Anatomy of a market correction: 19 years of Nifty falls

What the record shows

A correction feels like a unique event while you are inside it. Over a long record it looks routine. We went through every Nifty 50 closing price from 17 September 2007 to 1 October 2026, about 4,670 trading days, and marked every time the index fell 10% or more from its previous closing high.

There were ten such falls. Figures are for the price index on closing prices, so they exclude dividends; with dividends, every recovery would have come a little sooner.

The current one began at a close of 26,329 on 2 January 2026, reached its low on 30 March and, after a weak September, left the index at 22,422 on 1 October, 14.8% below the peak. The September 2026 recap has the month's detail.

Four things the table teaches

1. Corrections are normal, not rare. Ten falls of 10%+ in 19 years is one every two years or so. Look within calendar years and it is more striking: in 14 of the 18 full years from 2008 to 2025, the Nifty fell at least 10% from that year's own high at some point. Only 2014, 2017, 2023 and 2025 escaped.

2. A fall inside a year rarely decides the year. Of those 14 years with a 10%+ fall, 11 still ended higher than they began. 2020 is the clearest case: a 38% collapse by March, yet the year finished up about 15%. Only 2008, 2011 and 2015 ended in the red.

3. Depth and duration are not tightly linked. The 2020 crash was the second deepest on the list and one of the faster recoveries, at ten months. The 2010–11 fall was less than half as deep and took three years to heal. A sharp crash with a clear cause can reverse quickly; a slow grind driven by earnings and rates can take longer. Nobody knew which kind they were in at the time.

4. Most recent corrections have been moderate. Since 2016, every fall except 2020 has been between 10% and 18%, with recoveries of 6 to 15 months. That is not a promise; 2008 is in the same table. It is a reminder that "this is the big one" is a feeling far more often than a fact.

The best days live inside the worst periods

The strongest argument for staying put is not the recoveries. It is where the recovery days happen.

The Nifty 50 price index went from 4,495 on 17 September 2007 to 22,422 on 1 October 2026, about 8.8% a year. Here is what happened when we removed its best days:

Missing ten days out of about 4,670 halves the return. And all ten of the best days came during the 2008–09 and 2020 crashes, while the index was far below its previous high: the 17.7% jump of 18 May 2009, 8.8% on 7 April 2020, 7.0% on 31 October 2008, and others in October and November 2008 and March 2020.

That is the trap of selling in a correction. The investor who exits to "wait for things to settle" is out of the market on exactly the days that do most of the repairing, because those days arrive when things are least settled. The worst and best days cluster together; you cannot reliably keep one and skip the other. Best and worst Nifty days of 2026 looks at this year's biggest daily moves.

Staying calm, practically

Calm is easier with a plan made before the fall than with willpower during it.

• Know your number in advance. If a 35–40% fall in your equity would force you to sell, you hold too much equity. Fix the allocation in calm times, using something like the asset allocation calculator.
• Keep the SIP running. A SIP buys more units when prices are low. Stopping it during a correction is selling the discount. A twenty-year SIP through crashes walks through how that played out.
• Separate near-term money. Money needed within three years should not be in equity. Then a correction is a paper loss on long-term money, not a threat to next year's fees or down payment.
• Rebalance, don't react. If equity has fallen well below its target weight, topping it up from debt is a rule-based way to buy low.
• Read less news. The psychology of a market crash explains why the urge to act peaks just when acting is most expensive.

History cannot say how deep or long the current fall will run. It does say that falls like it have been common, that most years containing one still ended up, and that the costliest mistake in each of them was leaving.

This post is educational and not investment advice. Past index movements do not predict future returns.