Which AI Tool Can Check IDX Market Breadth Without Hiding the Denominator?

Trade Inc’s connected MCP research tools can retrieve stored breadth for completed IDX sessions, subject to account access. They are useful when you want the underlying counts inside an AI conversation—not just a bullish or bearish label. The important buying question is whether you can inspect the denominator.

In our read-only check on 14 October 2026, the 13 October session returned 352 advancing, 308 declining and 331 unchanged names. That means 53.33% of moving names advanced, but only 35.52% of all 991 covered names advanced. Both calculations are correct. They answer different questions.

Key takeaways

• Ask for advancing, declining and unchanged counts alongside the percentage.
• Separate the completed trading session from the time the data was computed.
• Treat the returned coverage as a stated universe, not proof that every IDX security is included.
• Check arithmetic before relying on a market-regime label.

What can the connected tool actually do?

Trade Inc’s public MCP connection guide documents completed-session breadth, trend mix and foreign-flow context. This is an account-authenticated connection to a compatible AI client. It does not establish that every feature is available as a website button or included in every plan.

For this guide, we tested only a read-only request for the last five completed sessions. We did not create a Screener, run a backtest or place an order.

A useful request is: “Show the last five completed IDX sessions. Include advancing, declining and unchanged counts, covered names, session dates and the calculation behind the advance ratio. Flag any totals that do not reconcile.”

Before choosing a subscription, confirm access for your account and AI client. The useful feature here is inspectable historical context, not a promise that an AI answer will always be correct.

A five-session example you can check

These are Trade Inc’s returned counts, read on 14 October 2026. They have not been independently reconciled against an official exchange security-by-security file.

In every row, advancing plus declining plus unchanged equals 991. The displayed percentages use advancing divided by advancing plus declining.

The latest record was computed at 23:50:34 WIB on 13 October. It describes that completed session, not live trading on 14 October.

Why the denominator changes the answer

For 13 October, the moving-name denominator is 352 + 308 = 660. Dividing 352 by 660 gives 53.33%. Among names whose prices changed, more advanced than declined.

The full covered denominator is 991. Dividing 352 by 991 gives 35.52%. Meanwhile, 331 divided by 991 gives 33.40% unchanged. Leaving that group out without saying so makes the first percentage easy to misread as a majority of the entire covered universe.

The difference between 15.73% on 12 October and 53.33% on 13 October shows a change in this particular participation measure. It does not establish a durable trend reversal. These are counts of names, not market-cap-weighted index returns, probabilities of profit or forecasts.

Where the evidence stops

The returned data does not, by itself, settle the eligibility rules for the universe or explain every unchanged name. Do not infer that unchanged necessarily means suspended or inactive.

We also found a separate consistency problem: the returned trend-category counts summed to 992, while reported coverage was 991, across these five records. We could not establish the cause. Consequently, this guide excludes those categories from its worked breakdown and does not use the supplied regime labels as validated conclusions. A tool should make such discrepancies visible, not quietly adjust them.

For repeat research, retain the raw counts and dates. If coverage changes, investigate the change before comparing percentages across periods.

Who is this useful for?

This workflow fits researchers who already work in an AI client and want a compact, auditable participation check before examining individual stocks. It is less necessary if your existing dashboard already supplies clear counts and definitions. It is not a substitute for an intraday feed, a full exchange-universe audit or security-level research.

The practical conclusion is simple: evaluate a breadth tool by whether it exposes counts, dates, coverage and exceptions. For this verified sample, the advancing/declining/unchanged arithmetic is reproducible; the separate trend distribution still needs reconciliation. Keep that distinction in the answer you use.

Publisher disclosure: Trade Inc Research publishes this guide and is affiliated with Trade Inc. This is a bounded product demonstration, not an independent ranking or investment recommendation. The cover is a conceptual illustration, not a chart of the observations.

This article is for research and information only. It is not personalized financial advice or an instruction to trade.