# What Share of Blackjack Rounds Do You Win, Push and Lose?

A 62 million round simulation measures how often a basic strategy player wins, pushes and loses a blackjack round, and shows why more losses than wins still costs only 0.44 units per 100 rounds.

Source: https://21trainer.app/articles/how-often-you-win-a-hand/
Updated: October 13, 2026
Publisher: 21 Trainer, makers of Blackjack Strategy: 21 Trainer, an educational blackjack trainer for iPhone and Android (App Store: https://apps.apple.com/app/blackjack-strategy-21-trainer/id6759527695, Google Play: https://play.google.com/store/apps/details?id=app.blackjacktrainer)

> **Educational content, not gambling advice.** 21 Trainer and this article teach blackjack strategy and card counting as skills, using virtual chips only. No strategy or counting system guarantees winnings, and nothing here encourages real-money play. If gambling is a problem for you or someone you know, call 1-800-522-4700 (National Council on Problem Gambling).

Played correctly, blackjack is a game you lose more often than you win: under ordinary six-deck rules a basic strategy player loses 47.89 percent of rounds and wins 43.34, with 8.77 pushed, per our simulation.

That sounds bleak, yet the same simulation puts the average result at just -0.44 percent of the starting bet. Losing 4.55 points more often than winning (47.89 minus 43.34) somehow costs less than half a unit per hundred rounds.

The resolution: in blackjack, a win is bigger than a loss on average.

## The simulation

The numbers come from the same engine behind our pieces on [how the player's edge moves with the count](https://21trainer.app/articles/player-edge-by-true-count/) and [whether counting changes how often you win a hand](https://21trainer.app/articles/do-you-win-more-hands-counting/): about 62 million rounds, one player, a flat one-unit bet, basic strategy from the app's 4 to 8 deck chart. The rules: six decks, dealer stands on soft 17 and peeks, blackjack pays 3:2, double on any two cards, double after split, split to four hands, one card to split aces with no resplitting, no surrender, cut card at 75 percent penetration.

A round is everything that happens to one starting bet, including any doubles and splits: won when its net result is positive, pushed when zero, lost when negative. A split into three hands that nets one unit won counts as one win, not two wins and a loss.

| Result | Share of rounds |
| --- | --- |
| Won | 43.34% |
| Pushed | 8.77% |
| Lost | 47.89% |
| Average result, percent of the starting bet | -0.44 |

A 414 million round run of the same engine and rules gave 43.35, 8.77 and 47.88, agreeing to a hundredth of a point, with an average result of -0.43 (-0.428 against -0.436 here). The second decimal is therefore the limit of what the simulation supports.

## The shape of a round's result

| Net result of the round | Share of rounds | Units per 100 rounds |
| --- | --- | --- |
| Won 3 or more units | 0.33% | +1.10 |
| Won 2 units | 5.87% | +11.74 |
| Won 1.5 units (a natural) | 4.53% | +6.79 |
| Won 1 unit | 32.62% | +32.62 |
| Pushed | 8.77% | 0 |
| Lost 1 unit | 43.42% | -43.42 |
| Lost 2 units | 4.20% | -8.41 |
| Lost 3 or more units | 0.26% | -0.85 |

One-unit results dominate, and they are lopsided: 43.42 rounds lost against 32.62 won. (The units column uses unrounded shares, 4.529 for naturals and 4.2029 for two-unit losses, hence 6.79 and 8.41 where the rounded shares give 6.80 and 8.40.)

Wins bring in 32.62 plus 6.79 plus 11.74 plus 1.10, which is 52.25 units across 43.34 winning rounds, an average win of 1.21 units. Losses take out 43.42 plus 8.41 plus 0.85, which is 52.68 units across 47.89 losing rounds, an average loss of 1.10. The net is 52.25 minus 52.68, or -0.43 from these rounded columns, -0.436 unrounded.

## Where the gap closes

Suppose every win paid one unit and every loss cost one: the player would be down exactly the frequency gap, 4.55 units per 100 rounds (4.54 unrounded). Two things give about 4.1 of them back.

First, the 3:2 payout on naturals: the extra half unit lands on 4.53 naturals per 100 rounds, worth 2.26 units (4.529 times 0.5).

Second, doubles and splits. Rounds decided for two units or more were won 6.20 times per 100 (5.87 plus 0.33) and lost 4.46 times (4.20 plus 0.26). They won 12.84 units (11.74 plus 1.10) and lost 9.26 (8.41 plus 0.85), so the units beyond the first come to 12.84 minus 6.20, or 6.64 won, and 9.26 minus 4.46, or 4.80 lost: a net of 1.84 in the player's favor. Together, 2.26 plus 1.84 equals 4.10, and -4.54 plus 4.10 equals -0.44.

A little over half of the recovery is the natural's bonus half unit; the rest comes from doubles and splits. Paying naturals at 6:5 instead of 3:2 removes 0.3 units from each of the 4.53 naturals, 1.36 units per 100 rounds, in line with the roughly 1.4 points our article on [the 3:2 versus 6:5 payout](https://21trainer.app/articles/blackjack-3-2-vs-6-5/) gives for that rule change.

## Doubles, splits and the bust race

A round included a double 10.27 percent of the time and a split 2.48 percent, both in 0.73 percent, so more than one unit was on the table in 10.27 plus 2.48 minus 0.73, or 12.02 percent of rounds; the average total wagered was 1.13 units. A doubled hand won 55.17 percent of the time, pushed 6.94 and lost 37.89, well above the 43.53 percent of all hands won.

Every player hand busted in 15.80 percent of rounds, and those rounds are lost before the dealer draws. The dealer played out the hand in 74.92 percent of rounds, not when every player hand had busted or a natural had settled things, and busted in 30.91 percent of those: 23.16 percent of all rounds.

Counted over all rounds, the dealer busts more often than the player: 23.16 percent of rounds against the 15.80 percent in which every player hand busts. The player still loses more rounds, and the order of play is one part of why: the player acts first, so a player bust loses even where the dealer would have busted too. Set 23.16 against the 43.34 percent of rounds won, and a dealer bust lies behind roughly half of the player's wins (only roughly: a split round can include a dealer bust and still not finish ahead).

## Why sources quote different numbers

Search for a blackjack win rate and you will meet figures that disagree by several points. Four choices are enough to move the answer that far.

| How the figures are counted | Won | Pushed | Lost |
| --- | --- | --- | --- |
| This article: rounds, no surrender | 43.34% | 8.77% | 47.89% |
| Same simulation, split hands counted separately | 43.53% | 8.57% | 47.90% |
| This article's run, pushes left out | 47.51% | left out | 52.49% |
| Rerun with late surrender allowed | 42.40% | 8.49% | 49.11% |
| Wizard of Odds simulation, nearly the same rules | 42.43% | 8.48% | 49.09% |

First, rounds or hands. Counting every split hand separately, 1.028 hands per round here, barely moves the shares, because only 2.48 percent of rounds split.

Second, pushes in or out. Of the rounds that do not push, the player won 47.51 percent and lost 52.49: a source quoting 47 or 48 percent and one quoting 43 can be describing the same game. Our article on [how often a blackjack hand pushes](https://21trainer.app/articles/blackjack-push-frequency/) covers where that roughly 8 to 9 percent comes from.

Third, the rules; surrender is the one tested here. We reran the simulation with late surrender allowed and nothing else changed, about 63 million rounds: won 42.40, pushed 8.49, lost 49.11, with 4.47 percent of rounds surrendered and counted as losses. That is 0.94 fewer points won, 0.28 fewer pushed and 1.22 more lost. The Wizard of Odds publishes a simulated table for almost the same rules (its list adds resplitting aces): positive rows sum to 42.43 percent, a push is 8.48 and the rest, 49.09, is lost, 4.47 of it surrendered. Our rerun lands within a few hundredths of a point of it.

Surrender is good for the player; our piece on [surrender as a strategy decision](https://21trainer.app/articles/blackjack-surrender-strategy/) gives it as worth roughly 0.08 percent under commonly published figures. Yet it lowers the win share by about a point and raises the loss share by more than a point, because the surrendered hands are ones the player would sometimes have won or pushed and now always loses, at half price. If the whole shift between the two runs comes from those hands, then of the 4.47 rounds per 100 surrendered, playing them out would have won about 0.94, pushed about 0.28 and lost about 3.25 (4.47 minus 0.94 minus 0.28), a net loss of about 2.31 units. Surrendering them instead costs 4.47 times 0.5, about 2.24 units. The difference, roughly 0.07 units per 100 rounds, is in line with the published value of the rule. Only roughly, because these are differences between two separate runs, good to a few hundredths. Share of rounds won, in short, is a poor guide to how good a set of rules is.

Fourth, how the hands are played. Every figure here assumes basic strategy on every hand; mistakes shift the shares by amounts this simulation did not measure.

## What the shares will not tell you

These are long-run numbers. They describe what a player gets per 100 rounds over tens of millions of them, not one evening, where a few hundred rounds swing widely around the averages; our piece on [streaks and randomness](https://21trainer.app/articles/streaks-and-randomness/) covers why short samples mislead. Nothing here is betting advice, and no session target follows from a percentage.

The shares also barely move with the count: in the counting article the win share moves by less than 1.7 points from the lowest count row to the highest, while the average result does most of the moving. Counting with your head is legal in the United States, casinos are private businesses that may refuse service, and no technique removes risk from the game.

## Frequently asked questions

### What percentage of blackjack hands do you actually win?

In our simulation, a basic strategy player under standard six-deck rules won 43.34 percent of rounds, pushed 8.77 and lost 47.89. A round is one starting bet with any doubles and splits, and it counts as won only when its net result is positive.

### How can the house edge be small if you lose more hands than you win?

Because an average win is bigger than an average loss. Naturals pay 3:2, and doubles and splits put extra money out, so the average winning round is worth about 1.21 units against an average loss of about 1.10. That nearly cancels the surplus of losses.

### Why do some sites say you win 48 percent of blackjack hands?

Often because pushes are left out. Counting only the rounds that are won or lost, our simulation shows 47.51 percent won and 52.49 lost. A source quoting 48 percent and one quoting 43 can be describing the same game with different denominators.

### Does surrender change how often you win a blackjack hand?

Yes, in a surprising direction. Our rerun with late surrender allowed dropped the win share from 43.34 to 42.40 and raised losses from 47.89 to 49.11 percent. Yet the rule helps the player, because the surrendered hands lose so often when played out that giving up half the bet costs less.

### Can you practice blackjack decisions without playing for real money?

Yes. 21 Trainer deals practice hands with virtual chips, gives instant feedback on every decision and has a house edge calculator for the rules you select. Nothing real is wagered, and these shares describe the long run, not any session.

## Keep reading

- [How Often Does a Blackjack Hand Push?](https://21trainer.app/articles/blackjack-push-frequency/)
- [Does Counting Change How Often You Win a Hand?](https://21trainer.app/articles/do-you-win-more-hands-counting/)
- [Expected Value in Blackjack, Without the Math Degree](https://21trainer.app/articles/expected-value-blackjack/)