Skip to main content

Expected Value (EV)

Expected value answers one question: if I made this exact trade thousands of times, would I come out ahead or behind?

It is the single most important concept in trading. Not technical analysis. Not chart patterns. Not "conviction." Every profitable trading desk in the world, from market makers to hedge funds, operates by finding and repeating positive EV trades.

One-sentence definition

EV is what you expect to make (or lose) on average per trade, calculated by multiplying each possible outcome by its probability and adding them up.

Start Simple: A Coin Flip

Before options, think about a bet. Someone offers you a coin flip:

  • Heads: you win $200
  • Tails: you lose $100

Should you take it? Your gut might say "it's 50/50, risky." But do the math:

Fair coin flip: should you take this bet?
50%+$200Heads (win)50%$-100Tails (lose)
50% × +$200 = +$100 (Heads (win))
50% × $-100 = $-50.00 (Tails (lose))
EV = +$50.00 per trade

The EV is +$50 per flip. You should take this bet every time it is offered. You will lose individual flips, but the wins are large enough to more than cover the losses.

The #1 Beginner Mistake

Win rate and EV are not the same thing. A trade that wins 80% of the time can have negative EV. A trade that loses 60% of the time can have positive EV. What matters is how much you win vs. how much you lose, weighted by how often each happens.

The Formula

EV = P(win) × gain + P(loss) × loss

Hover each term for a definition. "loss" is negative, so the second term subtracts from the first.

That is the whole formula. Everything else on this page is applying it.

Why Win Rate Lies to You

This is the trap that catches most beginners. Look at these two traders:

Trader A: 80% win rate

Sells OTM puts

  • Wins 80 out of 100 trades
  • Average win: +$50
  • Average loss: -$400
  • Total: (80 × $50) + (20 × -$400) = -$4,000
  • Net result: LOST $4,000

Trader B: 35% win rate

Buys OTM calls

  • Wins 35 out of 100 trades
  • Average win: +$600
  • Average loss: -$100
  • Total: (35 × $600) + (65 × -$100) = +$14,500
  • Net result: MADE $14,500

Trader A looks like a genius 80% of the time. Trader B looks like a loser 65% of the time. But Trader B is the one making money. This is why EV, not win rate, is the number that matters.

Trader A: high win rate, negative EV
80%+$50Win (80%)20%$-400Loss (20%)
80% × +$50 = +$40.00 (Win (80%))
20% × $-400 = $-80.00 (Loss (20%))
EV = $-40.00 per trade
Trader B: low win rate, positive EV
35%+$600Win (35%)65%$-100Loss (65%)
35% × +$600 = +$210 (Win (35%))
65% × $-100 = $-65.00 (Loss (65%))
EV = +$145.00 per trade
💡

Trader A's strategy is the classic "picking up pennies in front of a steamroller." Small, frequent wins feel good. But when the loss comes, it is catastrophic relative to the wins. Option sellers who ignore EV fall into this trap constantly.

EV Calculator

Plug in your own numbers. Try different combinations to build intuition for how probability and payoff size interact.

win_contribution = 40% × +$500 = +$200.00
loss_contribution = 60% × -$200 = -$120.00
EV = +$200.00 + (-$120.00) = +80.00
Expected Value per Trade
+$80.00
Positive EV. Repeating this trade makes money on average.

Applying EV to Options

Now let's use a real options trade. You buy a BTC call option for $200 premium. You estimate:

  • 40% chance BTC rallies enough for the option to be worth 700atexpiry(netgainof+700 at expiry (net gain of **+500**)
  • 60% chance the option expires worthless (you lose your $200 premium)
BTC call option: +$200 premium, 40% chance of a $500 gain
40%+$500Expires ITM60%$-200Expires OTM
40% × +$500 = +$200 (Expires ITM)
60% × $-200 = $-120 (Expires OTM)
EV = +$80.00 per trade

The EV is +$80 per trade. You will lose money on 60% of these trades. That can feel terrible. But the 40% that win pay enough to more than compensate.

Over 100 trades: 40 wins at +500 = 20,000. 60 losses at -200 = -12,000. Net profit: 8,000 dollars.

More Than Two Outcomes

Most option trades have more than "win or lose." A call could expire deep ITM, slightly ITM, or worthless. The EV formula handles this naturally by adding more rows:

BTC call with three possible outcomes
15%+$1.5kBig rally (deep ITM)25%+$300Moderate rally (ITM)60%$-200Expires worthless
15% × +$1.5k = +$225 (Big rally (deep ITM))
25% × +$300 = +$75.00 (Moderate rally (ITM))
60% × $-200 = $-120 (Expires worthless)
EV = +$180.00 per trade

Same logic, more granularity. The deep ITM outcome is rare but large, and it contributes meaningfully to total EV.

When Is Buying vs. Selling Positive EV?

Neither buying nor selling options is inherently positive or negative EV. It depends on whether the market is pricing the option correctly.

Strategy
Positive EV when...
Negative EV when...
Buying calls/puts
Implied vol is too low. The market underestimates how much the asset will move.
IV is already elevated. You overpay for protection or upside.
Selling calls/puts
Implied vol is too high. You collect more premium than the actual risk justifies.
A crash or rally wipes out months of collected premium in one event.
Straddles (buy vol)
Realized vol will exceed implied vol over the life of the option.
Theta bleeds faster than the underlying moves.
⚠️
EV Is Only as Good as Your Probability Estimates

The math is easy. The hard part is estimating the probabilities. The market's option prices already embed a consensus probability distribution (via implied vol). To have positive EV, you need to believe that consensus is wrong. If your estimate is no better than the market's, your EV after fees is slightly negative.

See It Over Time: The Convergence Simulation

This is where EV becomes intuitive. The chart below simulates 200 trades using the probabilities and payoffs you set. The blue line is your actual running P&L (random). The dashed green line is where EV says you should end up.

Click Re-shuffle multiple times. Notice how the blue line wanders randomly but keeps ending up near the dashed EV projection. That is the law of large numbers at work.

Cumulative P&L over 200 trades
40%
$500
$200
06k12k18k24kEV line050100150200Trade #
EV / Trade
+80
Wins / Losses
88 / 112
Win Rate
44%
Final P&L
+21.6k
The blue line is actual P&L (random). The dashed line is EV projection. Click Re-shuffle to run a new simulation.

Now try these experiments:

Try this
What you will see
Set Win Prob to 30%, Win to $700, Loss to $100
Low win rate, but the green EV line slopes up. Positive EV despite losing most trades.
Set Win Prob to 90%, Win to $20, Loss to $500
High win rate, but the EV line slopes down. Negative EV despite winning almost every trade.
Click Re-shuffle 10 times with the defaults
Some runs dip negative for 50+ trades before recovering. Variance is real, even with positive EV.
Set Win Prob to 50%, Win to $100, Loss to $100
EV is exactly zero. The blue line random-walks. No edge means no drift.
💡
Variance Is Not the Same as Negative EV

A positive EV strategy can lose money for 20, 50, or even 100 trades in a row by pure chance. That does not mean the strategy is broken. It means the sample size is too small. This is why professional traders care about volume: they need enough trades for the math to work out. If your strategy only produces 5 trades per month, it may take years before you can distinguish skill from luck.

Why Your P&L Chart Lies to You

There is a counterintuitive result from probability theory called the ArcSine law (first described by Paul Levy; Taleb applies it to trading P&L in Dynamic Hedging, Chapter 3). In a fair game (EV = 0), you would expect to spend roughly half the year in profit and half in loss. That is wrong. The most likely outcome is spending almost the entire year on one side: either 11 months in profit and 1 in loss, or 11 months in loss and 1 in profit. Spending exactly 6 months on each side is the least likely outcome.

Probability of time spent in profit (fair game, 1 year)

1m
2m
3m
4m
5m
6m
7m
8m
9m
10m
11m
Most likelyLeast likelyMost likely

This means: a trader who has been profitable for 10 straight months is not necessarily skilled. A trader who has been losing for 10 straight months is not necessarily unskilled. Cumulative P&L charts over short periods are deeply misleading. The only reliable signal is a large number of independent trades, which is why the convergence chart above uses 200.

⚠️
Track Records Are Mostly Noise

If you put infinite monkeys in front of trading screens, one of them will produce a 10-year track record that looks like Warren Buffett. That does not mean the monkey has edge. The ratio of luck to skill in a track record decreases with transaction frequency: for a market maker doing thousands of trades per day, a year of profitability is strong evidence of edge. For a fund manager making 10 trades per year, even a decade of profits could be chance. Always ask: how many independent bets does this track record represent?

This framing comes from Taleb's application of the Borel-Cantelli lemma to trader track records (Dynamic Hedging, Chapter 3). The original mathematical result proves that with enough trials, even extremely unlikely events become certain.

What Is "Edge"?

Edge is whatever makes your EV positive. Without edge, your EV on any trade is zero or slightly negative (after fees). Edge means you know something the market price does not fully reflect.

Real Edge

Where positive EV comes from

  • Better volatility estimates (you believe realized vol will differ from IV)
  • Faster information (you react to on-chain data or news before prices adjust)
  • Structural advantages (cheaper funding, more efficient margining, ability to warehouse risk)
  • Market making (you capture the bid-ask spread on each trade)

Not Edge

Things people mistake for edge

  • "I feel strongly that BTC will go up" (conviction is not analysis)
  • "This back-tested strategy returned 300%" (overfitting to historical data)
  • "Options usually expire worthless" (the premium already prices that in)
  • "I follow a whale wallet" (by the time you see it, so does everyone else)

Common EV Traps

Trap
What it looks like
What is actually happening
Picking up pennies in front of a steamroller
You win $50 every week selling OTM puts. Easy money.
One crash gives back 6 months of gains. EV was negative the whole time, but the feedback loop was delayed.
Ignoring fees and slippage
Your spreadsheet shows +$5 EV per trade.
After $6 in spread + fees, your real EV is -$1. You are paying the exchange to trade.
Confusing implied probability with actual probability
The option market says there is a 10% chance of a crash.
That 10% includes a risk premium. The actual probability might be 6%, but people pay extra for protection, inflating the implied number.
Sizing by conviction
You size up on your "best ideas."
A high-conviction trade with thin edge, sized too large, turns a +EV strategy into a path to ruin. One bad run and you are done.

Thinking in EV

The most important shift a trader can make is moving from judging trades by their outcome to judging trades by their process.

Outcome Thinking

How beginners evaluate

  • "That trade lost money, so it was a bad trade"
  • "I made money, so my strategy works"
  • "I should stop selling puts, I got burned last month"
  • Judges the decision by the result

EV Thinking

How professionals evaluate

  • "That trade had positive EV. The loss was expected variance."
  • "I made money, but was the trade actually positive EV or did I get lucky?"
  • "The put sale had positive EV. I'll keep selling, but size smaller."
  • Judges the decision by the process

EV Across Different Domains

Test your understanding before moving on.

Q: You sell a BTC put option and collect 150 in premium. The put has a 15% chance of expiring ITM with an expected loss of 1,200. What is the EV?
Q: Two traders both made +10,000 last month. Trader A took 200 small trades. Trader B took 3 big bets. Whose performance tells you more?
Q: A 7-day BTC ATM option has 65% implied vol, but you believe realized vol will be 55%. Is buying or selling positive EV?
Q: You have been running a positive EV strategy for 3 weeks and you are down 8%. Should you stop?

💡 Tip: Try answering each question yourself before revealing the answer.


Related: