Blackjack Odds Explained: Probability, Mathematics & Winning Chances
Why blackjack's odds are worth understanding directly
Most players learn blackjack strategy as a set of memorized actions — hit here, stand there, double this — without ever seeing the underlying probability that makes those actions correct. Understanding the actual numbers behind the game changes that: once you can see why a dealer showing a 6 busts far more often than one showing a 10, the strategy chart stops feeling like an arbitrary set of rules to memorize and starts feeling like the obvious conclusion it actually is. This guide walks through the core probabilities that drive every blackjack decision — the chance of specific starting hands, the chance of busting on a hit from any given total, the dealer's bust probability by upcard, and your overall win, loss, and push rates using correct strategy. For the money-focused counterpart to this guide — house edge, RTP, and expected return in dollar terms — see blackjack RTP explained.
A quick note on precision: the figures throughout this guide are the widely published, commonly cited approximations used across blackjack literature, based on a standard multi-deck composition. Exact figures shift very slightly depending on deck count and how many cards have already been seen in a given shoe, but the numbers here are accurate enough to be genuinely useful for understanding why basic strategy works the way it does.
It's also worth being upfront about what this guide is and isn't. It won't teach you to predict individual hands — nothing can, since each hand's specific cards are genuinely random within the bounds of the probabilities described here. What it will do is explain the mathematical machinery underneath every basic strategy recommendation you've encountered elsewhere on this site, so that "hit hard 16 against a dealer 10" stops being a memorized instruction and becomes a conclusion you can derive and verify for yourself.
What "odds" means in this context
Throughout this guide, "odds" and "probability" are used to mean the same thing: the mathematically expected frequency of an event across a very large number of repetitions, expressed as a percentage. This is distinct from betting odds in the traditional sense (like "3:2" or "6:5"), which describe payout ratios rather than the underlying likelihood of an outcome — though the two are related, since a fair payout ratio is one that matches the true probability of the event it's paying out on. Keeping this distinction clear matters specifically for insurance, where the payout odds (2:1) and the true probability (roughly 30.8%) don't match closely enough to make the bet favorable, exactly the comparison worked through earlier in this guide.
The building block: card composition
Before any specific probability makes sense, it helps to understand the deck composition itself. A standard 52-card deck has 13 ranks, and in blackjack, four of those ranks (10, Jack, Queen, King) are all worth the same value: 10. This means roughly 30.8% of all cards in the deck are worth 10 — nearly a third — while every other rank (2 through 9, and Ace) makes up roughly 7.7% each. This single fact — that 10-value cards are almost three times as common as any other individual rank — underlies almost every probability calculation in this guide, from bust rates to the odds behind insurance.
The probability of a natural blackjack
A natural blackjack requires an Ace plus a 10-value card, in either order, as your first two cards. Working through the combinatorics of a standard multi-deck shoe, this works out to approximately 4.8% of all hands — roughly 1 in 21. It's common enough that most players see several in an ordinary session, but rare enough that its enhanced 3:2 payout meaningfully rewards the moments it happens. Full detail on how the natural itself works, including the specific payout mechanics, is in what is a natural blackjack?
The probability of the dealer having a natural
This is the exact math behind why insurance is offered, and why it's usually a bad bet. When the dealer shows an Ace, the probability their hole card is a 10-value card — completing a natural — is approximately 30.8%, matching the overall proportion of 10-value cards in the shoe (since the Ace itself doesn't change the remaining proportion meaningfully). Insurance pays 2:1, which would need the true probability to be exactly 1-in-3 (33.3%) to break even. Since the actual probability (30.8%) sits below that threshold, insurance carries a built-in disadvantage for the average player. Full detail on this specific bet is in blackjack insurance explained.
The probability of busting on a hit, by your current total
This is one of the most directly useful probability sets in the entire game, because it's the mathematical foundation of every hit-or-stand decision, and it's worth committing to memory even in rough terms before moving on to the rest of this guide. The chance of busting on a single additional card depends entirely on how many points of "room" your current total has before exceeding 21:
| Your hard total | Cards that would bust you | Approximate bust probability |
|---|---|---|
| 12 | 10-value only | ~31% |
| 13 | 9 or 10-value | ~39% |
| 14 | 8, 9, or 10-value | ~46% |
| 15 | 7, 8, 9, or 10-value | ~54% |
| 16 | 6, 7, 8, 9, or 10-value | ~62% |
| 17 | 5 through 10-value | ~69% |
Look at this table alongside basic strategy and the logic becomes immediately clearer: even at a hard 16 — the hand basic strategy still recommends hitting against a dealer 10 — you're busting nearly two-thirds of the time on that single card. Hitting isn't correct there because it's safe; it's correct because the alternative (standing) is measurably worse against that specific dealer upcard, even accounting for how often the hit busts. This is exactly the tension covered in when to hit or stand in blackjack and hard hands in blackjack explained.
Why soft hands change this entire table
Every probability in the table above assumes a hard total, where any card that pushes you over 21 busts the hand outright. Soft hands break this relationship entirely — because the Ace can drop from counting as 11 to counting as 1, a soft hand has a 0% chance of busting on a single additional card, regardless of the total. This is the core mathematical reason soft hands are hit and doubled far more aggressively than hard hands at similar totals, covered in full in soft hands in blackjack explained.
Dealer bust probability by upcard
This is arguably the single most important probability table in the entire game, since it's the direct justification for why the dealer's upcard — not just your own hand — determines the correct strategy in the hard 12–16 range. Because the dealer must hit until reaching 17 and has no discretion to stop early, their bust probability is a fixed function of their starting upcard:
| Dealer upcard | Approximate bust probability |
|---|---|
| 2 | ~35% |
| 3 | ~38% |
| 4 | ~40% |
| 5 | ~43% |
| 6 | ~42% |
| 7 | ~26% |
| 8 | ~24% |
| 9 | ~23% |
| 10 | ~21% |
| Ace | ~12–17% |
Notice the sharp divide: upcards of 2 through 6 all carry a bust probability above 35%, while upcards of 7 through Ace all sit well below 30%. This divide is the entire mathematical basis for basic strategy's "weak dealer upcard vs. strong dealer upcard" framework — it's not a rough intuition, it's a direct reflection of these numbers. It's also why hard 12 specifically stands only against a 4, 5, or 6 rather than the full 2–6 range: the dealer's bust probability on a 2 or 3 (around 35–38%), while still elevated, isn't quite high enough to outweigh the risk of standing on such a weak total. Full detail on this specific range is in blackjack dealer rules explained.
Your overall win, loss, and push probability
Zooming out from individual decisions to a full session, a player using correct basic strategy on a standard table sees roughly the following distribution of outcomes across all hands played:
- Win: approximately 42–43%
- Loss: approximately 48–49%
- Push: approximately 8–9%
At first glance, this looks like blackjack is a losing proposition even with perfect play — you lose more hands than you win. But this is exactly where blackjack's structure matters: wins aren't all equal to losses in size. Natural blackjacks pay 1.5x, doubled hands pay 2x when they win (and only cost 2x when they lose, same as any bet), and split hands create additional independent winning opportunities. The reason correct-strategy blackjack carries only a small house edge despite a losing-hands majority is that the value of wins, weighted by these payout multipliers, comes close to balancing out the higher frequency of losses. This is a genuinely counterintuitive result worth sitting with: winning fewer than half your hands is entirely consistent with a low overall house edge, once you account for how wins and losses are actually sized.
The dealer's final total, blended across every upcard
The bust-probability table above breaks the dealer's outcome down by upcard, but it's also useful to see the dealer's overall final-total distribution blended across every possible upcard, weighted by how often each occurs. Across a large sample of hands on a standard multi-deck, dealer-stands-on-soft-17 table, the dealer's final outcome breaks down roughly as follows: bust around 28–29% of the time, finish on 17 around 14–15%, on 18 around 13–14%, on 19 around 13%, on 20 around 17–18%, and on 21 (including naturals) around 11–12%. Notice that 20 is the single most common non-bust final total — a direct consequence of how common 10-value cards are, since a huge share of dealer hands pass through a strong initial total built on a 10-value card. This is worth keeping in mind alongside the bust-probability table: even though the dealer busts more often than they land on any single specific total, "the dealer doesn't bust" is still the more likely outcome overall, at roughly 71–72% combined across every non-bust total.
The probability of being dealt a pair
Since pairs unlock the option to split, it's worth knowing roughly how often one comes up. Because four different ranks (10, Jack, Queen, King) all count as the same value for blackjack purposes, "pairs" in the blackjack sense are more common than in a game where every rank is distinct — a 10-Jack, Jack-Queen, or Queen-King starting hand all count as a matching pair of 10-value cards for splitting purposes, even though the specific card ranks differ. Taking this into account, a player is dealt some form of splittable pair on approximately 1 in 15 to 1 in 16 hands — meaningfully more often than the roughly 1-in-17 rate you'd expect if you only counted exact rank matches (like two literal Kings), precisely because of how many different 10-value combinations count as a pair for splitting purposes.
How many distinct starting hands are actually possible?
As a matter of pure combinatorics, there are 1,326 distinct two-card combinations possible from a standard 52-card deck (52 choose 2). But because suits don't matter in blackjack — only rank and, for 10-value cards, the shared value — the number of strategically distinct starting hands is much smaller: 10 hard totals with no pair option, a handful of soft totals from Ace combinations, and 10 distinct pair types (2-2 through 10-10, treating all 10-value pairs as one category, plus A-A). This is part of why basic strategy is learnable at all despite blackjack's underlying complexity — the astronomical number of raw card combinations collapses down to a genuinely manageable set of a few dozen strategically meaningful situations once you account for which distinctions actually matter to the correct decision.
How probability and strategy interact: a worked example
Take hard 16 against a dealer 7. Your bust probability on a hit is roughly 62%. That sounds terrible in isolation — but compare it to the alternative. Standing on 16 wins only when the dealer busts (dealer 7's bust probability is about 26%) or, more precisely, loses whenever the dealer doesn't bust and ends up with a total higher than 16, which — given a dealer showing 7 commonly resolves to 17 — is the large majority of the remaining 74% of outcomes. Weighing these two paths against each other across many repetitions is exactly the kind of calculation basic strategy has already solved for you: hitting produces a better long-run result than standing in this specific matchup, even though the bust probability alone looks alarming. This is the core lesson probability brings to blackjack strategy — no single number, viewed in isolation, tells you the correct decision. It's the comparison between the available options that matters.
Probability and doubling
The probability tables above also explain why certain doubling decisions are so strongly favored. Hard 11 is doubled against nearly every dealer upcard because busting on the next card requires drawing specifically nothing (an 11 can only reach a maximum of 21 on any single card, since even a 10-value card brings it to 21 exactly) — a hard 11 has a 0% chance of busting on the next card, the same structural safety as a soft hand, just via a different mechanism. This is why hard 11 sits alongside soft hands as one of the strongest doubling opportunities in the entire game: no bust risk, combined with a strong likely outcome. Full detail on every doubling situation is in when to double down in blackjack.
Probability and splitting
Pair-splitting probability is a little different in character, since it's less about a single card's bust risk and more about comparing the combined value of one hand against two. A pair of 8s (hard 16) has a rough win probability, played as a single hand, well below 50% against most dealer upcards — hard 16 is simply a weak total. Split into two hands, each starting from an 8, and each new hand has a meaningfully higher individual win probability than the combined 16 did, because an 8 is a fresh, average-strength starting card rather than a maxed-out weak total. This is the probabilistic logic behind "always split 8s," covered in full in when to split pairs in blackjack.
The probability behind surrender decisions
The surrender recommendations covered in blackjack surrender explained are themselves a direct product of probability comparison. Take hard 16 against a dealer 10: playing the hand out with the best available option (hitting) wins clearly under 25% of the time in this specific matchup, once you combine the roughly 62% bust probability on the hit itself with the further chance of losing even after successfully avoiding a bust against a dealer showing a strong 10. Compare that to surrender, which guarantees losing exactly half the bet, every time — mathematically equivalent to winning 50% and losing 50% on a hand played to completion. Since playing the hand out wins meaningfully less than 25% of the time in this matchup, guaranteeing a 50%-equivalent outcome through surrender comes out ahead on average, despite feeling like a concession in the moment.
How probability compounds across a full session
A single hand's probability doesn't tell you much on its own — the real value of understanding these numbers comes from how they compound across many repetitions, a principle sometimes called the law of large numbers. A dealer bust probability of 42% on a weak upcard means the dealer fails to bust more often than not on any individual hand, but across a thousand hands where that specific matchup comes up repeatedly, the actual bust rate converges very closely to that 42% figure. This is exactly why basic strategy is described as "correct in the long run" rather than "correct every time" — any individual hand's outcome is still essentially random within the bounds of its probability, but the aggregate result across a large enough sample reliably reflects the underlying numbers. This is also why a single losing session, even one where every decision followed basic strategy perfectly, says very little about whether you're playing correctly — the sample size of one session is far too small for the underlying probabilities to have meaningfully asserted themselves yet.
How blackjack's odds compare to other casino games
Context helps make these numbers meaningful. A coin flip has a 50% probability by definition — blackjack's roughly 42–43% win rate sits below that, which can look unfavorable in isolation. But unlike a simple even-money coin flip, blackjack's payout structure (naturals at 3:2, doubled and split hands scaling the bet size specifically in favorable spots) means the game's overall expected return sits much closer to break-even than the raw win percentage suggests — closer, in fact, than the equivalent raw win percentage would produce in a game with uniform, unweighted payouts. This is a big part of why understanding the underlying probability structure matters: judging blackjack purely by "do I win more hands than I lose" (no) gives a misleading picture compared to judging it by actual expected return (a small, manageable house edge with correct play), which is covered in dollar terms in blackjack RTP explained.
Why these odds shift slightly as a shoe is dealt
Every probability in this guide is calculated against a full, freshly shuffled shoe. As cards are dealt and not reshuffled (common on physical single- and double-deck tables, less common on continuously shuffled software-dealt or many live-dealer formats), the remaining composition shifts slightly — a shoe that's seen an unusually high number of low cards, for example, is now slightly richer in high cards than a fresh shoe would be, changing the true probabilities by a small amount. This is the exact principle card counting is built on, tracking these shifts to gain a real, if narrow, edge in favorable conditions. For the overwhelming majority of players on the overwhelming majority of online tables — where continuous shuffling or frequent reshuffles are standard — the fresh-shoe probabilities in this guide remain the accurate, practical reference.
Approximate win probability by starting total
It's useful to see, roughly, how your win probability shifts across different starting totals when played correctly against an average dealer hand (blended across all upcards). These figures are necessarily approximate, since the exact number depends heavily on the specific dealer upcard in any given hand, but they illustrate the overall shape:
- Hard 20: wins the clear majority of the time, commonly cited north of 70%.
- Hard 17: a positive but much closer figure, commonly in the low-to-mid 50% range against an average dealer hand.
- Hard 12–16 (played correctly): these hover much closer to even, and in several cases below it, which is precisely why this range is the most strategically demanding — no action available fully overcomes the weakness of the starting total itself.
- Soft 18–20: win rates well above even, benefiting from both a strong total and the safety of the flexible Ace.
The pattern across this list reinforces a point made earlier in this guide: no strategy, however correctly played, can turn a fundamentally weak starting total into a favorite. What basic strategy does is ensure you're extracting the best available result from whatever total you're dealt, not manufacturing an advantage that isn't there.
The gambler's fallacy and why past hands don't affect future odds
A common misconception worth addressing directly: the belief that a table is "due" for a natural blackjack after a long stretch without one, or that a dealer who has busted several hands in a row is somehow more or less likely to bust again on the next hand. In a continuously shuffled or software-dealt game, each hand draws from a probability distribution that's completely independent of previous hands — there's no mechanism by which past outcomes influence future ones. This mistaken belief, commonly called the gambler's fallacy, leads directly to some of the costly bet-sizing mistakes covered in common blackjack mistakes beginners make, particularly chasing losses under the assumption that a win is somehow "overdue." Even on a physical, dealt-down shoe where card removal does technically shift subsequent probabilities slightly (the basis of card counting), the effect is far too small on any single hand to justify the kind of dramatic bet-size changes the gambler's fallacy tends to inspire in players who believe in it.
Probability calculators and simulation tools
For players who want figures more precise than the standard approximations in this guide, dedicated blackjack probability calculators and simulators exist that let you specify an exact rule set — deck count, soft-17 behavior, doubling and splitting restrictions — and generate precise bust probabilities, dealer outcome distributions, and win rates for that specific combination. These tools work by running (or mathematically solving) millions of simulated hands under the specified rules, the same underlying approach used to originally derive basic strategy itself, covered in more detail in blackjack basic strategy guide. They're a useful way to move from the general, widely-applicable numbers in this guide to figures matched exactly to a specific table you're planning to play.
Odds, variance, and why a single session tells you very little
It's worth being direct about a common misunderstanding: knowing the exact odds behind a hand doesn't tell you what will happen in any specific session. A 62% bust probability still means you don't bust more than a third of the time; a dealer with a 42% bust probability on a weak upcard still fails to bust more often than not. Individual sessions are dominated by variance, not by the underlying probabilities directly — the odds in this guide describe the long-run average across a very large number of hands, not a prediction for your next ten. For a full discussion of how variance affects real sessions despite these fixed underlying probabilities, see blackjack variance explained.
Do the odds differ between American and European formats?
The core probabilities in this guide — bust rates, natural frequency, dealer outcome distribution — are essentially identical between American (hole-card) and European (no-hole-card) dealing formats, since both draw from the same underlying deck composition and follow the same dealer hit/stand rule. What differs isn't the raw probability of any given card or hand, but how that probability interacts with your decision-making: under the European format, you can't factor in an early dealer-blackjack check before deciding whether to double or split against a dealer Ace or 10, since that information simply isn't available yet when you act. This changes the optimal decision in a handful of specific spots without changing the underlying odds themselves. The full comparison is in European vs American blackjack.
Probability and playing multiple hands at once
Playing two or three hands simultaneously at a software-dealt table doesn't change the probability of any individual hand — each hand is dealt independently from the same underlying distribution, and the numbers throughout this guide apply identically to every hand you're playing at once. What does change is the combined variance of your round as a whole, since you're now exposed to multiple independent probabilistic outcomes simultaneously rather than one. Combining several independent hands with similar win probabilities tends to smooth out the extreme swings of any single hand slightly over a full session, though it also multiplies your total bet exposure per round — a bankroll consideration covered in blackjack variance explained, not a change to the underlying odds covered here.
Bringing the numbers together
Every probability covered in this guide — natural frequency, bust rates by total, dealer bust rates by upcard, and overall win/loss/push distribution — feeds into the same underlying structure that produces basic strategy's recommendations. None of these numbers need to be memorized precisely to play well; what matters is internalizing the shape of the relationships they describe, since that shape is what makes every individual strategy decision make sense on its own terms rather than as a rule taken on faith.
Frequently asked questions
Are these probabilities exactly the same on every table? They're very close, but not identical — deck count and specific rule variations shift the numbers by small amounts. The figures in this guide reflect a standard multi-deck game and are accurate enough to explain the strategic logic behind basic strategy, even if a specific table's exact numbers differ slightly.
Why does the dealer's bust probability matter more than my own hand in some situations? Because the dealer's upcard is a strong, calculable predictor of their entire hand's outcome, while your own total in the 12–16 range is inherently weak regardless of the dealer's card. Basic strategy weighs both pieces of information together, which is why the correct action in this range depends so heavily on the specific dealer upcard.
Is it true that I lose more hands than I win even with perfect strategy? Yes — this is one of the most counterintuitive but well-established facts about blackjack. The game's low house edge comes from how wins and losses are sized (naturals, doubles, and splits paying or costing more than a standard bet), not from winning a majority of hands.
Do these odds change if I'm playing live dealer versus software-dealt blackjack? No — the underlying probabilities are a function of the deck composition and the rules in play, not the format. What can differ between formats is how the shoe is managed (continuous shuffling versus a dealt-down physical shoe), which affects how much the odds shift over the course of a session, not the baseline fresh-shoe numbers themselves.
Where can I see these odds calculated for my exact table's specific rules? Dedicated blackjack probability calculators and strategy trainers exist that let you input a specific rule set (deck count, soft-17 behavior, and so on) and see the precise resulting numbers — useful if you want figures more exact than the standard approximations used throughout this guide.
Is it true that a table can be "due" for a natural after a long gap without one? No — this is the gambler's fallacy. In a shuffled or continuously-dealt game, every hand is dealt from a probability distribution independent of previous hands, so a long gap without a natural has no bearing on the likelihood of the next hand producing one.
Why does the dealer's overall bust rate (around 28%) seem low compared to some individual upcards in the table above? Because the blended figure averages across all ten possible upcards, including the strong ones (7 through Ace) where the dealer busts far less often than 28%. The individual upcard table gives the more useful, situation-specific number for any actual decision you're making.
Does knowing these probabilities help me beat the house edge? Not directly — the probabilities themselves don't create an edge; they're the reasoning behind why basic strategy is structured the way it is. Applying basic strategy correctly is what captures the value these probabilities describe, not simply knowing the numbers in isolation.
Are the win/loss/push percentages the same for every rule set? They shift slightly with different rules (deck count, soft-17 behavior, and so on), but the overall shape — losing a plurality of hands while maintaining a low house edge through payout weighting — holds true across virtually every standard blackjack rule combination.
