeinc.app/en) and your account before subscribing. If you already review only a handful of charts manually, a conventional charting tool may be enough. If you require scheduled alerts, complete point-in-time data or audited strategy performance, this demonstration does not verify those capabilities. The scan did not expose per-ticker input counts and date windows, and earlier answers displayed a missing-saved-evidence warning. Separate candle retrieval does not repair that provenance gap. Stored prices may not be corporate-action adjusted, and missing sessions can affect pattern labels. We therefore report the observed workflow and a cross-checked example, not exhaustive detection accuracy. A subsequent failure to hold the neckline would weaken the bullish interpretation; a breach of the trough area would weaken it further. Continued strength may support the interpretation, but cannot eliminate liquidity, execution or reversal risk. ## Conclusion Use AI to produce an inspectable shortlist—not an automatic trade instruction. The important sequence is scope, status, dated structure, chart inspection and a separate volume check. In this example, broadening a five-stock request uncovered candidates; separating statuses prevented seven invalid rows from being presented as opportunities. For adjacent workflows, see [candlestick analysis through chat](https://tradeinc.app/en/blog/ai-chat-candlestick-analysis) and [daily versus weekly chart inspection](https://tradeinc.app/en/blog/compare-daily-weekly-idx-charts). This is educational product research, not investment advice. [1]: https://tradeinc.app/en "Trade Inc product overview" [2]: https://www.interactivebrokers.com/campus/glossary-terms/double-bottom/ "IBKR: Double Bottom"