Monday 2 Feb 2026. The Nifty closed at 25,088.40. Eight weeks later, on 30 Mar 2026, it closed at 22,331.40, down 10.99%, inside the one deep fall in the file. A Nifty put struck at 23,000 for the 30 Mar expiry cost 51.45 a unit on 2 Feb, Rs 3,344 a lot, and was worth 668.60 at expiry. The put struck at 22,000 cost 27.75 and expired worthless. Same expiry, same fall, two strikes 8.3% and 12.3% below the index. That gap is what tail hedging is about. Dated NSE data, nothing live, nothing a view on the next fall.
3By the end you can
- Say what a tail hedge pays, what it costs and why the strike matters more than the month.
- Compare the same strike on a short and a long expiry by premium, vega and daily decay.
- Read a protective put at expiry and the yearly drag of buying one every month.
The idea, plainly
What is a tail hedge?
A tail hedge pays in rare, large falls and costs a premium in all the other months. The simplest version is a bought put struck well below the index, which asks for no margin and cannot lose more than its price. Far puts carry a higher implied volatility than puts near the money, because buyers pay extra for crash protection. On the 10 Jun 2026 chain, the 22,000 put of the 30 Jun expiry is priced at 17.59% against 15.03% at the money.
The fall went through one strike and stopped short of the other
Read one more paragraph, 56 words
Tail hedges are insurance. Most months the insurance expires unused, and the cost is real: every rupee of premium is a rupee off the portfolio. A hedge that is too far out is cheap and pays only in a collapse. One too close pays more often and costs more. There is no setting that is free.
Why look at long dated options?
A longer option costs more in total but not much more per day, and it has more vega: it gains more when volatility rises, which is what happens in a fall. It also saves the roll. Beyond the weekly and three monthly expiries, NSE lists Nifty 50 options on three quarterly expiries and the next eight half yearly ones, in June and December (NSE circular NSE/FAOP/73928 of 28 Apr 2026). My data set holds no chain beyond 57 days, so the longest example here is 48 days, against 20 for the monthly put it is compared with.
The long put costs three times as much, decays no faster, and has more than twice the vega
The maths: why a long dated put decays slowly
Black 76 put, same strike:
at the money, price grows
with sqrt(T)
far out, price grows faster
theta a day: near flat in T
vega = e^(-rT)*F*pdf(d1)
* sqrt(T) * 0.01
vega grows with sqrt(T)
Cover D days, rolling n puts
of life L: n * premium(L)
One put of life nL:
premium(nL)
At the money, premium rises with the square root of time, so a put with a life 2.4 times as long would cost well under 2.4 times as much. Here it cost 3.1 times as much, because a far strike gains from extra time faster than the square root rule says. Rolling cheap short puts instead buys cover at a lower cost per day, but each new put is struck off the new index, so the protection resets after a fall.
Read 2 more paragraphs, 144 words
The 22,000 put of 28 Jul 2026 cost 122.70 a unit against 39.50 for the 30 Jun put, 3.1 times as much, on the same day. Its daily decay was about the same, 3.58 against 3.81 rupees a unit, because the extra time is spread over many more days. Its vega was 20.40 against 8.67 rupees a unit for each vol point. For one lot, a rise of five vol points would add about 6,631 rupees to the longer put and 2,819 to the shorter, with nothing else changing.
Two more differences. The longer put has a higher delta, so it moves more with the index from the first day. And it has more days to run after a fall, so it keeps value that a monthly put loses at expiry. The cost is capital held for longer, and a chain with less trading.
What sets the strike?
Distance is a poor yardstick when expiries differ. The 22,000 put is 5.2% below the index on the 30 Jun expiry and the same distance on the 28 Jul expiry, yet its delta is 0.09 on the first and 0.16 on the second. Delta is a rough reading of how likely the market thinks the put is to finish in the money, so the longer put, at the same distance, is the more likely to pay. Reading by delta puts two expiries on a common footing.
Read one more paragraph, 57 words
The extra volatility of the far strike is the skew. The 22,000 put is priced at 17.59% on the 30 Jun expiry against 15.03% at the money, a premium of 2.56 vol points for being far out. A tail hedge pays that skew every time it is bought. The skew lesson reads it as a risk reversal.
Worked example: the 2026 fall, strike by strike
What did each strike do at expiry?
The 30 Mar 2026 expiry settled on the index close, 22,331.40, the rule for expiries before 7 Dec 2026. The 24,000 put, 4.3% below the index, cost 130.25 and finished 1,538.35 better. The 23,000 put cost 51.45 and finished 617.15 better, Rs 40,115 a lot. The 22,000 put cost 27.75 and lost the premium, Rs 1,804 a lot.
The put sets a floor: a fall stops costing at the strike, less the premium
Read 2 more paragraphs, 102 words
The 23,000 put capped the loss on a long future at 9.3% of its value, premium included, against 11.7% for the future alone. It did not remove the loss. A put struck below the index leaves the first slice of any fall with the holder, which is the price of a cheaper hedge.
One window is a worked example, not a track record. The file has one deep fall in fifteen months. A longer record would hold other falls, some deeper, some quicker, and some that reversed before expiry and left a put worth less than it had been a week earlier.
What does carrying the hedge cost?
Suppose a holder buys the 22,000 put of the 30 Jun expiry every month at the 10 Jun 2026 price. That is 30,810 rupees a lot over twelve months, about 2.04% of the lot's value a year. A put with 20 days left covers about two thirds of a month, so a full month of cover at this strike would cost more. Prices change from month to month, and the real cost can be higher or lower. The figure shows the size of the drag at one dated price, and why it should be compared with what the holder thinks a fall would cost.
Model, not market data
The seller gains small most of the time. The hedge buyer is the mirror.
Read 5 more paragraphs, 332 words
Compare it with what a fall costs. A holder with one lot of the future who sat through the 2026 fall unhedged lost 11.7% of its value in eight weeks. The hedge bill for a year at the dated price is 2.04%. Whether the second is worth paying to soften the first is a choice about risk, not arithmetic. Cash, a smaller position or a collar are other ways to a similar result, each at a different price.
In the model, the buyer of the 21,500 put of 28 Jul 2026 wins about 1 repeat in 10, and the average win is around 9 times the premium paid. Real falls cluster in a few months, so a run of unlucky repeats can last longer than a model with independent draws shows.
The far chain is also thin. On 10 Jun 2026 the 22,000 put of the 30 Jun expiry traded 48,904 contracts with 40,85,400 units of open interest. The same strike on 28 Jul traded 8,666 with 11,34,120. A settlement price at a strike that did not trade is set by the exchange's own method and not by a trade, so the figure I see there is a reference and not a price I could deal at.
Rolling has its own mechanics. Each roll pays a new premium and sets the strike off the index of that day. After a fall, the next put is struck lower in rupees, so a second fall is covered from the new, lower base. A hedge held for several months saves the rolls and keeps its strike. On 10 Jun 2026 it also cost more for each day of cover, and it ties up capital in a contract that trades less.
Collars and put spreads cut the carry by selling part of the protection back. The collar lesson shows how, and the lesson on hedging with Nifty puts sizes the lots. Both give up some of the tail to pay for the hedge.
The common mistake
The claim: Buy a far out of the money put every month and your portfolio is protected from crashes.
Protected from what size? In the 2026 fall of 10.99%, the 23,000 put, 8.3% below the index, paid Rs 40,115 a lot, and the 22,000 put, 12.3% below, lost its Rs 1,804. The strike set the answer. The monthly bill at one dated price is 2.04% of the lot's value a year. A put every month is a cost with a payout that depends on strike, timing and a fall that may not come. Say how far down the cover starts, and what it costs to hold.
Key takeaways
- A tail hedge pays in rare large falls and costs a premium otherwise.
- In the 2026 fall, one strike paid many times its cost and another expired worthless.
- A long put costs more in total but decays about the same each day.
- A longer put has more vega: it gains more when volatility rises.
- A put every month is a yearly drag at a dated price.
- Far chains are thin: settlement figures there are references, not prices.
2 terms this section turns on
Before you trade. Derivatives trading carries a high risk of loss. SEBI's study "Profitability of Individual Traders in the Equity Derivatives Segment (FY25 to FY26)" (August 2026) found that 87.7% of individual traders in equity F&O made net losses in FY26, with aggregate net losses of Rs 91,685 crore. Read the SEBI study