Every Indian trading group repeats the same handful of beliefs. Almost nobody checks them. We have 6.5 years of NSE data and a research process built to falsify things, so here are the answers, including the ones that were uncomfortable for us. Each test shows its sample size, the full distribution and a year-by-year breakdown, because a statistic that lives in one market episode is not a pattern.
MOSTLY TRUE ON AVERAGE, BUT THE TAIL IS BRUTAL · 1,015 observations
Across 1,015 results events since 2024, the at-the-money straddle cost 4.8% on average and the stock actually moved 4.6%.
| Average straddle cost (option buyer pays) | 4.77% |
| Average actual move to expiry | 4.60% |
| Average seller profit per event | +0.17% |
| Share of events the seller won | 60% |
| Median seller profit | +0.87% |
| WORST single seller loss | -24.4% |
| Losses bigger than 3x the average win | 42 events |
Year by year (the fragility check: a result that lives in one year is not a pattern)
| 2024 +0.39 | 2025 +0.15 | 2026 +0.02 |
The seller wins often and small, and loses rarely and large: the classic insurance payoff. One event at the wrong end of the distribution erases many wins, and this sample covers only 3 years without a severe market shock. Costs, slippage and margin are NOT deducted here; they subtract from the seller's edge.
BUSTED · 338 observations
Over 338 sampled days since 2020, the highest-delivery fifth of the F&O universe returned +2.07% over the next month versus +2.27% for the lowest fifth.
| Top delivery quintile, next 21 days | +2.07% |
| Bottom delivery quintile | +2.27% |
| Difference | -0.20% |
| Universe average | +1.89% |
| Days the top quintile won | 48% |
Year by year (the fragility check: a result that lives in one year is not a pattern)
| 2020 +0.76 | 2021 +0.54 | 2022 +0.04 | 2023 -1.47 | 2024 +0.43 | 2025 -0.55 | 2026 -1.74 |
Delivery percentage tells you ownership genuinely changed hands, which is real information about the past. It does not, on its own, tell you what happens next.
BUSTED · 1,534 observations
Across 769 ban entries and 765 ban exits since 2020, the next five days looked almost exactly like the market: +1.10% and +0.72% on average.
| After entering the ban, next 5 days | +1.10% |
| After exiting the ban, next 5 days | +0.72% |
| Positive after entry | 55% |
| Worst single outcome after entry | -40.7% |
| Spread of outcomes after entry (sd) | 7.0% |
Year by year (the fragility check: a result that lives in one year is not a pattern)
| 2020 +1.12 | 2021 +0.80 | 2022 +0.31 | 2023 +2.15 | 2024 +1.13 | 2025 +0.50 | 2026 +1.32 |
Our pre-registered study of 1,028 ban episodes found the 'forced unwind' story was actually wrong-signed: the crowd tends to persist through the ban. The real cost of a ban is practical, not directional: you cannot open or add positions, only reduce.
BUSTED · 1,710 observations
Since 2020, the month after the highest-VIX days averaged +3.68% versus +1.49% after the calmest days.
| Next 20 days after HIGH VIX days | +3.68% |
| After MEDIUM VIX days | +0.14% |
| After LOW VIX days | +1.49% |
| Positive after high VIX | 76% |
| Spread of outcomes after high VIX (sd) | 6.6% |
Year by year (the fragility check: a result that lives in one year is not a pattern)
| 2020 +5.28 | 2021 +4.04 | 2022 +0.68 | 2024 +6.26 | 2025 +5.92 | 2026 +3.73 |
High VIX days have historically been followed by ABOVE-average returns, not below. But look at the spread: outcomes after high-VIX days are far wider in both directions. The honest lesson is about position size, not about exit.
BUSTED · 1,729 observations
Average daily return in expiry week was -0.010% versus +0.115% in all other weeks. The difference is noise.
| Average daily return, expiry week | -0.010% |
| Average daily return, other weeks | +0.115% |
| Daily volatility, expiry week | 1.23% |
| Daily volatility, other weeks | 1.10% |
| Expiry-week days in sample | 402 |
Year by year (the fragility check: a result that lives in one year is not a pattern)
| 2020 +0.07 | 2021 -0.05 | 2022 -0.27 | 2023 -0.01 | 2024 +0.10 | 2025 +0.03 | 2026 +0.09 |
Our pre-registered expiry study went further: it compared expiry-week reversals against identical trades in placebo mid-month weeks, and the extra effect was statistically indistinguishable from zero (t = 1.57). Expiry feels dramatic because volume is loud, not because returns are different.
BUSTED · 1,729 observations
Best day averaged +0.205% and worst -0.062% per day since 2020, a gap far smaller than a single day's typical swing.
| Monday | -0.062% (58% positive, n=346) |
| Tuesday | +0.201% (59% positive, n=347) |
| Wednesday | +0.205% (62% positive, n=346) |
| Thursday | +0.071% (54% positive, n=345) |
| Friday | +0.013% (51% positive, n=345) |
Wednesday came out best and Monday worst in this sample, but daily returns vary by roughly 1% around these averages: the weekday pattern is a rounding error next to the noise, and it changes sample to sample.