Expiry straddle — real NSE daily premiums

Long ATM CE + PE, held from k trading days before expiry to the expiry close · 275 ATM entries · 2021-11-18 .. 2026-10-16 · 1 lot = 65

Headline number below is the 5-DTE row: +353,720 INR on 275 trades. Read the caveats before believing it — its median trade is exactly 0.00% and the mean is carried by a right tail (99th percentile +195%, 5th percentile −88%). It is published because it is the only cell that is positive, not because it is an edge.

Drawdown is measured on cumulative trade P&L against an assumed ₹5,00,000.00 of capital, one lot at a time, no compounding.

Every entry offset

Days before expiry	n	Mean %	Median %	Win %	Long P&L	Worst	Best
5	275	6.74%	0.00%	48.0%	₹3,53,720.25	-99.2%	267.6%

5 DTE, by year

Year	n	Mean %	Win %	Long P&L
2021	7	-33.7%	14%	-₹42,705.00
2022	52	7.7%	46%	₹61,480.25
2023	52	18.5%	48%	₹1,44,137.50
2024	52	11.6%	50%	₹1,00,129.25
2025	54	-9.8%	41%	-₹1,00,649.25
2026	58	11.2%	59%	₹1,91,327.50

How this was measured

Source: data/fyers/options_daily_nse/NIFTY/ — 434 expiry files of real NSE daily CE/PE premiums. Nothing here is synthesised.
Strike = nearest 50 to spot on the entry date. Spot comes from the option file's own underlying_price where populated (only from ~2024-07), otherwise from data/fyers/index/NIFTY50_daily.csv. 832 of 275 entries used the index fallback.
settle_price is deliberately NOT used: on the expiry date that column holds the underlying's settlement for both legs (identical CE and PE values), so it is not an option price. close_premium on expiry day is the intrinsic value and is used instead.
Close-to-close prices, no slippage. Costs modelled only on the short side (brokerage 20/order, STT 0.1% of sell premium, exchange 0.035%, SEBI, GST, stamp).
Lot 65 throughout; NIFTY lots were larger in earlier years, so INR totals are not era-exact while percentage returns are unaffected.
Coverage starts 2021-11 because that is where the index daily file starts — not 2019.