How we score

Every trading day, a program reads the news and market data, scores each candidate event on seven parameters, and publishes the highest. No person chooses the event, and no AI writes the text: it is assembled from fixed sentences filled with the computed numbers.

When it runs

Twice on each NSE trading day: a morning edition at 08:30 IST, from overnight news and the previous session, and a closing edition at 15:45 IST, which adds the day’s market reaction. Weekends and NSE trading holidays are skipped.

Finding candidate events

We read headlines published in the last 18 hours from The Economic Times, Mint, Business Standard and BusinessLine. Each headline is matched against fixed keyword rules for the kinds of event below. Results and corporate actions are kept only when they name a Nifty 50 company, and each company is its own candidate. The market move itself is always a candidate too.

The seven parameters

Each is scored from 0 to 100. The event score is the weighted average, rounded.

Market impact25%

How far the market moved, measured against its own normal swing. We take the biggest of three moves: the Nifty 50, the sector indices the event touches, or (for a single company) that company’s share price. Each is divided by its typical daily move over the last 20 sessions. A move of 2.5 times the usual swing or more scores 100.

Company impact15%

How many Nifty 50 companies the event reaches. For a theme such as an RBI decision, it is the share of Nifty 50 companies in the industries affected, scaled so that about 60% of the index scores 100. For one company’s news it scores 28: a base of 20, because one heavyweight can still move the index, plus four times that company’s share of the 50.

Investor relevance15%

A fixed weight for each kind of event, set once and the same every day, for how directly it changes what an Indian investor holds. The table below lists them.

News velocity15%

How much is being written: 8 points per report in the last 18 hours plus 12 points per separate publisher, capped at 100.

Search interest10%

There is no free source of search volume for India, so we use a stated stand-in: how many separate publishers ran the story within six hours of the first report, at 25 points each.

Historical significance10%

Whether days like this have led to unusually large moves. We compare the spread of the Nifty’s five-session outcomes after comparable days with the spread after all days. The same spread scores 50; twice the spread scores 100.

Data availability10%

How complete our inputs were: headlines, the Nifty’s move, sector or company prices, and enough history. Each available input is a quarter of the score.

Bands

75 and above is a major event, 60 to 74 significant, 45 to 59 notable, and below 45 minor.

Investor relevance by kind of event

Kind of event	Sectors it moves first	Relevance
RBI policy	Private banks, PSU banks, Financials, Realty, Auto	90
Inflation data	Private banks, FMCG, Consumption	75
Foreign fund flows	Private banks, IT, Financials	70
Growth data	Infra, Metals, Auto	65
US Federal Reserve	IT, Financials, Metals	65
Oil prices	Oil and gas, Energy, Auto, FMCG	60
Quarterly results	The company named	60
Government and SEBI policy	Financials, Infra, PSU banks	60
The rupee	IT, Pharma, Oil and gas	55
Corporate action	The company named	50
IPOs	Financials	45
The market move	The whole market	60

What followed similar days

We use ten years of the Nifty 50’s daily closes. “Similar days” are sessions when the Nifty moved in the same direction by a similar amount, within 0.6 of its usual daily swing. Once our archive holds eight or more past events of the same kind, we use those dates instead. At least 20 similar days are needed; with fewer we say so and give no cases.

We then show how often the Nifty ended more than 1% higher, within 1% either way, or more than 1% lower five sessions after those days, with the median move and the moves a quarter of them beat or fell short of. These are past outcomes only. We do not project a future level for the Nifty or any stock.

Other figures we work out

FII and DII monthly totals add up NSE’s provisional daily net figures that we saved. A month shows how many of its trading days are included, and days we did not capture are left out, never estimated.

IPO lead managers and registrars: for each firm named in NSE’s issue details, we take the median (middle) listing-day open gain against the issue price across the issues it handled that year, and count how many opened above the issue price. An issue with several lead managers counts once for each.

Options backtest: the strategy is opened at the strike nearest the index’s close on the entry day, at each contract’s closing price in NSE’s end-of-day F&O file, and valued at expiry from the index’s final value. A contract that did not trade on the entry day stops the test, because its closing price is only theoretical.

What this is not

The score describes how big and how widely felt an event is, and the history describes what has happened before. Neither is a forecast, and nothing here is a recommendation to buy or sell. Prices are delayed and come from public sources that can be wrong or late. See the disclaimer.