Expected value (EV)
Expected value is what a bet is worth on average across every way it could end, the one number that says whether a wager makes money over time. Positive EV, repeated with discipline, is the entire business.
Expected value is what a bet is worth on average once you weigh the payoff by the odds. A positive-EV bet still loses often, the money shows up over many repetitions, which is why EV is your fund’s native language.
Any single bet can win or lose. That part is noise. Expected value is the signal underneath: the average result if the same bet could be replayed a thousand times. It is one multiplication and one subtraction, and it quietly runs every decision a serious fund makes.
01The arithmetic, on one bet
Take a Yes share at 62¢ on a market where your research supports a 74% chance. Two futures exist: 74% of the time the share pays $1; 26% of the time it pays $0. Weigh each outcome by its probability:
| Outcome | Chance | Result per share | Weighted |
|---|---|---|---|
| Yes lands | 74% | +38¢ | +28.1¢ |
| Yes misses | 26% | −62¢ | −16.1¢ |
| Expected value | +12.0¢ per share |
On average, this bet earns 12¢ per 62¢ staked. Notice what made it positive: not the payout, not the probability alone, but the gap between your 74% and the market’s 62%, the edge, converted into cents.
02Positive EV still loses, often
That +12¢ bet loses outright one time in four. The favorite fumbles, the poll was noise, the coin lands wrong, and the bet was still correct. EV is a property of the decision, not the outcome. Judge decisions by the number that was true when they were made; judge outcomes over a season, not an afternoon.
03Repetition is where the money is
One +12¢ bet is a coin with a friendly lean. Hundreds of them, each sized to survive its losses, is a machine: the misses and hits average out, and what remains is the sum of the edges. This is why a fund cares more about the pipeline of bets than any single one, and why it can report a losing week without a losing process.
Negative-EV bets also win, just often enough to keep their buyers coming back. A lottery ticket pays someone every week. The math still collects from all of them.
04EV as your fund’s native language
Every candidate trade your CEO evaluates is forced through this exact arithmetic, its estimate versus the price, weighted, netted, and compared against fees and alternatives. Bets below zero don’t get sized small; they don’t get placed. The one-paragraph version is in the FAQ: what is expected value?
How your CEO runs the numbers
Your CEO is looking at a market priced at 40¢ for “yes,” while its own research puts the true chance at 50%. Each $1 share pays $1 if right, $0 if wrong, so the expected value per share is (0.50 × $1) − $0.40 = +$0.10 A 25% edge on cost. Over a single bet that edge is invisible; across a hundred bets at the same edge, it’s what the ledger shows. Move the price to 55¢ on the same read and EV flips to −$0.05 So the CEO passes.
Common misreads
- Judging a bet by its outcome. A losing positive-EV bet was still the right bet; a winning negative-EV bet was still luck.
- Expecting each bet to pay. EV shows up over many repetitions, not on any single trade.
- Forgetting fees. A thin edge can vanish after costs. Your CEO computes EV net of the 0.5%/1% fee.
What to remember
- EV weighs payoff by probability The average outcome of a bet.
- A positive-EV bet still loses individually, and often.
- The edge compounds only over many repetitions, not one bet.
- Your CEO thinks in EV by default, never betting a price it believes is fair.