Methodology

The headline breadth figure is the percentage of the Nifty 100 trading above its own 200-day moving average. Daily closes come from Yahoo Finance; the moving average is computed over the 200 sessions before the day being measured, so the day’s own close never sits on both sides of the comparison.

A constituent that fails to fetch is dropped from that day’s count rather than assumed unchanged, so the denominator can move by a name or two between days.

Nothing here is proprietary. In rough order of how much the site leans on them:

• Yahoo Finance ↗ — daily closes for index constituents and portfolio holdings, with an NSE quote fallback.
• NSE Indices ↗ and NSE ↗ — index levels, published constituents, and end-of-day participant flow data.
• BSE ↗ — corporate announcements for tracked holdings.
• Reserve Bank of India ↗ and data.gov.in ↗ — policy rates and macroeconomic series.

Where a page shows a single indicator, its own source is named on that page. If a source and this site disagree, the source is right.

Data is refreshed by scheduled jobs that run on weekdays only, after the Indian market closes — breadth first, then index levels, holdings prices, flows and announcements as each upstream source publishes. Flow data in particular lands late, because NSE publishes it well after the close.

This means figures are end-of-day, not live. Nothing on this site is a real-time quote, and none of it should be used as one. Weekends and market holidays show the last trading day’s values.

These are the constraints a backtest has to satisfy before it gets published here:

• No lookahead. A signal computed from a day’s close is acted on at the next session’s open at the earliest. Rolling statistics exclude the day being evaluated.
• Costs are charged. Brokerage, STT and slippage assumptions are stated in the write-up. A result quoted gross of costs is labelled as such.
• The universe is fixed in advance. Testing a rule on today’s index members over ten years quietly tests it on the survivors; where a point-in-time universe isn’t available, the write-up says so and treats the result as an upper bound.
• Parameters are reported honestly. If several variants were tried, the write-up says how many and shows the spread, not just the best one.
• The code is in the repository. Every published backtest is reproducible from the scripts in the public repo linked in the footer.

Errors get fixed in place and the page notes what changed. If you find one, the fastest route is an issue on the GitHub repository ↗.