How Far Does NIFTY Move by Expiry?

The short answer

A typical NIFTY expiry week moves about 1.2% from open to expiry close, in either direction. Four weeks in ten finish within 1% of where they started, three in four finish within 2%, and only about one in eight moves more than 3%. Up weeks barely outnumber down weeks. The handful of very large weeks almost all sit in the spring of 2020.

Median weekly move, either way: 1.19%
Weeks closing within ±1%: 41.6%
Weeks moving beyond ±2%: 26.1%
Weeks that closed up: 52.9%

Dashboard

Pick weekly or monthly expiries, choose which move buckets to count, and every chart below redraws. With all buckets on, each column shows the full spread of moves. Turn some off and each column shows only the share of weeks that landed in the buckets you kept: switch on just the two largest down buckets and you are looking at how often each month delivered a fall of more than 3%. Hover or tap any column or cell for its full breakdown.

Expiry

Weeks

Show

Move buckets (move_pct_bucket)

Overall distribution

By calendar month

By month and year

Month-year calendar grid

By year

Show the numbers by calendar month

Colours run from deep cerise (the largest falls) through pale tints near zero to deep blue (the largest rises); rises are blue rather than green so the two sides stay distinct for colour-blind readers. Monthly cycles use wider buckets than weeks, because a month moves about twice as far.

The shape of an expiry week

Weekly moves bunch tightly around zero. The two buckets either side of zero, a fall of up to 1% and a rise of up to 1%, hold 166 of the 399 weeks between them. Extend that to ±2% and you cover 295 weeks, or 74%. The middle half of all weeks sits between about −1.0% and +1.4%, and only one week in ten falls by more than 2.4% or rises by more than 2.4%.

The tails are thin but long. Fifteen weeks moved more than 5%, eight down and seven up, and nine of those fifteen came in 2020. The worst was the week to 12 March 2020, down 12.4%; the best was the week to 9 April 2020, up 9.0%, a reminder that the biggest rallies tend to arrive in the same weeks of panic as the biggest falls.

Move from open to expiry close	All weeks	Monthly cycles
Beyond ±1%	58.4%	–
Beyond ±2%	26.1%	69.7%
Beyond ±3%	13.0%	–
Beyond ±5%	3.8%	29.7%
Beyond ±8%	–	6.3%
Median move, either way	1.19%	2.98%

Where it closes is not how far it travels

The expiry close hides the journey. Measured from the same starting price, the median week reached 1.0% above it and 1.1% below it at some point before expiry, and its full high-to-low range was 2.5%, twice the median close-to-open move. Half of all weeks touched a level 2% away from the start at some point, but only a quarter closed beyond it, and in one week in twelve the index went more than 2% one way and still finished within 1% of where it began. For anyone holding a position through the week, the path matters as much as the destination.

Does a bad week predict the next one?

Barely. The correlation between one week’s move and the next is −0.06, indistinguishable from none. After the 48 weeks that fell by more than 2%, the following week rose 44% of the time, with a median move of −0.3%: no reliable bounce, and no reliable follow-through either.

Calendar months: patterns that do not survive a test

Pool every week by the month its expiry fell in and some months look distinctive. June closed up in 25 of its 35 weeks (71%), November in 19 of 29 (66%) and August in 21 of 35 (60%). January looks weakest: up in only 11 of 30 weeks (37%), with a median move of −0.4%. March has the widest spread, with a standard deviation of 3.4% against 1.4% to 1.5% for November and December, though nearly all of that comes from the four March weeks of 2020.

None of this is strong enough to trade on. Each month holds only 29 to 36 weeks, and with twelve months to compare, one or two will look unusual by chance alone. A chi-square test of up versus down weeks across the twelve months gives p = 0.34, and a Kruskal–Wallis test of the moves themselves gives p = 0.57: the monthly differences are well within what random variation produces. The calendar-month chart is best read as a description of what happened, not a forecast.

Years and month-years

Years differ more than months do, mostly in how wide the spread is. 2020 is in a class of its own: a weekly standard deviation of 3.9% and both the largest fall and the largest rise in the series. 2023 and 2025 were the calmest years, with typical weeks moving about 1.1% and 0.9%. 2026 so far is the weakest year in the data: only 15 of its 39 weeks closed higher (38%), and the median week fell 0.7%.

The month-year view splits each calendar month by year, and here the sample gets small. A month holds three to five expiry weeks, so its percentages move in steps of 20% to 33%, and a single week can swing a month from “all up” to “mostly down”. Read the timeline or calendar grid for clusters rather than single cells. Set the filter to the two largest down buckets and February to May 2020 stand out as the longest run: four months in a row, each with at least one fall beyond 3%. Set it to the two largest up buckets and the rebound lights up instead, with a rise beyond 3% in every month from March to July 2020 and again in October and November.

Thursday and Tuesday expiries

NIFTY’s expiry day moved from Thursday to Tuesday from September 2025. The 57 Tuesday weeks so far closed up only 46% of the time, against 54% for the 342 Thursday weeks, and moved a little less (median 1.0% either way against 1.2%). Do not read that as an effect of the weekday: two-thirds of the Tuesday weeks fall in 2026, the weakest year in the data, so the comparison mostly measures the market of the past year.

Monthly-expiry weeks: a different perspective

Once a month the weekly expiry coincides with the monthly expiry, when the month’s futures and options settle along with the weekly contracts. An earlier version of this analysis left those weeks out. Putting the 92 monthly-expiry weeks back in, and comparing them with the other 307, shows they run hotter:

Weekly expiries	Monthly-expiry weeks	Other weeks
Weeks	92	307
Median move, either way	1.46%	1.12%
Moved beyond ±1%	67.4%	55.7%
Moved beyond ±3%	16.3%	12.1%
Closed up	55.4%	52.1%
Median move	+0.14%	+0.16%

Direction is no different: monthly-expiry weeks closed up about as often as other weeks, and their median move is the same. Size is different. A typical monthly-expiry week moved about 30% further, and two in three moved more than 1%. The gap in size is borderline rather than conclusive (a Mann–Whitney test on the size of the moves gives p = 0.07), but it runs the same way at every threshold. Leaving these weeks out understated how far a typical expiry week travels; with them in, the median weekly move rises from 1.12% to 1.19%. Use the “Monthly-expiry” and “Other” buttons on the dashboard to compare the two groups month by month.

Monthly expiry cycles since 2012

Monthly expiries give a longer view, because they predate weekly options by seven years. A monthly cycle runs from the first open after one monthly expiry to the close on the next, and the data covers 175 of them from March 2012. Over that span the index rose in 100 cycles (57%), with a median move of +0.8% and a typical size of 3.0% either way. Seven cycles in ten moved more than 2%, three in ten more than 5%.

The extremes are concentrated again. The cycle to 26 March 2020 fell 24.1%, twice as far as any other; the second worst is recent, the cycle to 30 March 2026, down 12.5%. The best was the cycle to 26 November 2020, up 11.2%.

By calendar month the monthly cycles show the same temptation as the weeks. April rose in 12 of 15 years (80%) and October in 11 of 14 (79%), while January and February each rose in only 5 of 14. With 14 or 15 cycles per month, these are small samples: a chi-square test across the twelve months gives p = 0.18 and a Kruskal–Wallis test p = 0.53. April and October have been good months for this index over this period; that is not the same as a reason to expect the next one to be.

Method and limits

Source: Daily NIFTY 50 open, high, low and close from 21 February 2012 to 1 October 2026.
Expiry calendar: Weekly expiries from 14 February 2019, on Thursdays until 28 August 2025 and Tuesdays from 2 September 2025. The monthly expiry is the last expiry weekday of the month. When the expiry day was a holiday, the previous trading day’s close is used (17 weekly and 9 monthly expiries).
Period: Each week or cycle starts the day after the previous expiry, at that day’s open. When that day was a holiday, the start price is the previous trading day’s close. Periods never overlap and leave no gaps.
Buckets: Weekly: below −5%, −5% to −3%, −3% to −2%, −2% to −1%, −1% to 0%, and the same steps above zero. Monthly: below −8%, −8% to −5%, −5% to −3%, −3% to −1%, −1% to 0%, and the same steps above zero. No move landed exactly on a boundary.
Grouping: A week or cycle belongs to the month and year of its expiry.
Validation: The build script rebuilds the earlier weekly-only analysis from the same daily prices and matches 306 of its 307 weeks exactly. All four of its distribution tables also match. The one difference is the first Tuesday week, which now starts on 29 August 2025 so that it does not overlap the 28 August monthly week. Every distribution sums to 100%, and every count adds up to the number of weeks.

The index is measured on its price alone, without dividends. Weekly options and the monthly expiries are a calendar for slicing the data, not the price of any contract, so these moves are the underlying index, not option premiums. Statistical tests here are descriptive checks on a few hundred non-overlapping observations from one market and one era. Nothing on this page is investment advice.