Home / Decision Making / Expected Value
Decision scienceIntermediate

Expected Value

Expected value is the probability-weighted average of all the possible outcomes of a decision, the sum of each outcome multiplied by its probability, and it tells you what a trade is worth on average even though any single result will differ.

Quick answer: Expected value is the probability-weighted average of all the possible outcomes of a decision, the sum of each outcome multiplied by its probability, and it tells you what a trade is worth on average even though any single result will differ.

In simple words

Expected value answers one question: if you could take this trade a thousand times, what would you make or lose per trade on average? You list each possible outcome, multiply it by how likely it is, and add them up. A trade with positive expected value makes money over the long run; a negative one loses, no matter how good a single result feels. It does not predict the next trade, which can land anywhere, but it tells you whether the bet is worth repeating, which is the only question that matters over a career.

Purpose

Expected value exists to reduce a messy set of uncertain outcomes to a single comparable number, so a trader can rank very different opportunities on the same scale and know whether an approach makes money on average rather than merely feeling good in the moment.

Professional explanation

The definition and the formula

Expected value is the sum, over every possible outcome, of that outcome's payoff multiplied by its probability. Written compactly it is EV equals the sum of probability times outcome across all outcomes. For a simple two-outcome trade it becomes the win probability times the average win, minus the loss probability times the average loss, since a loss is a negative outcome. The probabilities must be genuine estimates that sum to one, and the payoffs should be net of costs to be honest. The number that results is not a prediction of the next trade; it is the long-run average per trade, the anchor around which real results will vary.

A worked options example

Suppose a Nifty option trade has a 40 percent chance of making Rs 6,000 and a 60 percent chance of losing Rs 3,000. The expected value is 0.40 times 6,000 minus 0.60 times 3,000, which is 2,400 minus 1,800, equal to plus Rs 600 per trade before costs. That positive figure means the trade is worth repeating even though it loses more often than it wins, because the wins are large enough to more than offset the frequent smaller losses. Subtract brokerage, exchange fees, STT and slippage, say Rs 150, and the net expected value is about Rs 450, still positive but thinner, a reminder that costs are part of every honest calculation.

Positive EV does not mean it wins often

A crucial and counterintuitive point is that expected value is silent about win rate. A trade can be positive-EV while losing most of the time, if its rare wins are large, and negative-EV while winning most of the time, if its rare losses are huge. Naked option selling is the textbook trap: an 85 percent win rate can still be negative expected value once the occasional large loss is weighted in. Traders drawn to high win rates for the emotional comfort of frequent small wins routinely accumulate negative-EV positions, because the pleasant hit rate distracts from the payoff asymmetry that actually determines the average.

Why EV needs the law of large numbers

Expected value is a long-run average, and it only manifests over many independent trades. On any single trade you receive an actual outcome, not the average, and that outcome can sit far from the expected value. The law of large numbers guarantees that as trades accumulate, the realised average per trade converges toward the true expected value, but only if the trades are numerous and the edge is real. This is why a positive-EV trader can be underwater for a long stretch and must be capitalised and sized to survive it. Expected value without survival is a promise the market may never let you collect.

Garbage in, garbage out: the estimate problem

Every expected-value number rests on probabilities and payoffs that are estimated, not known. If the probabilities are wishful, the EV will be optimistic by construction, and a trader can convince themselves a losing approach is profitable simply by nudging the win rate. Payoffs, too, assume stops and targets fill at their levels, which gaps break, making realised losses larger than modelled. Honest expected-value work therefore uses conservative, evidence-based probabilities, subtracts realistic costs, and treats the output as a range rather than a point. The formula is arithmetic; the difficulty and the judgement live entirely in the inputs.

EV, sizing and the tail

Expected value tells you whether to take a trade, not how large to make it. A trade can be positive-EV yet carry a tail outcome large enough to threaten the account, and repeating it at full conviction can produce ruin before the average arrives. This is why expected value must be paired with position sizing that caps the worst plausible loss at a small fraction of capital. The Kelly criterion formalises the link, giving the growth-maximising fraction to stake given the edge and odds, but even Kelly is usually scaled down because the inputs are uncertain. Expected value chooses the bet; risk sizing decides how much of it you can survive.

Formula

EV = Σ ( probability × outcome )

For each possible outcome, multiply its payoff (a gain is positive, a loss negative, and best measured net of costs) by its probability, then add the results; the probabilities must sum to one. For a simple win-or-lose trade this is EV = P(win) × average win − P(loss) × average loss. A positive EV means the trade profits on average over many repetitions, though any single result will differ, and the number is only as reliable as the estimated probabilities behind it.

Practical example

Illustrative example (Indian market)

Take the worked case directly: a Nifty options trade has a 40 percent chance of +Rs 6,000 and a 60 percent chance of -Rs 3,000. EV = (0.40 x 6,000) + (0.60 x -3,000) = 2,400 - 1,800 = +Rs 600 per trade before costs. Even though it loses 60 percent of the time, it is worth repeating because the wins are twice the size of the losses. Net Rs 150 of brokerage, STT and slippage and the edge is about +Rs 450 per trade. Over 100 such trades the expected total is roughly +Rs 45,000, but any given block of ten could easily be negative, which is why the position must be sized to survive the losing clusters that reaching the average requires.

A common NSE trap is selling deep out-of-the-money Bank Nifty weekly options that expire worthless perhaps 88 percent of the time for a small premium, while the 12 percent adverse expiry can lose eight to ten times that premium. The comforting win rate hides a negative expected value once the tail is weighted in, and a single volatile expiry, around RBI policy or a budget, can erase months of small credits. The math, not the hit rate, tells the truth.

Advantages

  • Reduces many uncertain outcomes to one comparable number for ranking trades
  • Reveals profitable trades that lose more often than they win, and vice versa
  • Exposes the payoff asymmetry that a seductive win rate hides
  • Provides the objective basis for whether an approach makes money on average
  • Forces costs and the tail into the calculation instead of ignoring them

Limitations

  • Only as accurate as the estimated probabilities and payoffs, which are noisy
  • Silent on how large to size a position, so it needs a separate risk rule
  • A long-run average that may not appear over a small or unlucky sample
  • Assumes stops and targets fill at their levels, which gaps can violate
  • Can be gamed by wishful probabilities that make a bad trade look positive

Why it matters in practice

  • Separates trades worth repeating from those that merely feel good
  • Stops a high win rate from disguising a negative-edge strategy

Common mistakes

  • Confusing expected value with win rate, and preferring frequent small wins
  • Using optimistic probabilities that make the EV positive by construction
  • Ignoring costs, so the modelled EV overstates the real edge
  • Treating the EV as a prediction of the next trade rather than a long-run average
  • Taking positive-EV trades so large that the tail loss ruins the account first
  • Forgetting that a positive EV needs many trades and survival to be realised

Professional usage

Professional traders and quant desks think natively in expected value net of costs, ranking opportunities by expectancy rather than by how often they win. They estimate probabilities conservatively, stress the tail branches, and pair every positive-EV decision with position sizing, often a scaled-down Kelly fraction, so the worst plausible outcome stays a small fraction of capital. They treat the EV as a decision input whose reliability depends entirely on the honesty of its probabilities, and they never assume the long-run average will arrive without the survival to wait for it.

Key takeaways

  • Expected value is the sum of each outcome times its probability, EV = Σ(p × outcome)
  • A positive-EV trade can lose most of the time if its wins are large enough
  • EV is a long-run average, realised only over many trades, so survival is required
  • The formula is easy; the honesty of the probabilities and costs is the hard part

Frequently asked questions

What is expected value in trading?
Expected value is the probability-weighted average of all possible outcomes of a trade: each outcome's payoff multiplied by its probability, summed up. It tells you what the trade is worth on average over many repetitions, so a positive expected value means an approach makes money in the long run even though any single result will differ.
What is the expected value formula?
EV equals the sum of probability times outcome across all outcomes. For a simple win-or-lose trade it is the win probability times the average win minus the loss probability times the average loss. The probabilities must sum to one, and payoffs are best measured net of costs to be honest.
Can you give a worked options example?
Suppose a Nifty option trade has a 40 percent chance of +Rs 6,000 and a 60 percent chance of -Rs 3,000. EV = 0.40 x 6,000 minus 0.60 x 3,000 = 2,400 minus 1,800 = +Rs 600 per trade before costs. It loses 60 percent of the time but is still worth repeating because the wins are twice the losses.
Does a positive expected value mean I will win the trade?
No. Expected value is a long-run average, not a prediction of the next result. A positive-EV trade can and often does lose on any given attempt; the positive figure only means that if you repeat the trade many times, the average outcome is a profit.
Can a trade be profitable while losing most of the time?
Yes. Expected value is silent about win rate. If the rare wins are large enough, a trade that loses most of the time can be positive-EV. Conversely, a trade that wins most of the time can be negative-EV if its rare losses are huge, as with naked option selling.
Why does win rate not equal expected value?
Because expected value weights outcomes by both probability and size, while win rate counts only how often you win. A high win rate with tiny wins and rare huge losses can be negative-EV, so focusing on win rate for emotional comfort routinely leads traders into losing positions.
How do costs affect expected value?
Costs reduce it directly. Brokerage, exchange fees, GST, stamp duty, STT and slippage must be subtracted from each outcome, and at high turnover they can turn a positive gross EV negative. An honest expected-value calculation is always net of realistic costs, not gross.
Why do I need many trades for expected value to work?
Because it is a long-run average and any single outcome can sit far from it. The law of large numbers says the realised average converges to the true EV only as trades accumulate, so you must be capitalised and sized to survive the variance before the average arrives.
How do I estimate the probabilities for EV?
Usually from the historical frequency of similar setups reaching target before stop, adjusted for current conditions and treated as approximate. Because the EV is only as good as these inputs, use conservative, evidence-based figures rather than optimistic ones, and view the result as a range, not a precise point.
What is the difference between expected value and expectancy?
They are essentially the same idea. Expectancy is the term traders often use for the average profit or loss per trade of a strategy, which is exactly its expected value. Some define expectancy per rupee risked; both express the probability-weighted average result.
Can I make expected value positive just by hoping?
No, and trying to is a classic self-deception. Nudging the win rate upward on paper makes the EV look positive by construction, but reality uses the true probabilities. Honest EV work resists the temptation to assume the numbers that produce the answer you want.
Does expected value tell me how much to trade?
No. Expected value tells you whether a trade is worth taking, not how large to make it. A positive-EV trade can still carry a tail loss big enough to ruin the account, so you must pair it with position sizing that caps the worst plausible loss at a small fraction of capital.
What is the Kelly criterion's link to expected value?
The Kelly criterion uses the edge and odds, which come from the expected-value inputs, to compute the fraction of capital that maximises long-run growth. It formalises the jump from deciding a bet is positive-EV to deciding how much to stake, though most traders use a fraction of Kelly because the inputs are uncertain.
Why can a positive-EV trader still go broke?
Because expected value is realised only over many trades, and a run of losses or an oversized tail loss can exhaust the account before the average arrives. A positive edge you cannot survive to realise is worthless, which is why EV must always be paired with survival-based sizing.
How does expected value handle rare big losses?
By weighting them by their probability, provided you include them. A one-in-twenty catastrophic loss still enters the sum as its size times 5 percent, which can dominate the average. The danger is leaving the tail outcome off the calculation entirely, which flatters a fragile trade.
Is a higher expected value always better?
Not necessarily, because a higher EV can come with a larger tail risk or greater variance. A trader optimising for long-run survival may prefer a slightly lower EV with a much smaller chance of ruin, since staying in the game is the precondition for any edge to compound.
How is expected value used in options trading?
It weighs the premium collected or paid against the probability-weighted payoff across the range of underlying outcomes at expiry. Because option payoffs are non-linear, EV is especially useful for exposing when a high-probability short option carries a low-probability but large loss that makes its true expected value negative.
Can expected value be negative and still tempt me?
Yes, and that is its value as a check. Lotteries, most tips and many high-win-rate option strategies are negative-EV but emotionally attractive because of a jackpot dream or a comforting hit rate. Computing the EV strips away the feeling and shows the average is a loss.
How often should I recalculate expected value?
Whenever the inputs change materially: a shift in volatility regime, a change in costs, or evidence that your setup's win rate has drifted. Expected value is not a one-time number but a living estimate, and stale probabilities from a different market regime can quietly turn a positive edge negative.
Does expected value apply beyond individual trades?
Yes. You can compute the expected value of a whole strategy, a portfolio, or a business decision by weighting its outcomes by probability. In trading it also guides whether to take on a position at the book level, where the relevant EV is the marginal contribution net of correlation with existing risk.

Voice search & related questions

Natural-language questions people ask about Expected Value.

What is expected value?
It is the average result of a trade if you could repeat it many times. You multiply each outcome by its odds and add them up; positive means you make money over the long run.
What is the expected value formula?
Add up each outcome times its probability. For a simple trade it is your win chance times the win size, minus your loss chance times the loss size.
Can a trade make money if it loses most of the time?
Yes. If your wins are big enough, you can lose more often than you win and still come out ahead. The size of the wins and losses matters, not just how often you win.
Does positive expected value mean the next trade wins?
No. It is a long-run average, not a forecast. The next trade can easily lose; the point is that repeating the bet many times makes money on average.
Why does a high win rate fool people?
Because winning often feels safe, but if the rare loss is huge it can wipe out lots of small wins. Do the expected-value math and the truth shows up.
Why do I need lots of trades for it to work?
Because any single trade can land far from the average. Only over many trades does your real edge show through, so you have to survive the rough patches to get there.
Do costs change the expected value?
Yes. Subtract brokerage, STT and slippage from every outcome. At high trading frequency those small costs can turn a winning edge into a losing one.

Sources & references

Last reviewed 12 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

← Previous
Probability Thinking
Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Risk-management techniques reduce but never remove risk, and trading derivatives involves substantial risk of loss. See our Risk Disclosure and SEBI Disclaimer.