What Is EV in Blackjack?
Every rule in the strategy chart — hit 16 vs. 10, split 8s, never take insurance — is the answer to one question: which action has the highest expected value? Understanding EV turns the chart from a list to memorize into a system you can reason about.
Expected value in one sentence
EV is the average outcome of a decision if you could repeat it millions of times. An EV of −0.54 on a 1-unit bet means that choice loses 0.54 units on average per attempt. The single hand in front of you will win or lose entirely — EV describes the long run, which is the only thing a decision can actually control.
A worked example: hard 16 vs. dealer 10
The most hated spot in blackjack is a clean EV comparison:
| Action | Approx. EV (per 1 unit) |
|---|---|
| Stand | −0.54 |
| Hit | −0.51 |
| Surrender (if offered) | −0.50 |
Every option is bad — you're a big underdog either way. But hitting loses less than standing, and surrender (giving up exactly half) beats both. "Correct play" never meant "winning play"; it means the option that bleeds least. That reframing is the core mental shift EV training produces.
How EV is actually computed
For a given hand, dealer upcard, rule set and shoe composition, a solver walks the full tree of what can happen:
- Stand EV: enumerate every way the dealer's hand can finish (bust, 17, 18, 19, 20, 21) with its probability, and sum the win/lose/push outcomes.
- Hit EV: for each card you might draw, compute the EV of the resulting hand played optimally from there — a recursive calculation.
- Double EV: like a one-card hit, with doubled stakes.
- Split EV: play each new hand optimally, accounting for re-split and double-after-split rules.
The result is exact, not simulated — and it shifts as cards leave the shoe. A 16 vs. 10 in a ten-rich shoe has a different hit EV than the same hand off the top. That's precisely the information card counting approximates with one running number, and why counting and EV are two views of the same underlying math.
Why train with EV visible?
Traditional trainers tell you "correct" or "wrong". EV feedback tells you how much a decision cost, which changes what you learn:
- Mistakes get sized. Standing on soft 18 vs. 9 (−0.10 vs. hitting) and mis-playing a marginal 12 vs. 4 (~−0.01) are both "wrong" — but one is ten times the leak. You fix the expensive habits first.
- Close calls stop feeling arbitrary. Seeing that two actions differ by 0.003 units explains why the cell was hard to memorize — and why deviations flip exactly there when the count moves.
- Variance stops fooling you. You hit the 16, busted, and the table groaned — but the EV panel shows it was right. Training on decision quality instead of hand outcomes is the entire discipline.
House edge is just summed EV. Add up the best-action EV of every possible hand, weighted by how often each occurs, and you get the game's overall expectation — about −0.5% for perfect basic strategy on a standard 6-deck game. Every strategy error you make subtracts its EV gap from that number.
EV beyond the playing decisions
The same lens applies to money management. Bet sizing by true count, risk of ruin, and the Kelly criterion are all EV-and-variance calculations over a bankroll instead of a single hand. A disciplined practice loop treats them identically: simulated chips, measured expectation, no outcome-chasing.
See the math, live
Blackjack Edge computes exact EV for hit, stand, double, split and surrender on every hand, from the actual shoe state — as you play.
FAQ
Is EV the same as odds or probability?
No. Probability says how often something happens; EV combines probabilities with payoffs into one number — the average result. A 30% chance to win 2 units and a 70% chance to lose 1 unit has EV −0.10, even though "you usually lose" and "the win pays double" are both true.
If a play is +EV, why did I lose?
Because a single hand is one draw from a distribution. EV is the average of that distribution, not a promise about any draw. Over thousands of hands, results converge toward EV; over one hand, anything happens. Judging decisions by outcomes is the most expensive habit in the game.
Does EV change during a shoe?
Yes — that's the whole basis of card counting. As the shoe's composition shifts, so do the probabilities behind each action's EV. A few borderline plays flip when the count moves; those flips are the index deviations counters memorize.