A blackjack house edge calculator turns the house edge into money. Enter your bet per hand, your hands per hour, the hours you play and the edge, and it shows the average amount the table keeps. With the starting numbers (10 units a hand, 60 hands an hour, 2 hours, 0.5% edge) you stake 1,200 units and the expected cost is 6.00 units.
Many people searching for a blackjack calculator want one that says whether to hit or stand. This tool does something different: it prices a session. The edge depends on the rules of the table, so this page also lists the edge for common rule sets, each worked out by us, so you can pick the right number. If gambling has stopped feeling like a choice, the Safer play toolkit is a good place to start.
What does a blackjack house edge calculator tell you?
A blackjack house edge calculator tells you the average cost of playing: bet per hand times hands per hour times hours times the house edge. At 10 units a hand, 60 hands an hour and a 0.5% edge, that is 3.00 units an hour. It is a long-run average, not a forecast of tonight’s result.
The house edge is the share of each bet the casino keeps on average over many hands. A 0.5% edge means 0.5 units kept per 100 units bet, even though you win some hands and lose others along the way. The edge applies to the total you bet, so speed and bet size matter as much as the rules. Two players with the same edge can have very different costs if one bets five times as much.
This tool is a cost calculator, not a strategy calculator. It does not look at your cards. A strategy tool picks the best play for one hand; this one adds up what many hands cost, and it is best used before you sit down.
How do you use the blackjack house edge calculator?
Enter four numbers and read the expected cost. Bet per hand, hands per hour and hours played set how much you bet in total. The house edge sets how much of that total the table keeps on average. If you are unsure of the edge, take it from the rule tables on this page.
- Enter your bet per hand. Use the amount you put in the betting circle before the cards are dealt, in units. Do not add doubles or splits; the edge figures here already allow for them.
- Enter hands per hour. Use your own pace if you know it. Otherwise use the hands-per-hour table below as a planning figure.
- Enter hours played. Use the time you plan to sit, not the time you hope to stay.
- Enter the house edge. Start with 0.5 for a 3 to 2 table with common rules and basic strategy, or about 1.8 where blackjack pays 6 to 5. Use the rule table below for a closer match.
- Read the expected cost. Change one number at a time to see which one moves the cost most. The line under the result shows the sum, so you can check it by hand.
Example: You plan 3 hours at 25 units a hand and 70 hands an hour. At a 0.5% edge, total staked is 25 x 70 x 3 = 5,250 units and the expected cost is 5,250 x 0.5 / 100 = 26.25 units. The same three hours at a 6 to 5 table, using an edge of 1.8%, is 5,250 x 1.8 / 100 = 94.50 units.
What is the blackjack house edge under different rules?
With correct basic strategy the blackjack house edge runs from about 0.3% to about 0.8% when a natural pays 3 to 2, depending on the rules, and about 1.8% to 2.1% when it pays 6 to 5. The figures below assume six decks, a dealer who checks for blackjack, and no card counting.
Researchers writing on blackjack mathematics describe the edge under basic strategy as about 1% in typical games and about 0.5% with the best rules (an arXiv paper on the standard deviation effect in blackjack). The table gives finer detail. Terms used: S17 means the dealer stands on a soft 17 (a hand with an ace counted as 11), H17 means the dealer hits it, DAS means you may double down after splitting a pair, surrender means you may give up a hand for half your bet, and resplit aces means you may split aces again.
| Rules (six decks, basic strategy) | House edge | Average cost per hour at 10 units, 60 hands |
|---|---|---|
| 3 to 2, S17, DAS, late surrender | 0.33% | 1.98 units |
| 3 to 2, S17, DAS, resplit aces | 0.34% | 2.04 units |
| 3 to 2, S17, DAS | 0.41% | 2.46 units |
| 3 to 2, H17, DAS, late surrender | 0.53% | 3.18 units |
| 3 to 2, S17, no DAS | 0.55% | 3.30 units |
| 3 to 2, H17, DAS | 0.62% | 3.72 units |
| 3 to 2, H17, no DAS | 0.77% | 4.62 units |
| 6 to 5, S17, DAS | 1.77% | 10.62 units |
| 6 to 5, H17, DAS | 1.98% | 11.88 units |
| 6 to 5, H17, no DAS | 2.12% | 12.72 units |
| Even money (1 to 1) on naturals, S17, DAS | 2.67% | 16.02 units |
Deck count matters less than the payout, but it is not zero. With a 3 to 2 payout, S17 and DAS, we measured the edge at 0.20% with two decks, 0.35% with four, 0.41% with six and 0.43% with eight. Fewer decks help the player slightly, but the other rules usually matter more.
How were these edges worked out?
We wrote a program that deals 800 million rounds per row from a freshly shuffled shoe, with the dealer checking for blackjack. The player follows the strategy that maximizes expected value. Our results for the baseline rules matched exact figures published by independent calculators to within about 0.01 percentage points, and the 6 to 5 effect was also computed exactly (next section). Casinos deal with a cut card, so live results differ by a few hundredths of a percentage point. Single-deck games are left out because our method is less precise there. Our editorial standards explain how we check numbers.
Why does 6 to 5 blackjack cost three to four times more?
A 6 to 5 payout on a natural, instead of 3 to 2, adds about 1.36 percentage points to the house edge in a six-deck game. With S17 and DAS the edge goes from about 0.4% to about 1.8%, roughly four times higher. With H17 and DAS it goes from 0.6% to 2.0%, roughly three times.
The extra cost is plain arithmetic. In six decks, the chance your first two cards are an ace and a ten-value card is 2 x 24 x 96 / (312 x 311) = 4.75%. The chance the dealer then also has a natural is 4.56%, so about 4.53% of hands are a player natural that the dealer does not match. Each such hand pays 1.5 units per unit bet at 3 to 2 and 1.2 at 6 to 5, a loss of 0.3 units per unit bet. So 0.3 x 4.53% = 1.36 percentage points, whatever the other rules are.
In money, that is about 8 extra units an hour at 10 units a hand and 60 hands an hour (600 units staked x 1.36%). Regulators treat the payout as a basic term of the game. New Jersey’s rules set the blackjack payout at 3 to 2, and in 2024 its Division of Gaming Enforcement approved only a time-limited test of 6 to 5 blackjack, with conditions including a placard reading Blackjack Pay 6-5 at each such table. Wherever you play, the payout printed on the layout is the fastest way to know which row of the table above applies.
How many hands per hour do you really play?
Pace depends on how many players share the dealer. Figures attributed to Jim Kilby’s book Casino Operations Management put it at about 209 hands an hour for one player, 105 for three, 84 for four and 52 for seven, as tabulated by the Wizard of Odds. The Wizard’s own house-edge comparison assumes 70 an hour. No regulator publishes one official rate, so treat every figure here as a planning estimate and replace it with your own count when you can.
Manitoba’s GameSense program, a public-health body, uses 50 hands an hour for three of its four blackjack examples (and 40 for a double-deck example). The same figure of 50 appears as an assumption in the arXiv paper cited above. The 400 an hour for fast software-dealt games and the 40 for a slow dealer are our assumptions, not measured data. To find your own pace, count the hands you play in 15 minutes and multiply by four.
| Hands per hour (your own) | Typical setting | Bet per hour at 10 units | Cost at 0.5% edge | Cost at 1.8% edge (6 to 5) |
|---|---|---|---|---|
| 40 | Full table, slow dealer | 400 units | 2.00 units | 7.20 units |
| 50 | Standard full table (GameSense planning figure) | 500 units | 2.50 units | 9.00 units |
| 70 | Five players (Kilby: 70; also the Wizard’s average) | 700 units | 3.50 units | 12.60 units |
| 100 | Three players (Kilby: 105) | 1,000 units | 5.00 units | 18.00 units |
| 200 | One player against the dealer (Kilby: 209) | 2,000 units | 10.00 units | 36.00 units |
| 400 | Fast software-dealt game (our assumption, no source) | 4,000 units | 20.00 units | 72.00 units |
Speed is a cost multiplier that is easy to miss. Going from a full table to a quiet one multiplies your hourly cost by two to four without changing the rules or your bet, because you meet the edge on more hands. The cost per hand is the same; the cost per hour is not.
What does a blackjack session cost per hour at different bets?
At a 0.5% edge and 60 hands an hour, the average cost is 0.3 times your bet each hour: 3 units at a 10-unit bet, 7.5 at 25 and 30 at 100. Double the bet and the cost doubles. Double the speed and it doubles again. The grid below shows common bets and speeds.
| Bet per hand | 40 hands an hour | 60 hands an hour | 100 hands an hour | 200 hands an hour |
|---|---|---|---|---|
| 5 units | 1.00 | 1.50 | 2.50 | 5.00 |
| 10 units | 2.00 | 3.00 | 5.00 | 10.00 |
| 25 units | 5.00 | 7.50 | 12.50 | 25.00 |
| 50 units | 10.00 | 15.00 | 25.00 | 50.00 |
| 100 units | 20.00 | 30.00 | 50.00 | 100.00 |
For scale, a single-zero roulette wheel keeps 2.70% on every bet, as the roulette odds and payouts page shows, so blackjack with 3 to 2 and good rules costs about one seventh as much per unit bet. Tips and comps are not in these numbers. For other games, the planned gambling session cost calculator uses the same formula. Setting a time limit and a money limit before you play turns the figure into a plan; the Safer play toolkit explains how.
What if you play by gut or add side bets?
Playing by instinct costs far more than basic strategy. In our simulation, copying the dealer’s rule of hitting until 17 gives up about 5.7%, and never hitting a hand that could bust gives up about 6.0%, against about 0.4% for basic strategy. Side bets add edges of about 4.6% to 7.4% on the money put into them.
Basic strategy is a table of best plays for every hand and dealer card. The first accurate version was published in 1956 by four mathematicians known as the Four Horsemen of Aberdeen, according to the American Statistical Association’s history page. A Cornell mathematics handout puts the player’s disadvantage at about 6% for both mimic-the-dealer and never-bust play (a Cornell mathematics blackjack handout), which agrees with our results. GameSense uses 2.00% for an average player in its double-deck example, against 0.51% for its base game.
| Choice | House edge | Average cost per hour (10 units a hand, 60 hands) |
|---|---|---|
| Basic strategy, 3 to 2, S17, DAS | 0.41% | 2.46 units |
| Average-play figure used by GameSense (double deck) | 2.00% | 12.00 units |
| Copy the dealer (hit to 17, no doubling or splitting) | 5.7% | 34.20 units |
| Never bust (stand on any hand that could bust) | 6.0% | 36.00 units |
| Insurance, taken every time the dealer shows an ace (six decks) | 7.40% of the insurance bet | 1.71 units (about 4.6 offers an hour) |
| Perfect Pairs side bet, 10 units every hand | 6.11% | 36.66 units, on top of the main bet |
| 21+3 side bet (100/40/30/10/5 paytable), 10 units every hand | 3.70% | 22.20 units, on top of the main bet |
Paytables used: Perfect Pairs pays 25 to 1 for a pair of the same suit, 12 to 1 for the same color and 6 to 1 for mixed colors, which gives 6.11% at six decks according to the Wizard of Odds Perfect Pairs page; 21+3 pays 100 to 1 for suited three of a kind, 40 to 1 for a straight flush, 30 to 1 for three of a kind, 10 to 1 for a straight and 5 to 1 for a flush, which gives 3.70% at six decks according to the Wizard of Odds 21+3 page. Insurance costs half your bet and pays 2 to 1, and the hidden card is worth ten in 96 of the 311 unseen cards, which gives the 7.40% edge at six decks (our arithmetic; other deck counts give other figures). A different paytable or deck count changes these figures, so check the printed one.
How far can a real session stray from the estimate?
Far further than the average suggests. One hand has a standard deviation of about 1.15 times your bet, so after 60 hands a typical swing is about 9 bets either way, while the expected cost is only 0.24 of a bet. Most short sessions end closer to the swing than to the average.
| Hands played | Average cost (bets) | Typical swing (1 standard deviation, bets) | Ended ahead | Down 10 bets or more | Up 10 bets or more |
|---|---|---|---|---|---|
| 60 | 0.24 | 8.9 | 48% | 14% | 13% |
| 120 | 0.49 | 12.6 | 48% | 23% | 21% |
| 500 | 2.03 | 25.8 | 46% | 38% | 32% |
At 10 units a hand, 120 hands is the default two-hour session. The expected cost is about 5 units, but the typical swing is about 126 units either way. A session that ends ahead is variance, not a sign the edge is gone, and a bad one does not make the next hand more likely to win: each hand is dealt from a shuffled shoe, independent of the last. Gambler’s fallacy, the belief that past results change the next hand, is one of the most common reasons people raise their bets after a loss.
The average cost only becomes visible over thousands of hands. For one night, treat it as the price of playing and the swing as the risk. If swings of that size would hurt, a smaller bet or a shorter session is the change that helps.
What mistakes make a blackjack cost estimate wrong?
The most common mistake is using the wrong edge. A 0.5% figure belongs to a 3 to 2 table played with basic strategy; at a 6 to 5 table or with side bets the cost is several times higher. These errors add up.
- Using 0.5% at a 6 to 5 table. Check what a natural pays before you enter the edge. The cost per hour is about 3.6 times higher at 1.8%.
- Counting the table’s hands, not yours. A dealer deals far more hands an hour than any one player sees. Use the number of hands you play.
- Adding doubles and splits to your bet. The edges here are per initial bet and already include the extra money from doubling and splitting (about 1.13 units in play per unit bet in our simulation). Enter the bet you place before the deal.
- Assuming perfect play. Every error raises the edge. The figures for 0.4% assume the right play on every hand.
- Forgetting side bets. Calculate them as their own line, with their own stake and edge, and add the results.
- Treating the estimate as a limit. It is an average. Real sessions can cost many times the figure, as the swing table shows.
- Leaving out tips. They are spending the edge does not include.
Card counting is not covered here as advice. It is not a reliable route to a lower cost, it takes sustained practice, and it may breach a casino’s house rules. The calculator assumes you do not count.
What should you do with your estimate next?
Compare the number with what you decided you can afford to spend, not with what you hope to get back. If the cost already feels too high, change one input and rerun it: a smaller bet, a slower pace, or a table where a natural pays 3 to 2. Set a time limit and a money limit before you start.
Our Safer play toolkit lists practical limits and tools. If gambling is costing more than you planned, the Get help page lists free, confidential support. To track what you actually spend, the planned gambling spend tracker keeps a running total, and the planned expected loss calculator works through other games.
How to read the result
Expected cost of the session is the average amount the game keeps, in units, over many sessions like this one. It is not a prediction for one night. The line under it shows the total bet (bet per hand x hands per hour x hours), and the bar shows your house edge on a scale of 0 to 30%. The last line shows the sum so you can check it by hand. Typical edges with basic strategy and a 3 to 2 payout are 0.3% to 0.8%, so 10 units a hand at 60 hands an hour costs about 2 to 5 units an hour. Where blackjack pays 6 to 5, edges are about 1.8% to 2.1%, which is about 11 to 13 units an hour for the same play.
How it is calculated
Total staked = bet per hand x hands per hour x hours played. Expected cost = total staked x house edge / 100. Written in one line: cost = bet x hands x hours x edge / 100. The tool shows the working, for example 10.00 x 60 x 2.0 x 0.50% = 6.00. Bet per hand must be at least 0.01, hands per hour a whole number of at least 1, hours played at least 0.1, and the house edge between 0 and 30. Results are rounded to two decimals. The calculator does not work out the house edge from the rules: you enter it, and the rule tables on this page show typical values.
Worked example
Worked example
With the starting values (bet per hand 10 units, 60 hands per hour, 2 hours, house edge 0.5%), total staked is 10 x 60 x 2 = 1,200 units. Expected cost = 1,200 x 0.5 / 100 = 6.00 units, or 3.00 units an hour. The same session at a 6 to 5 table, with the edge set to 1.8, costs 1,200 x 1.8 / 100 = 21.60 units.
Limits
The result is an average over many sessions, not a forecast: a real session can end well ahead or far behind, with a typical swing of about 1.15 times your bet times the square root of the number of hands. The edge is an input, so the result is only as good as the figure you choose; it depends on the payout, the dealer's soft 17 rule, doubling and splitting rules, deck count and how well you play. The tool assumes a constant bet and edge for the whole session. Enter the edge per initial bet, as quoted in this page's tables; the extra money from doubling and splitting is already included in those figures. Side bets, insurance, tips, comps, taxes and card counting are not modeled. For side bets, run the tool again with the side bet as its own bet and edge. The edge field accepts 0 to 30 only.