The behavioural biases that cost Indian retail traders money

The short answer

The losses that show up in Indian retail accounts are not random. They trace to a small set of documented behavioural biases: the disposition effect (selling winners early, holding losers), loss aversion (a loss hurts about 2.25 times as much as an equal gain), overconfidence and the overtrading it drives, recency and the hot-hand fallacy, lottery preference (chasing cheap far out-of-the-money options), herding, anchoring, sunk-cost get-evenitis, and confirmation bias. Each has a source in behavioural finance and a clear signature in the SEBI data. None is cured by awareness. Each is disarmed by a rule that removes the moment of discretion where it acts.

The base rate is the place to start, because it is the number every bias here helps produce. In its July 2025 study of the equity derivatives segment, SEBI reported that over 91 percent of individual traders were net loss-making in FY25, with aggregate individual net losses of about 1,05,603 crore rupees. The earlier September 2024 study, covering FY22 to FY24, put the figure at 93 percent of more than one crore individual traders, aggregate losses above 1.8 lakh crore, and only about one percent clearing more than one lakh rupees of profit. A failure rate that consistent across years and samples is not bad luck distributed at random. It is behaviour, repeated, and behaviour has structure.

This page is a catalogue. For each major bias it states three things: what the bias is, grounded in the behavioural-finance research that named it; how it specifically shows up in Indian retail trading; and the structural fix that disarms it. The order runs from where biases strike in perception through the holding decisions they corrupt to the activity they inflate. Read it as a reference, not a lecture. The companion piece on trading psychology at scale covers how an institution builds the rituals that hold these fixes in place; this page is the underlying map of what those rituals defend against.

The map: where each bias strikes

Biases are easier to defend against when you know where in the decision they live. Some distort what you perceive and predict before a trade exists. Some corrupt the holding decision once a position is open, which is where the disposition effect and its relatives do their damage. Some simply inflate how often you act, turning a small negative edge into a large one through sheer frequency. The map below groups the catalogue that way.

The bias map: perception, holding, activity Ten behavioural biases sorted into three zones. Perception and prediction: recency and hot-hand, overconfidence, confirmation bias, anchoring. The holding decision: disposition effect, loss aversion, sunk-cost get-evenitis. Activity and frequency: overtrading, lottery preference, herding. Where each bias strikes A trade flows left to right: what you predict, what you hold, how often you act. 1 · PERCEPTION & PREDICTION Recency & hot-hand fallacy Overconfidence Confirmation bias Anchoring Distort the estimate before the position exists. 2 · THE HOLDING DECISION Disposition effect Loss aversion Sunk-cost get-evenitis Corrupt when to exit, once you are in. 3 · ACTIVITY & FREQUENCY Overtrading Lottery preference Herding Inflate how often, and how wildly, you act.
Group by where the bias acts, not by name. The same fix rarely covers two zones. A rule that protects the holding decision, both exits placed at entry, does nothing for a distorted prediction; a rule that governs prediction, write the thesis before opening any feed, does nothing for how often you trade. The catalogue below follows these three zones in order.

The holding decision: the disposition effect

What it is. Shefrin and Statman named the disposition effect in 1985: the disposition to sell winners too early and ride losers too long. It is the single most studied bias in retail behaviour, and prospect theory explains it precisely. Value is felt relative to a reference point, usually the entry price, and the value function is concave in gains and convex in losses. A modest gain sits in a region where locking it in feels satisfying and holding for more feels greedy, so the winner is sold. A modest loss sits in the loss-averse region where crystallising the loss is disproportionately painful, so the loser is held in the hope it returns to the reference point.

The prospect-theory value function and the disposition effect Felt value plotted against objective gain or loss relative to entry. The gain arm is shallow and concave; the loss arm is steep and convex, about 2.25 times as steep. The asymmetry drives selling winners early and holding losers. Why losers are held and winners are sold Prospect-theory value function, felt value against gain or loss from entry. Gain Loss Felt value + Felt value − entry (reference) a gain of X feels like +v an equal loss feels like about 2.25v loss arm is steeper: lambda about 2.25 (Tversky and Kahneman, 1992)
The asymmetry is the whole mechanism. Because the loss arm is roughly 2.25 times as steep as the gain arm, the pain of realising a loss dwarfs the pleasure of realising an equal gain. Holding the loser postpones that pain; selling the winner banks a certain, if smaller, pleasure. The disposition effect is simply this curve, acted out one trade at a time.

How it shows up in Indian retail. The winner is booked at the first green candle, often well before the level that the plan said to target, because the certain small gain feels safer than the uncertain larger one. The loser is not cut at the stop. It is averaged down to lower the break-even, or rolled into the next weekly expiry, or simply watched in the hope it recovers to the entry. In derivatives this is especially costly, because the loser that is held is decaying on time value while the trader waits for a reference point that the market has no obligation to revisit.

The structural fix. Pre-commit both exits at entry. Write the stop and the target before the position is live, place both, and do not modify them mid-trade. The disposition effect acts at the exit decision, so the fix is to have made that decision before the emotion is present. This is exactly why a mechanical exit wrapper matters: the value of pre-setting a stop and target is not that the levels are clever, it is that the discretion has been removed from the one moment where the bias would otherwise win.

The holding decision: loss aversion and get-evenitis

What it is. Loss aversion is the engine underneath the disposition effect. Tversky and Kahneman, in their 1992 formulation of cumulative prospect theory, estimated the loss-aversion coefficient at about 2.25: losses are weighted roughly 2.25 times as heavily as equal gains. This is not the same as risk aversion. Loss aversion is a kink at the reference point that makes people risk-averse over gains but risk-seeking over losses, which is why the same trader will grab a small sure profit yet gamble to avoid booking a small sure loss.

Get-evenitis is the sunk-cost version of the same pull. Once a trade is under water, the money already lost feels like it must be recovered before the position can be closed, so the trader holds, averages down, or adds risk purely to reach break-even. It converts a bounded, planned loss into an open-ended one, and it usually adds size at the worst possible moment. Sunk cost is a logical error compounded by loss aversion: money already gone is irrelevant to whether the position deserves to stay open, yet it dominates the decision.

How it shows up in Indian retail. Loss aversion is why so many retail traders gravitate to structures with a high hit rate and a small, comfortable win, even when a lower-hit-rate structure would carry a better expectancy, because frequent small wins minimise the number of painful loss experiences. Get-evenitis is the account that doubles the option position after it has halved, or the intraday trade converted to a delivery hold the moment it moves against the plan, so the loss need never be admitted.

The structural fix. Judge every structure on expectancy, the probability-weighted rupee outcome, not on how often it wins, and let the stop be defined by the level that invalidates the idea rather than by the entry price. Money already lost is excluded from the decision by rule. This is the discipline that the upstream work of setting an edge, an invalidation level and a size is meant to protect, and that upstream work is exactly what the method we teach is built around.

Activity and frequency: overconfidence and overtrading

What it is. Overconfidence is the systematic overestimation of one's own skill, information or prediction accuracy. Its most expensive consequence in markets is overtrading. Barber and Odean made the case directly in their 2000 paper, titled Trading Is Hazardous to Your Wealth, studying 66,465 households at a large discount broker from 1991 to 1996. The households that traded most turned over their portfolios heavily and, after costs, earned net annual results of about 11.4 percent against a market that returned about 17.9 percent over the same window, while the average household did somewhat better than the most active. The mechanism they identified was overconfidence: traders who believed their information justified frequent action paid for that belief in transaction costs and worse timing.

How it shows up in Indian retail. The India base rate is the sharpest possible illustration. SEBI's July 2025 study found over 91 percent of individual equity-derivatives traders net loss-making in FY25, with individual net losses of about 1,05,603 crore rupees; the September 2024 study found 93 percent loss-making across FY22 to FY24. Trading is a domain where feedback is delayed, noisy and confounded, so a win on luck is easily filed as a win on skill, and the self-model calibrates to the best recent outcomes rather than to the full distribution. High activity then multiplies brokerage, taxes and slippage on top of a base rate that is already negative.

The structural fix. Separate process from outcome. Grade every trade on whether it followed the plan, entry at the planned level, stop honoured, size correct, independently of whether it made or lost money. An A-grade losing trade is a good trade; a C-grade winning trade is a warning. Because the base rate is what it is, the honest question is never how much a method makes but whether the process was clean, and grading the process is what breaks the feedback loop that overconfidence feeds on.

Activity and frequency: lottery preference and skewness seeking

What it is. Prospect theory has a second edge beyond loss aversion: people overweight very small probabilities. That overweighting produces lottery preference, or skewness seeking, the willingness to overpay for a small chance of a large, positively skewed payoff. It is the same impulse that keeps lottery tickets and low-priced stocks in demand: a tiny outlay, a slim probability, and a payoff large enough that the imagination fixes on it rather than on the odds.

How it shows up in Indian retail. The clearest signature is the demand for cheap, far out-of-the-money options. SEBI has documented retail traders placing bets on strikes several percent away from the current price, where the premium is small and the probability of profit is low. A far out-of-the-money weekly option costs a few rupees and can multiply many times over if a large move arrives before expiry, which makes it feel like a cheap ticket rather than what it is, a low-probability bet decaying quickly on time value. The low-priced stock exerts the same pull for the same reason.

The structural fix. Price the ticket. Expectancy accounting forces the bet to be evaluated as probability multiplied by payoff minus cost, and most cheap lottery tickets are negative on that arithmetic once the true odds are used rather than the imagined payoff. The illustrative table below shows the shape of the reasoning: the numbers are hypothetical and label themselves so, but the method of deriving an expected value rather than staring at the payoff is the transferable point.

Illustrative expectancy of a cheap lottery-style bet (hypothetical figures, for method only, not a claim about any real instrument)
What the trader seesWhat expectancy accounting addsResult
Ticket costs 5 rupeesCost is certain and paid up front−5 always
Could pay 100 rupeesOnly if a large move arrives before expiry+95 net, rarely
The move looks possibleAssign it an honest probability, say 4 in 1000.04 chance
Expected value per ticket0.04 times +95, plus 0.96 times −5about −1 rupee

The point is not the specific figures, which are invented for the illustration. It is that the payoff the mind fixes on, the 100, is not the number that governs the decision. The number that governs it is the expected value, and when the true probability is small enough the expected value is negative even though the payoff is large. Skewness seeking is the habit of looking at the payoff and ignoring the probability.

Perception and prediction: recency and the hot-hand fallacy

What it is. Recency bias is the over-weighting of the most recent outcomes when estimating what comes next, so the last three or five trades feel far more informative than they are. The hot-hand fallacy is the specific belief that a streak will continue because it is a streak; the gambler's fallacy is its mirror, the belief that a run is due to reverse. Both misread ordinary variance. In a process with a stable edge, streaks of wins and of losses occur regularly through chance alone, and reading a signal into them is reading noise.

How it shows up in Indian retail. Position size drifts with the recent run: it is raised after a few wins, when confidence is high, and cut after a few losses, when nerve fails, despite a stated rule of constant fractional sizing. Because size is largest exactly when a hot streak is most likely to end and smallest exactly when a cold streak is most likely to turn, recency-driven sizing tends to buy high conviction at the wrong times. The trader experiences this as reading the market; it is reacting to the last few outcomes.

The structural fix. Size from a long rolling sample, not from the last few trades, and review sizing on a fixed calendar rather than after each result. A rule that references a rolling window of many trades smooths the recency signal, and a review that happens on a schedule prevents the intra-week drift that recency produces. The trade-journal practice that records size and outcome for every trade is what makes a rolling rule possible in the first place.

Perception and prediction: anchoring and confirmation bias

What it is. Anchoring is over-reliance on a reference number, typically the entry price or a recent high, when making later decisions about a position. The anchor persists even in traders who understand it intellectually. Confirmation bias is the tendency to seek and over-weight evidence that supports a view already held and to discount evidence against it, so conviction grows as agreeing voices accumulate rather than as the case genuinely strengthens.

How it shows up in Indian retail. Anchoring is the trader who bought near a round number and treats a return to that number as the natural exit, even when the chart structure argues for a different level, because the entry price has become a psychologically privileged reference. Confirmation bias is the long position whose holder scrolls only for bullish takes, files a contrary signal as noise, and mistakes the resulting comfort for analysis. The two compound: an anchored target plus a feed full of agreement can keep a broken trade alive well past the point the plan called it wrong.

The structural fix. Set stops and targets by structure, not by the entry price, so the decision is anchored to where the idea is genuinely invalid rather than to an arbitrary number, and write the falsification condition in advance, the specific price or event that would prove the thesis wrong, before opening any external source. Reaching that condition is then a signal to act, not to rationalise. A stop placed just beyond the level that invalidates the idea is structural; a stop placed a fixed distance below the entry is anchored.

Activity and frequency: herding

What it is. Herding is aligning actions with what other traders are doing rather than with independent analysis. Being wrong alone feels worse than being wrong alongside a crowd, so the crowd provides a kind of emotional insurance at the cost of analytical independence. Most people who herd do not experience it as herding; they experience it as conviction, because the shared view feels like consensus evidence rather than social pressure.

How it shows up in Indian retail. The idea arrives from a group, a channel or a trending post rather than from the trader's own process, and the position is sized by how loud the crowd is rather than by the risk plan. The feed supplies both the idea and, through confirmation bias, an endless stream of voices that make it feel validated. When the crowd reverses, the herded position is held longest, because the same social signal that justified entering now justifies waiting.

The structural fix. Generate the thesis before opening any external source. A hard rule that the idea and its invalidation level are written down before any feed, channel or research is opened, with external sources permitted only for price verification and never for idea generation, cuts the herding channel at its root. It is one of the highest-leverage rules available to a retail trader precisely because it changes what a trade is made of, not merely how it is managed.

The catalogue, at a glance

The table below is the reference form of everything above: each bias, the behavioural-finance source that named it, its Indian retail signature, and the structural fix that disarms it. It is deliberately compact so it can be used as a checklist against a real trade log.

The bias catalogue: source, Indian retail signature, and structural fix
BiasSourceHow it shows up in Indian retailStructural fix
Disposition effectShefrin & Statman, 1985Winners booked at the first tick; losers averaged down or rolled to next expiryPlace both exits at entry; do not modify mid-trade
Loss aversionTversky & Kahneman, 1992 (lambda about 2.25)Preference for high-hit-rate structures with small wins over better-expectancy onesJudge structures on expectancy, not hit rate
Sunk-cost, get-evenitisSunk-cost bias, amplified by loss aversionAdding size to a loser purely to reach break-evenStop set by invalidation level; ignore money already lost
Overconfidence, overtradingBarber & Odean, 2000High activity multiplying costs on a negative base rateGrade process separately from outcome
Lottery preferenceProspect theory, small-probability overweightingBuying cheap far out-of-the-money options and low-priced stocksPrice the bet by expectancy before taking it
Recency, hot-handAvailability and streak misreadingSize raised after wins, cut after losses, against a stated ruleSize from a long rolling sample; review on a calendar
AnchoringAnchoring and adjustmentTreating the entry price as the natural exit levelSet stops and targets by structure, not by entry
Confirmation biasMotivated reasoningScrolling only for takes that agree with the open positionWrite the falsification condition in advance
HerdingSocial conformityIdeas and size driven by a channel or trend, not a planWrite the thesis before opening any external source

From bias to fix: the disarming map

Read together, the fixes share a single shape. None of them tries to argue the trader out of a feeling. Each one moves a decision earlier, to a point before the emotion is live, and freezes it there. The diagram maps each bias to the rule that disarms it so the pattern is visible at a glance: awareness on the left, mechanism on the right, and the arrow from one to the other is always the same move, pre-commitment.

Each bias arrowed to the rule that disarms it Biases on the left connect to structural fixes on the right. The exit biases map to pre-set exits and expectancy; overconfidence maps to process grading; lottery preference maps to expectancy accounting; recency maps to rolling-sample sizing; anchoring maps to structural stops; confirmation and herding map to writing the thesis first. The move is always the same: decide earlier THE BIAS THE RULE THAT DISARMS IT Disposition effect Loss aversion Sunk-cost get-evenitis Overconfidence Lottery preference Recency, hot-hand Anchoring Confirmation bias Herding Place both exits at entry;judge structures on expectancy Stop set by the invalidation level Grade process separately from outcome Price the bet by expectancy Size from a long rolling sample Set stops by structure, not entry Write thesis and falsificationbefore opening any feed
Nine fixes, one pattern. Every arrow encodes the same move: take the decision the bias would corrupt and make it before the emotion is present. The exit is decided at entry, the size is decided by rule, the thesis is decided before the feed. Awareness is on the left because it is necessary; the rule is on the right because awareness alone is not sufficient.

The SEBI evidence, mapped to the biases

The base-rate numbers are not just context. Each headline finding is the aggregate footprint of one or more of the biases above, which is why the fixes and the statistics belong on the same page. The table connects the two.

SEBI evidence mapped to the biases it reflects
SEBI findingStudyBiases it reflects
Over 91 percent of individual traders net loss-making; net losses about 1,05,603 crore rupeesFY25 study, July 2025Overconfidence and overtrading against a negative base rate
93 percent loss-making across FY22 to FY24; only about 1 percent cleared over one lakh rupeesStudy, September 2024The full bias cluster, aggregated over three years
Retail bets concentrated on strikes several percent from spot, where probability of profit is lowSEBI derivatives findingsLottery preference and small-probability overweighting
Losses persist year after year at a similar rateFY25 and prior studiesRecency and get-evenitis defeating the intent to stop
What the numbers do and do not say. The SEBI figures are cited here as regulator statistics, facts about a population. They describe how a large group fared; they are not a prediction about any individual, and nothing on this page promises a different outcome. The fixes in this catalogue are designed to reduce specific, named errors. Reducing error is not the same as producing a gain, and no rule here should be read as a claim about profit.

Where this fits

This catalogue is the map; the discipline that uses it is a separate skill. Knowing that the disposition effect lives at the exit does not, by itself, place the exit; that requires a plan, a rule and the habit of following it. The biases sit inside a wider psychology of trading that scales from a single account to an institution, and the fixes here are the atoms that the larger rituals are built from. In the Bharath Shiksha curriculum, thirty volumes across six stages, this material runs from the journaling and sizing discipline of the early stages through the institutional review rituals covered later, always structurally: the aim is never to will a bias away, but to build the protocol that removes the moment where it would act.

Common questions

Frequently asked questions

The costliest cluster is the disposition effect (selling winners early and holding losers), loss aversion (a loss hurts about 2.25 times as much as an equal gain), overconfidence and the overtrading it drives, recency and the hot-hand fallacy, lottery preference (chasing cheap out-of-the-money options and low-priced stocks for a slim shot at a large payoff), herding, anchoring, sunk-cost or get-evenitis, and confirmation bias. Each is documented in behavioural finance and each has a clear signature in Indian retail data. None is fixed by awareness alone; each is disarmed by a pre-committed rule that removes the moment of discretion where it acts.

The disposition effect, named by Shefrin and Statman in 1985, is the tendency to realise gains too early and to hold losses too long. Prospect theory explains it: a small gain sits in the concave region of the value function where locking it in feels good, while a loss sits in the convex, loss-averse region where holding on to avoid crystallising the pain feels rational. In Indian retail it appears as winners booked at the first green tick and losers averaged down or rolled to the next expiry rather than cut. The fix is to set both exits, target and stop, at entry and to leave them alone.

Risk aversion is a smooth dislike of uncertainty across all outcomes. Loss aversion is a kink at zero: Tversky and Kahneman estimated in 1992 that losses are weighted about 2.25 times as heavily as equal gains. That kink makes people risk-averse over gains yet risk-seeking over losses, which is exactly why a trader who would gladly book a small profit will gamble to avoid booking a small loss. It is not that losses are disliked in general; it is that the pain is disproportionate right around the reference point, usually the entry price.

The evidence points the other way. Barber and Odean, studying 66,465 households at a large discount broker from 1991 to 1996, found the most active traded far more yet earned materially less after costs than the market and than quieter accounts, a finding they titled trading is hazardous to your wealth and attributed to overconfidence. In Indian equity derivatives the base rate is starker still: SEBI reported that over 91 percent of individual traders were net loss-making in FY25, with individual net losses of about 1,05,603 crore rupees. Activity multiplies costs; it does not multiply edge.

This is lottery preference, or skewness seeking: people overpay for a small chance of a large payoff, the same impulse that sells lottery tickets. A far out-of-the-money option costs a few rupees and can multiply many times if a big move arrives, so it feels like a cheap ticket. SEBI has documented retail placing bets on strikes several percent away from spot, where the probability of profit is low and time decay is fast. The structure that disarms it is expectancy accounting: price the bet by probability times payoff minus cost, and most such tickets are negative on that arithmetic.

Both over-read short runs. Recency bias is weighting the latest few outcomes far above the base rate, so three wins feel like proof of skill and three losses feel like the system is broken. The hot-hand fallacy is the specific belief that a streak will continue because it is a streak; the gambler's fallacy is its mirror, expecting a reversal because a run feels due. In a system with a stable edge, ordinary variance produces streaks in both directions, so a trader who resizes after every run is reacting to noise. The fix is to size from a long rolling sample and review on a fixed calendar, not after each trade.

You do not fix the feeling; you remove the moment where it can act. Every bias in this catalogue expresses itself at a specific decision point: the disposition effect at the exit, recency at the sizing choice, herding at idea generation, anchoring at stop placement. A structural fix pre-commits that decision before the emotion is live: both exits placed at entry, size drawn from a rolling sample by rule, the thesis written before any external source is opened, the stop set by chart structure rather than by the entry price. The rule holds when willpower does not, which is the entire point.

Get-evenitis is the sunk-cost bias applied to a losing trade: the urge to hold, average down or add risk purely to return to break-even, because exiting would make the paper loss permanent. It is dangerous because it converts a bounded, planned loss into an open-ended one, often by adding size at the worst moment. It compounds with anchoring, since break-even is defined by the entry price rather than by whether the idea is still valid. The disarming rule is that the stop is defined by the level that invalidates the thesis, and money already lost is irrelevant to whether the position deserves to stay open.

Confirmation bias is the tendency to seek and weight evidence that supports a position already held and to discount evidence against it. In practice a trader who is long scrolls for bullish takes, treats a contrary signal as noise, and feels growing conviction that is really just accumulated agreement. It pairs with herding, since social feeds supply an endless stream of confirming voices. The counter is to write the falsification condition in advance: state, at entry, the specific price or event that would prove the idea wrong, and treat reaching it as a signal to act, not to rationalise.

Sources

Where the facts come from

  • SEBI, equity derivatives study (July 2025). Establishes the FY25 base rate: over 91 percent of individual traders net loss-making, aggregate individual net losses of about 1,05,603 crore rupees. sebi.gov.in
  • SEBI, updated study (September 2024). Establishes the FY22 to FY24 figures: 93 percent of more than one crore individual traders loss-making, aggregate losses above 1.8 lakh crore, only about one percent clearing over one lakh rupees of profit. sebi.gov.in
  • Shefrin, H. and Statman, M. (1985). The disposition to sell winners too early and ride losers too long: theory and evidence. Journal of Finance 40(3), 777 to 790. Names and models the disposition effect.
  • Tversky, A. and Kahneman, D. (1992). Advances in prospect theory: cumulative representation of uncertainty. Journal of Risk and Uncertainty 5, 297 to 323. Source of the loss-aversion coefficient of about 2.25 and the value function used in the second diagram.
  • Barber, B. and Odean, T. (2000). Trading is hazardous to your wealth: the common stock investment performance of individual investors. Journal of Finance 55(2), 773 to 806. The 66,465-household study linking overconfidence, overtrading and worse net results.
Educational note. This guide explains behavioural biases and structural ways to reduce the errors they cause. It is not a recommendation to trade or invest, and it is not investment advice. The fixes described are intended to reduce specific decision errors; they do not promise any gain, and no result is implied. Bharath Shiksha is an educational publisher, not a SEBI-registered investment adviser or research analyst.

Related guides

Learn the discipline these fixes require

Recognising a bias is the easy half. Placing the exit, sizing from a rule and writing the thesis first is a trained skill. Bharath Shiksha is a thirty-volume curriculum across six stages, from chart reading through institutional risk process, priced from ₹14,999 to ₹1,49,999, built around the upstream judgement that these fixes protect.

Take the free diagnostic →