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.

Almost all of it is public. 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.
• NSE corporate filings ↗ — announcements for tracked holdings. The same filings are also requested from BSE ↗ as a backup.
• Reserve Bank of India ↗ — policy rates.
• CCIL ↗ — the India 10-year government bond yield, with the RBI as the fallback.
• MoSPI ↗ — CPI inflation, monthly.
• CMIE Prowess ↗ — the long Nifty 50 price and P/E history (January 1995 to May 2026) behind the history charts and the backtests. This one is a licensed database, not a public source: its figures appear in charts and results, and its raw series are not offered for download.

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 stated. Each write-up lists the costs it charges and what it leaves out. Most charge 0.03% a side and do not model STT or slippage, and say so; with a thin edge, those costs alone could close it.
• 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 the result should be read 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.

Errors get fixed in place and the page notes what changed. If you find one, email shivanshforcollege@gmail.com.