On this page
This house edge calculator compares 48 supported blackjack rule combinations and converts the selected percentage into an expected game cost. Choose the rules, an average original wager and a number of original rounds. The result also compares 3:2 and 6:5 blackjack payouts while keeping the other selected rules unchanged.
The figures use Wizard of Odds’ basic-strategy-with-cut-card results, checked November 1, 2026. They describe specified mathematical profiles, not verified rates for a named Alberta casino. Expected loss is a long-run average; it is not a forecast of what a particular session will lose.
Calculate house edge and expected cost
Match every selected and fixed rule. Enter starting wagers only: the source house-edge measure already accounts for doubles and splits under proper strategy. Include original rounds that push. All dollar illustrations use Canadian dollars.
Enable JavaScript to run the calculator. The formula, worked examples and references remain readable.
For the underlying ideas, read blackjack odds and house edge. Use the Blackjack Calculator for a completed hand’s settlement and the Hand Decision tool for its stated S17 opening-hand profile. These tools answer different questions.
How to run a useful comparison
- Record the game’s deck count, soft-seventeen rule, double-after-split rule, surrender rule and blackjack payout.
- Compare the fixed rules printed above with the game’s complete rules.
- Select the matching options; do not assume a table matches because its headline says blackjack.
- Enter an average original wager and the number of original rounds in the illustration.
- Select Calculate house edge and read the percentage separately from the dollar cost.
- Change one rule at a time to identify which difference is being compared.
When an input changes, the old result is replaced by a reminder to recalculate. Reset calculator restores eight decks, S17, doubling after splits, no surrender, a 3:2 payout, a $10 original wager and 100 original rounds. This creates a repeatable starting point for the worked examples below.
What the house-edge percentage measures
House edge expresses expected net loss relative to a defined wager base. This calculator uses an original-wager basis for blackjack. Its percentage belongs to the selected rules, the fixed assumptions and the stated strategy and dealing method. It is not the proportion of hands the dealer wins.
A percentage cannot be interpreted properly without that denominator. Multiplying an original-wager edge by a total that also includes every double and additional split stake changes the basis. The result may look precise while counting exposure that the source measure has already incorporated.
OLG PlaySmart explains that rules and playing choices affect blackjack’s house edge. This tool therefore does not assign one universal “blackjack rate.” The rules controls choose among specific source profiles, while the dollar inputs change only the illustrated volume to which that selected rate is applied.
The expected-cost calculation
Original-wager volume = average original wager × original rounds. The expected game loss then equals that volume multiplied by the house-edge percentage divided by one hundred. The calculator keeps the intermediate result unrounded and rounds the displayed dollar amounts to cents.
For the default profile, $10 × 100 gives $1,000 in original-wager volume. Applying 0.44686% gives $1,000 × 0.0044686 = $4.4686, displayed as $4.47. That is an expectation across the mathematical model, not an instruction to set aside exactly $4.47 for a session.
The result is labelled expected game loss rather than a guaranteed bill. A particular session can finish ahead or behind by much more. The form does not simulate card sequences, estimate the chance of exhausting a bankroll or calculate a confidence interval around the displayed average.
Read the fixed assumptions as well as the selectors
The seven form fields do not describe every possible blackjack rule. The stored profiles additionally allow doubling on any first two cards and splitting to a maximum of four hands. They do not allow resplitting aces or hitting split aces. The original-wager-only dealer-blackjack loss condition is also fixed.
Those conditions are part of every result. A game allowing doubles only on ten or eleven, for example, does not become a match simply because its deck count and payout appear in the form. Use an analysis with the actual restrictions instead of treating an unsupported rule as a small invisible detail.
The calculator is also specific to the source’s basic-strategy-with-cut-card measure. It does not silently substitute the source’s composition-dependent optimal result or continuous-shuffler result. Keeping the dealing and strategy basis consistent makes the selected percentages meaningfully comparable.
Selected rule-profile examples
| Profile | House edge | Expected loss per $1,000 original volume |
|---|---|---|
| 4 decks, S17, DAS, no surrender, 3:2 | 0.38699% | $3.87 |
| 6 decks, S17, DAS, no surrender, 3:2 | 0.42622% | $4.26 |
| 8 decks, S17, DAS, no surrender, 3:2 | 0.44686% | $4.47 |
| 8 decks, H17, DAS, no surrender, 3:2 | 0.66083% | $6.61 |
| 8 decks, S17, no DAS, no surrender, 3:2 | 0.58840% | $5.88 |
| 8 decks, S17, DAS, late surrender, 3:2 | 0.37104% | $3.71 |
| 8 decks, S17, DAS, no surrender, 6:5 | 1.80485% | $18.05 |
DAS means doubling after a split is allowed. Each line is a complete stored profile, with the remaining fixed rules unchanged. These are examples of mathematical rule sets. The table does not claim that any particular casino currently offers them or that a player uses the required strategy correctly.
Compare 3:2 and 6:5 without changing other rules
The result always includes a matched payout comparison. It looks up both payout versions with the same selected decks, dealer rule, DAS setting and surrender setting. This is more informative than comparing two percentages from different tables while overlooking several other rule changes.
In the default eight-deck example, 3:2 produces 0.44686% and 6:5 produces 1.80485%. Subtracting gives an increase of 1.35799 percentage points. At $1,000 of original-wager volume, the unrounded extra expected cost is $13.5799, displayed as $13.58.
This is a percentage-point difference, not a 1.35799% relative increase. Keep the terms separate when saving a comparison. The tool prints the two edges, their percentage-point gap and the extra dollar cost at the entered volume, so the calculation can be checked without guessing which meaning was intended.
The winning payment on an individual natural is a separate calculation. Our blackjack payouts guide compares the profit paid at 3:2 and 6:5. The rule’s effect on overall expectation also depends on how often the relevant outcomes occur under the model.
S17 and H17 are dealer rules
S17 means the dealer stands on soft seventeen; H17 means the dealer hits it. A soft total has an ace counting as eleven without exceeding twenty-one. The label describes the dealer’s compulsory action, rather than telling a player to stand on every seventeen.
Start from the default result and change only the dealer rule to H17. The edge becomes 0.66083%, giving $6.61 expected game loss at the same $1,000 volume. The dollar comparison is meaningful because the wager, round count and other rules have been held constant.
Return the selector to S17 before trying another comparison. Recording the full output is useful when testing several rules in sequence; otherwise an unnoticed setting can remain active. The dealer rules guide explains stopping rules and blackjack checks in more detail.
Doubles, splits and surrender affect the profile
The DAS selector changes whether the model permits a double on a hand formed by splitting. It is not a setting for how often you intend to double. Proper strategy for the resulting rule set remains part of the source’s calculation; the tool does not hold an unsuitable strategy fixed while changing the rules.
Late surrender is likewise a rule option, not a choice to surrender every difficult hand. Enabling it selects the profile where that option is available and used according to the model’s strategy. It does not insert early surrender or assume a surrender option in a game that lacks one.
Use the double-down, split and surrender guides to understand these actions. An edge comparison and an individual-hand decision are related, but they are not interchangeable outputs.
Count original rounds, including pushes
One original round begins with one initial main-hand wager. If that round later creates two hands through a split, do not enter two original rounds for it. Likewise, a double increases the stake during the existing round rather than creating a second initial round.
A push still belongs in the original-round count even though its net result is zero. Removing pushed rounds changes the denominator of the average. The form does not ask for the number of resolved wins and losses alone; it asks for the original decisions included in the illustration.
If you play more than one starting hand during a dealer deal, count the initial wagers consistently. Two separately started $10 hands contribute $20 of original volume. A later split of either hand remains inside that starting hand’s round for this purpose.
Use a weighted average when starting wagers vary
Suppose a hypothetical record contains forty original rounds at $5 and sixty at $10. Original volume is $200 + $600 = $800. Dividing by 100 original rounds gives an average starting wager of $8. Entering $8 and 100 therefore represents the same original-wager total.
Do not use the simple midpoint of five and ten, which is $7.50, because the two wager amounts occur different numbers of times. The weighted average preserves their actual contribution to volume. The calculator accepts cents, so a record producing an average such as $8.25 can also be represented.
This volume shortcut assumes the same applicable edge across the grouped wagers. It does not model a card-counting relationship between wager size and changing shoe composition, or combine different games under one invented average edge. Analyse those situations with their appropriate model and data.
More volume changes cost, not the selected edge
| Average original wager | Original rounds | Original volume | Expected game loss |
|---|---|---|---|
| $10 | 50 | $500 | $2.23 |
| $10 | 100 | $1,000 | $4.47 |
| $20 | 100 | $2,000 | $8.94 |
| $25 | 200 | $5,000 | $22.34 |
| $8 | 100 | $800 | $3.57 |
These are arithmetic illustrations rather than recommendations about how much or how long to play. Doubling the original volume doubles the model’s expected cost, while the chosen edge stays 0.44686%. Changing stakes does not improve the game’s rules or erase its mathematical disadvantage.
A deposit is not original-wager volume
A $100 deposit does not automatically mean $100 has been wagered. It can remain partly unused, or the same money can fund multiple starting wagers as returns are played again. The expected-cost formula needs the relevant original-wager volume rather than the amount transferred into an account.
For example, ten $10 original rounds create $100 of original volume. One hundred such rounds create $1,000, even if the player began with a smaller deposit and reused returned stakes. The calculator does not establish whether a bankroll would last long enough to complete that number of rounds.
Keep banking records separate from the illustration. The session-balance calculator can reconcile deposits, withdrawals and closing balances. Our withdrawals guide covers account movement, which should not be mistaken for a direct measure of the game’s theoretical edge.
Rounding, limits and small amounts
The stored source percentages retain their published precision. Dollar outputs round only after applying the rate to the full original volume. Multiplying a per-round amount that was first rounded to cents can give a different answer, particularly for small wagers or many rounds.
In the default $10 example, the unrounded expected cost per original round is $0.044686. Rounding that to four cents and multiplying by 100 produces $4.00, rather than the correctly rounded full-volume result of $4.47. Preserve the intermediate decimals when checking the calculation by hand.
The form accepts average original wagers from $0.01 to $1,000,000.00 and whole original-round counts from zero to one million. Enter numbers without commas or currency signs. A zero-round illustration has zero expected game loss while the rule profile still has a house-edge percentage.
Small positive expectations may display $0.00 after rounding. That does not turn the underlying rate into zero. The input limits support arithmetic examples; they are not statements about a casino’s minimum wager, maximum wager or permitted session length.
Expected loss is not a session prediction
The displayed average does not say what will happen next, how large a losing streak can be or whether the next session will finish ahead. The calculator has no probability distribution for session results, and it does not present the expectation as a likely narrow range.
A session can lose its available funds before reaching the entered round count. In that situation, the originally entered fixed-round illustration describes a volume that was not completed. This form does not adjust for a stop rule, a bankroll constraint or ending play after a particular win or loss.
For a meaningful review, preserve the question being asked. A theoretical cost for a specified volume and an observed result from one session are different records. A lucky session does not establish that the edge was zero, and a poor one does not reveal a new edge by itself.
The strategy assumption matters
These profiles assume proper basic strategy for their particular rules. Selecting H17 in this form does not make an S17-only decision chart suitable for that game. The site’s basic strategy guide and cheat sheet state their S17 scope explicitly.
The form does not collect a player’s mistakes or price them into the edge. It also does not assume composition-dependent perfect play, count cards or recommend changing stakes according to the remaining shoe. Its role is to compare the stated source profiles on one consistent measurement basis.
Betting progressions change starting amounts and therefore can change the volume used in a cost illustration. They do not, by themselves, improve the stored rule-set percentage. Read betting systems versus basic strategy for the separation between hand actions and starting stakes.
Fees, side bets and unsupported rules
The expected game loss excludes side wagers, participation charges, banking fees and other external costs. A side wager has its own amount, outcome rules and expectation; it should not be folded into the original main-hand volume while retaining the main-hand percentage.
For a simple cost illustration, a fixed twenty-five-cent participation fee on 100 rounds adds $25 of fees. Add that separately to the model’s expected game loss if those fees genuinely apply. This does not change the displayed main-game edge or make it an all-in percentage on a different denominator.
Special variants, different resplit limits, unusual split-ace permissions, restricted doubles and additional-stake losses to dealer blackjack are outside these profiles. The side-bets guide and blackjack terms glossary help identify questions requiring their own analysis.
Compare rates and costs separately
A lower percentage can still produce a larger expected dollar loss if the original volume is much higher. For example, a hypothetical 0.5% edge applied to $2,000 gives $10 of expected loss. A hypothetical 1% edge applied to $500 gives $5. The first percentage is lower, but its dollar cost is higher because the volumes differ.
To compare rule quality, hold the original volume constant. To compare two proposed arithmetic scenarios, retain each scenario’s own volume and report both the rate and expected cost. Labelling the comparison prevents an apparently favourable rate from hiding a much larger amount of play.
These two hypothetical percentages explain the formula; they are not extra stored blackjack profiles or claims about any operator. The form’s matched payout comparison uses the same entered volume on both sides, so its extra-cost figure isolates the selected payout difference within the supported rules.
What the displayed decimal places mean
Five decimal places preserve the source figure for reproduction and arithmetic. They do not mean a casino’s actual results have been measured to that precision. The output remains conditional on the rules and strategy assumptions, even when the percentage contains several digits.
When comparing a rounded summary such as 0.45% with the default 0.44686%, first check whether both refer to the same profile and measurement basis. A rounding difference is different from a rule difference. Preserve the stored value for calculations, then round the final dollar illustration only once.
Save enough detail to reproduce a result
Record the five selected rules, all fixed assumptions, the dealing-and-strategy basis, the average original wager, the original-round count and the date the figures were checked. Saving only “0.45%” discards the precision and conditions needed to reproduce the example.
For a screenshot, include the selectors and the result. For a written note, use the full profile rather than a casino name alone. Two tables in one venue can differ, and a source-checked mathematical profile does not verify which rules a named operator is offering today.
Use our Tools hub to move between expected cost, committed stakes, hand decisions and settled returns. Choosing the appropriate question first makes the output easier to interpret and avoids treating every number labelled a calculation as the same kind of evidence.
House edge calculator questions
It is expected net loss relative to the original main-hand wager under the selected profile. It is not the percentage of hands lost. For the default 0.44686% profile, $1,000 of original-wager volume gives $4.4686 of expected game loss, displayed as $4.47. A particular session can differ substantially. It compares 48 combinations of four, six or eight decks, S17 or H17, double-after-split permission, late surrender and 3:2 or 6:5 natural payment. Fixed rules include any-two-card doubles, at most four split hands, no resplitting or hitting split aces, and original-wager-only loss to dealer blackjack. It uses the source’s basic-strategy-with-cut-card figures. No. This measure uses original wagers, and its edge already incorporates doubles and splits under the stated strategy. Count the initial main-hand stakes and original rounds, including pushes. A separate commitment worksheet can add every stake, but that total uses a different denominator. Only the natural-blackjack payout changes; the other selected rules and entered volume stay the same. In the default eight-deck profile, the source edges are 0.44686% and 1.80485%. Their gap is 1.35799 percentage points, giving $13.58 extra expected cost at $1,000 of original-wager volume. The stored figures use the source’s cut-card measurement. Its continuous-shuffler and composition-dependent optimal outputs are separate results. Use an analysis matching the game’s actual dealing method and rules rather than treating one stored percentage as a universal blackjack rate. Use total original-wager volume divided by original rounds, provided the same edge applies throughout. Forty rounds at $5 and sixty at $10 create $800 of original volume, or an $8 average over 100 rounds. The simple midpoint of the two stakes would not preserve their different frequencies. No. It is a long-run mathematical average for the entered volume. The calculator does not estimate a session-result range, completion probability or chance of running out of funds. A planned round count may never be reached, and the displayed expectation is not a spending limit or a safe bankroll amount. No. The result covers the stated main-hand model. Record side wagers and external fees separately, using their own applicable amounts and rules. Deposits and withdrawals are account movements rather than original-wager volume. A deposit can remain unused or fund repeated wagers as money is returned. Dollar outputs round to cents after applying the source rate to the full volume. A positive amount smaller than half a cent can display as $0.00 without making the underlying edge zero. Zero rounds also produces zero expected cost while leaving the selected profile percentage unchanged.What does the house-edge percentage measure?
Which rules and dealing method does this calculator support?
Should I add double and split stakes to the wager volume?
What changes in the matched 3:2 and 6:5 comparison?
Can I use a continuous-shuffler game with these percentages?
How should I enter different starting wagers?
Does the expected cost predict my session or a safe bankroll?
Are banking fees, side bets and actual withdrawals included?
Why can a small expectation display as $0.00?
References
All 48 rule profiles checked again in the live source calculator November 1, 2026. The 48 rule profiles use the source calculator’s basic-strategy-with-cut-card output. Dollar examples are calculations from those percentages and the stated original-wager volume.