SEBI's August 2026 behaviour study followed individual F&O traders quarter by quarter. In the last quarter of FY26 about twelve lakh of them stopped trading, and more than eight in ten of those had lost money in the quarter before. Look at the same study from the other side and a second group appears: traders whose losses had passed ten lakh rupees in total, nearly nine in ten of whom kept trading. Losses push some people out and keep others in. This lesson reads what the study measured, what it did not, and where chasing a loss fits. It describes groups and says nothing about any one trader. Every study figure carries its table, and every modelled figure is labelled as a model.

3By the end you can

  1. Say what SEBI's behaviour study measured after a loss, and what it did not.
  2. Work out the gain needed to recover a loss, and why it exceeds the loss.
  3. Explain why a high win rate and a bigger stake after a loss do not make a plan.
An old illustrated book lying open on a wooden desk

The idea, plainly

Why do 9 out of 10 F&O traders lose money?

Nine in ten is the share of individual traders with a net loss in a year: 87.7% in FY26, in SEBI's profitability study. The behaviour study explains the pattern behind the count. Almost all individual traders mainly buy options. They trade often compared with their capital. Their losing quarters are more frequent and larger than their winning ones, and traders who lose tend to lose again. Costs come out of every trade, win or lose. None of this is a verdict on one trader. It describes how a group behaves. The question page has the short answer, and module F03 reads the companion study of losses.

What chasing a loss means

Chasing a loss is raising the stake, or the pace of trading, to win back what was lost. A trader on a bad run puts on a bigger position, trades on more days or moves to cheaper options with bigger swings, because only a big swing gets the account back to where it began. My own check is simple: if the size of a stake depends on the last result, that is chasing, whatever it is called.

Read one more paragraph, 56 words

The study does not see stake sizes after a loss. Its data is quarterly and yearly totals. What it can see is who kept trading after losing, and who lost again. That is persistence. Chasing can sit behind it, along with hope, habit, and an account that cannot step back. The study does not say which.

What SEBI found

Six findings are worth carrying, each from a named table of the study.

  1. Losses repeat. Of continuing traders who lost in each of the two years before, about nine in ten lost again, in FY24, FY25 and FY26 alike.
  2. A big loss did not mean leaving. Close to nine in ten traders with losses above ten lakh rupees in FY22 to FY24 were still trading in FY25 and FY26, and more than nineteen in twenty of them lost again.
  3. Leaving follows losing. More than eight in ten of the traders who stopped in a quarter had lost in the quarter before.
  4. Losses outweigh gains. More than eight in ten trader quarters lost money, and the median loss in a losing quarter was more than twice the median gain in a winning one.
  5. Intensity. Traders active on more than a hundred days in a year were fewer than half of all, and took almost all of the turnover and nearly nine tenths of the losses.
  6. Capital cuts two ways. Small accounts lost more often. Larger accounts lost less often, and much more per loser.

The most active traders were fewer than half of all and took almost all of the losses

Traders active on more than 100 days in a year were 42 percent of the sample but 94 percent of turnover and 87 percent of lossesShare of traders 42; Share of turnover 94; Share of losses 87.Traders active on more than 100 days, percent of allShare of traders42%Share of turnover94%Share of losses87%
SEBI's random sample of individual traders, which SEBI calls indicative. The group is traders active on more than a hundred days in the year. A link between activity and loss, not proof that activity causes it. Source: SEBI, Trading Behaviour of Individual Traders in the Equity Derivatives Segment (FY25 to FY26), Aug 2026, Table 18 (random sample, indicative). As of 20 Aug 2026.

After a big loss most traders kept trading, and most lost again

After a big loss most traders kept trading and most lost again, and among traders who stopped most had lost the quarter beforeLost again after 2 losses 90; Kept trading, 10L+ loss 88; 10L+ losers, loss again 95; Quitters who had lost 86.Percent of traders in each group; L is lakh of rupeesLost again after2 losses90%Kept trading,10L+ loss88%10L+ losers, lossagain>95%Quitters who hadlost86%
Population data of individual traders. First bar: lost in each of the two years before, lost again in FY26. Second and third: cumulative loss above ten lakh rupees in FY22 to FY24, then kept trading and stayed loss making in FY25 to FY26. Fourth: of those who stopped trading in FY26 Q4, the share who lost the quarter before. Source: SEBI, Trading Behaviour of Individual Traders in the Equity Derivatives Segment (FY25 to FY26), Aug 2026, Tables 14 and 24, sections 5.23 and 5.24. As of 20 Aug 2026.
Read 2 more paragraphs, 119 words

One more finding shows the cost. Of the traders who lost in FY22 to FY24, more than three quarters held an equity portfolio worth less than a quarter of what they had lost in derivatives, and fewer than one in five held more than they had lost. The study links the two and does not claim cause, since it cannot see other savings or debts. Still, the money that went into losses was not there to compound in shares.

None of these says what a trader should do. They say what groups of traders did. A study of outcomes can show that most who lose keep going. It cannot show whether stopping would have helped any one of them.

The seller's shape: win often, lose big

The study splits traders by how they trade. Those who mainly sold options had the lowest share of loss makers, fewer than half of them, against nine in ten for options buyers. But when a seller lost, the average loss was more than eleven times that of a mainly buying trader. This part comes from SEBI's random sample, which it calls indicative. A model on a dated Nifty chain shows why the two facts sit together. Take one lot of a weekly Nifty 23,500 put, sold for Rs 29.25 a unit at a fair price. In the model the seller wins about 88 weeks in a hundred. The 3 or so worst weeks make up 57% of the losses. Module F13 walks the same shape on a dated week. This is not a case for or against selling. It is the reason a win rate says little by itself, and the reason a loss can be many times the premium taken in.

Model, not market data

A seller wins most weeks, and a few large losses give it all back

Model, not market data. Modelled weeks for the seller of one Nifty put: a green bar of many small wins and a red bar of a few large losses, the same lengthProfit: 88 weeks in 100 1,65,742; Loss, all 12 weeks in 100 −1,65,931; of which up to 3x premium −9,931; of which 3 to 10x premium −60,958; of which over 10x premium −95,043.Per 100 weeks, one lot, before costs; L is lakhProfit: 88 weeksin 100+1.66LLoss, all 12weeks in 100−1.66Lof which up to 3xpremium−0.10Lof which 3 to 10xpremium−0.61Lof which over 10xpremium−0.95L
Each bar is the total in rupees per hundred weeks for one lot. Prices are fair, so nobody starts with an edge. SEBI finds the same shape in the data: fewer loss makers among sellers, far larger losses when they lose. Inputs: one lot of 65 units of the Nifty 23,500 put sold at its settlement premium of Rs 29.25 a unit on 16 Jun 2026, expiry 23 Jun 2026; the forward 23,983.33 and the option's own implied volatility 13.24% over 7 calendar days; the index at expiry follows a lognormal curve around the forward; 100 weeks; before costs and interest. Source: RupeeCase model, inputs from NSE end of day data. As of 16 Jun 2026.

Worked example: what chasing does to one account

Start with the arithmetic of getting back. A loss is a percentage of the old balance, and the gain to recover it is a percentage of the new, smaller one. Take an account of Rs 5,00,000. After a loss of one fifth it holds Rs 4,00,000, and recovering the Rs 1,00,000 needs a gain of 25% on what is left.

LossGain
One tenth11.1%
One fifth25%
Three tenths43%
One half100%
Three quarters300%

Model, not market data

Same odds, bigger stake: the chance of losing half the account climbs from nothing to near certainty

Model, not market data. Chance of losing half the account in 250 trades for a trader with a near zero edge, by risk per trade, from almost none at one percent to nearly certain at twenty percentRisk 1% a trade 0.0%; Risk 2% a trade 7.6%; Risk 5% a trade 60.9%; Risk 10% a trade 89.9%; Risk 20% a trade 98.9%.Paths that hit the ruin level, percent of 2,000Risk 1% a trade0.0%Risk 2% a trade7.6%Risk 5% a trade60.9%Risk 10% a trade89.9%Risk 20% a trade98.9%
A trader with a win rate and payoff that nearly break even before costs. Each bar is a separate run with the same seed, at the risk per trade on its label. Raising the stake after a loss moves a trader to the right on this chart. The inputs line gives the five percent run. Inputs: win rate 45%, average win 1.20 times the average loss, 5% of the account risked on each trade, ruin at a 50% loss of the starting account, 250 trades, 2,000 simulated paths, seed 1. Source: RupeeCase model, seeded simulation. As of 17 Oct 2026.

Model, not market data

At a bigger stake the middle path halves, and only a lucky tail comes back

Model, not market data. Account as a multiple of its start across 250 trades risking five percent a trade: the median path halves, and only a thin upper band ends above the start5th percentile: from 1 at 0 to 0.477 at 250, high 1 at 0, low 0.477 at 250. Median: from 1 at 0 to 0.496 at 250, high 1 at 0, low 0.496 at 250. 95th percentile: from 1 at 0 to 2.39 at 250, high 2.56 at 238, low 1 at 0.5th percentileMedian95th percentile0.501.001.502.002.50050100150200
The 5th, median and 95th percentile paths of the same trader at five percent risk per trade. The upper band is why chasing feels possible: some paths do recover. Most do not. Inputs: win rate 45%, average win 1.20 times the average loss, 5% of the account risked on each trade, ruin at a 50% loss of the starting account, 250 trades, 2,000 simulated paths, seed 1. Source: RupeeCase model, seeded simulation. As of 17 Oct 2026.
Read 3 more paragraphs, 216 words

The deeper the hole, the more than proportionally harder the climb. A trader who answers a loss of one fifth by raising the stake from Rs 10,000 a trade to Rs 50,000 is betting that the next few trades will be good ones. Now run the same trader through a model. Win rate and payoff are set so that the edge before costs is almost zero. Only the stake changes. Module F29 covers sizing in full.

Risking one percent of the account a trade, the chance of losing half the account in 250 trades is 0%. At two percent it is 7.6%. At five percent it is 61%, at ten percent 90%, and at twenty percent 99%. Nothing about the trader's skill changed. The line chart shows why it still feels possible: at five percent of the account a trade, the 95th percentile path ends at 2.39 times the start, while the median path ends at 0.50.

The model is a tidy picture. Real losses come with costs and gaps, and the study's traders are in the data, not in a simulation. The point is the direction. Raising the stake after a loss does not change the odds of the next trade. It changes how much the odds can take.

The common mistake

The belief: I will trade bigger until I am back to even, and then I will stop. Three things say otherwise. The arithmetic of recovery asks for a bigger gain than the loss. The model shows the chance of ruin climbing with the stake while the edge stays the same. And SEBI's data shows that big losses were followed mostly by more trading and more loss, not by a clean stop at even. A plan made after a loss is a plan made by the part of the mind that wants the loss undone. A fair test of any plan is whether the stake rule was written before the first trade, in terms of the account, and not after the last result.

Key takeaways

  • SEBI's study shows persistence in groups. It does not see stake sizes after a loss.
  • Most traders with big losses kept trading, and most of them lost again.
  • Winning back a loss needs a bigger percentage gain than the loss itself.
  • A seller can win most weeks and still lose over time.
  • A bigger stake after a loss raises the chance of ruin, not the trade's odds.
  • Write the stake rule before trading, in terms of the account.

The term this section turns on: Persistence

The term this section turns on: PersistenceThe term this section turns on: Persistence. Persistence: The share of a group that keeps trading, or keeps losing, from one period to the next. SEBI measures it in groups.PersistenceThe share of a group that keeps trading, orkeeps losing, from one period to the next.SEBI measures it in groups.
From the module glossary; every term is in the Learn glossary too. Source: Module 12.9, Behaviour in F&O: loss chasing and what SEBI's behaviour study found.

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