Blackjack FAQ: 50 Questions Answered

Blackjack is a card game where you try to get a hand totaling closer to 21 than the dealer's, without going over. You play against the dealer (not other players). Each card has a value: number cards count their face value, face cards (J/Q/K) count 10, and Aces count 1 or 11 (whichever helps you most). You're dealt two cards, then choose to take more (hit) or keep your hand (stand). Closest to 21 without busting wins.

A blackjack (also called a "natural") is your first two cards totaling exactly 21 — meaning one Ace plus any 10-value card (10, J, Q, or K). It's the strongest possible hand. In most games, blackjack pays 3:2 (a $10 bet wins $15), though some casinos pay only 6:5 ($12 on $10), which is much worse for the player.

It varies. Single-deck games still exist but are rare and often have worse rules to compensate. Most modern blackjack uses 6 or 8 decks dealt from a "shoe." More decks slightly favor the house (about 0.5% extra edge for 8-deck vs single-deck with otherwise identical rules), partly because card counting becomes less effective with more decks.

Hit: take another card. Stand: keep your current hand. Double down: double your bet and take exactly one more card (usually only allowed on your first two cards). Split: if your first two cards are a pair, separate them into two hands with a matching bet on each. These four (plus surrender at some tables) are your only decisions in standard blackjack.

Surrender lets you give up your hand and lose only half your bet, rather than playing it out. Late surrender (after the dealer checks for blackjack) is the standard form. Surrender is mathematically correct in specific situations — hard 16 vs dealer 9, 10, or Ace, and hard 15 vs dealer 10. Most casinos don't offer it; if your table does, use it correctly. It saves about 0.07% of house edge over the long run.

Insurance is a side bet offered when the dealer's up-card is an Ace. You bet half your original wager that the dealer has blackjack. It pays 2:1 if the dealer does. Mathematically, insurance is a bad bet — the house edge is about 7%, much worse than the base game. The only exception is if you're counting cards and the count is highly positive. For everyone else: never take insurance.

The dealer follows fixed rules and has no choices. The two main variants: "S17" tables where the dealer stands on all 17s including soft 17 (Ace-6), and "H17" tables where the dealer hits on soft 17. S17 is better for the player by about 0.2% house edge. The dealer must hit any hand totaling 16 or less and must stand on hard 17 or more. The dealer doesn't double, split, or take insurance.

A tie is called a "push." Your bet is returned — you don't win or lose. The one exception: if both you and the dealer have natural blackjack (two-card 21), it's still a push at most tables. A few rare tables call this "player wins" but it's vanishingly uncommon and shouldn't affect your decisions.

Basic strategy is the mathematically optimal way to play every possible hand against every possible dealer up-card. It was derived in 1956 by four U.S. Army mathematicians (the "Four Horsemen") and refined by Edward Thorp in the 1960s using computers. Following basic strategy perfectly gives the player about 0.5% house edge with standard rules — the best you can do without counting cards.

Yes, basic strategy says hit (or surrender if available). It feels wrong because you're likely to bust, but the math is clear: 16 stands almost zero chance against the dealer's likely 20. Hitting gives you a small chance to improve to 17-21. Standing gives you almost certain loss. Surrender (losing half your bet) is even better if your table offers it. The intuition to stand is wrong.

Always double on 11 vs dealer 2-10. Double on 10 vs dealer 2-9. Double on 9 vs dealer 3-6. Double on soft hands (A-2 through A-7) against dealer 4, 5, or 6 in most cases. Never double on 12+ hard hands or on Ace-8 / Ace-9. The exact chart varies slightly with rules. The basic principle: double when you're likely to make a strong hand and the dealer is in a weak position (showing 4-6).

Always split Aces and 8s. Never split 5s or 10s. The rest depends on the dealer's up-card. Split 2s, 3s, or 7s against dealer 2-7. Split 6s against dealer 2-6. Split 9s against dealer 2-9 except 7. Split 4s only if you're allowed to double after splitting and dealer shows 5-6. The 8s are the most-important rule: a hard 16 is the worst hand in blackjack, so splitting it gives you two new chances.

No. Basic strategy minimizes your loss rate to about 0.5% of what you bet — it doesn't make you a winner. You'll still lose in most sessions due to variance. What basic strategy does: ensures you're not making mathematical errors that cost more. Over thousands of hands, it produces the best possible outcome given the house edge. Many players assume basic strategy makes them money; it doesn't. It just minimizes loss.

Yes, soft hands (containing an Ace counted as 11) require different play than hard hands of the same total. Soft 18 (A-7) is actually a marginal hand — you should double against dealer 3-6, stand against 2/7/8, and hit against 9/10/A. Soft 17 (A-6) should generally be hit or doubled. The flexibility of the Ace changes the math significantly compared to hard 17 or hard 18.

In physical casinos: generally yes for basic strategy charts. Most casinos tolerate them as long as you don't slow the game down significantly. Some pit bosses may comment but rarely enforce removal. Online: absolutely yes — keep a chart open beside you. The chart isn't cheating; it's using public mathematical knowledge. Casinos prefer recreational players over skilled ones, but using a chart is firmly in the recreational-player range.

With basic strategy and standard rules (6-deck, dealer stands soft 17, blackjack pays 3:2, double after split allowed), the house edge is about 0.5%. This means for every $100 you bet, you expect to lose 50 cents on average over the long run. Blackjack has one of the lowest house edges of any casino game when played correctly — making it appealing for skilled players.

Roughly: you win about 42% of hands, lose 49%, and push (tie) about 9%. The reason the house wins more often: when both you and the dealer bust, you lose first. This is the structural advantage built into the rules. Your better-than-49% win rate plus the 3:2 blackjack payout almost — but not quite — compensates for the busting-first disadvantage.

RTP (Return to Player) is the inverse of house edge. A game with 99.5% RTP has a 0.5% house edge. RTP terminology is more common for slots and online RNG games. For blackjack with basic strategy on standard rules, RTP is approximately 99.5%. With suboptimal play (most casual players), actual realized RTP drops to about 97-98% because of strategy mistakes.

Variance is the swing range around your expected loss rate. The house edge is small (0.5%) but variance is large — standard deviation of about 1.15 betting units per hand. Over 100 hands you might be up or down 30+ betting units just from variance. Over 10,000 hands, variance shrinks relative to expected value. Most player "wins" are variance, not skill. Most player "losses" are variance plus house edge.

About 4.83% in a single-deck game (4.75% in 6-deck). Roughly 1 in 21 hands. The math: 4 Aces out of 52 cards × 16 ten-value cards out of remaining 51 cards, times 2 (Ace first or ten first). The slight decrease with more decks is one of several reasons single-deck has slightly better player odds when all other rules are equal — though casinos usually compensate single-deck games with worse rules.

On a 3:2 game, a $10 blackjack pays $15. On a 6:5 game, the same blackjack pays only $12. You get $3 less per blackjack — and blackjacks happen about 4.8% of the time. The math: 6:5 increases the house edge by approximately 1.4 percentage points. A 0.5% game becomes a 1.9% game — roughly quadrupling the house's expected take from you. 6:5 tables are designed to extract more from recreational players who don't notice.

Hard 16. It's mathematically terrible because: you're likely to bust if you hit (any card 6 or higher busts you), and you're almost certain to lose if you stand (the dealer will probably make 17-21). Worst of all is 16 against a dealer's 10 — about 77% of hands lose regardless of decision. Hitting is marginally better than standing, surrender is best if available. There's no good play; only the least-bad one.

Yes, in essentially every jurisdiction. Counting cards is using publicly available mathematical knowledge and your own memory — it's not cheating in any legal sense. The 1982 New Jersey Supreme Court ruling Uston v. Resorts International Hotel established this conclusively for Atlantic City. Casinos can refuse to serve you (and will, if they suspect counting) but they cannot have you arrested. Practically, getting backed off from casinos is the main risk, not legal consequences.

Each card gets a value: 2-6 count as +1, 7-9 count as 0, 10/J/Q/K/A count as -1. You keep a running count as cards are dealt. To convert to true count (which determines your edge), divide running count by decks remaining. When true count is positive, the remaining deck has more high cards than low cards — favoring you. You raise your bets in positive counts and reduce them when neutral or negative.

No, not at RNG (computer-generated) tables. Every hand starts a fresh shuffle — there's no persistent deck state to track. Counting is structurally impossible. Live-dealer online tables theoretically allow counting (real cards, real shoe) but in practice are unprofitable: shallow penetration, eight-deck shoes, and automated bet-pattern monitoring make it not worth the effort. Realistic counting still requires hand-shuffled physical casino tables.

Far less than movies suggest. A skilled solo counter with adequate bankroll might achieve a 0.5-1.5% edge. At $100 average bets and 80 hands/hour, that's $40-$120/hour gross — minus variance losses, casino backoffs cutting sessions short, travel costs, and the time to find playable games. Net annual earnings for full-time counters are typically $30-80k, not the millions implied by movies. The MIT Team's total over 20+ years was "several million," not the inflated Mezrich figures.

If you're serious about counting and play often, eventually yes. Modern casino surveillance is sophisticated: facial recognition databases (Griffin, Biometrica) shared between properties, bet-pattern monitoring software, and pit bosses trained to spot the signature changes in betting that indicate counting. Most counters extend their playing time through "cover" — looking like a recreational player while counting accurately. Even with cover, eventual backoffs are normal.

The Hi-Lo system itself is straightforward — you can learn the basic mechanic in an afternoon. What's hard is executing it accurately at casino-table speed while pretending to play recreationally, managing the math under social pressure and table noise, and maintaining the discipline to vary your bets correctly. Most people who try to count fail at the execution layer, not the math. Months of practice are realistic to be casino-ready.

At licensed sites (UKGC, MGA, NJ DGE, AGCO, etc.) — yes. The random number generators are audited by independent firms like eCOGRA, iTech Labs, and GLI. Audit reports are public. The math is verifiable. At unlicensed sites — no guarantees, and a real history of rigging. The fairness question depends entirely on regulator and auditor, not on the platform being "online."

At licensed and audited sites: essentially no. The math would be detected by audit, and casinos that get caught lose their license — a catastrophic outcome they avoid. At unlicensed offshore sites: history shows multiple documented cases of rigging. The risk is in operator selection, not in the underlying technology. The PRNG and Fisher-Yates shuffle architecture used by legitimate sites is mathematically sound when properly implemented.

RNG: computer generates the card sequence using algorithms (typically ChaCha20 or AES-CTR) plus the Fisher-Yates shuffle. Every hand starts fresh. Live dealer: a real human deals real cards from a real shoe in a studio, captured by OCR and streamed to your screen. RNG is faster (200+ hands/hour vs 60-80) and counter-proof. Live dealer feels more like physical casino but at slower pace.

Through skilled basic strategy alone: no — you'll lose 0.5% of your wagers over time. Through counting: only at live-dealer tables, and only marginally because of structural limits. Through bonus hunting: occasionally yes for sophisticated bonus hunters, but most bonuses restrict blackjack contribution to 10-20% of wagering requirements, making the math weak. For most players, online blackjack is entertainment with a known cost, not a profit source.

Sometimes, with careful reading. A typical "100% match up to $500" bonus with 30x wagering requires you to wager $15,000. At 0.5% house edge, expected loss is about $75 — netting the bonus to roughly $425 net value. BUT many bonuses contribute blackjack at only 10-20%, meaning you need 5-10x more wagering, which makes most bonuses worth nothing on blackjack. Always read the specific terms before depositing.

Most reputable RNG sites offer $0.10 to $1 minimums on standard tables. Live dealer tables typically start at $1-5. Compare to physical casinos where $5-25 minimums are common, and $50+ at quality 3:2 tables on the Vegas Strip. The low online minimum is genuinely useful for learning basic strategy without risking much — though it also enables faster losses for problem gamblers.

Much faster. Online RNG blackjack can run 200-400+ hands per hour. Physical casino tables run 60-80 hands per hour. Live-dealer online sits in between at about 60-80. The speed difference matters: same house edge × more hands = more dollars flowing through the math, both ways. Online lets you experience variance much faster — both the upswings and the downswings.

Yes — including the ones on this site. Look for: no signup required, no credit card requested, no "play with real money" pressure during play. Free tables on regulated casino sites ("demo mode") are also genuine — the same RNG, just without money. Some sketchy sites disguise registration funnels as "free play." Real free blackjack should feel like opening a webpage and clicking deal.

In the US: Nevada, New Jersey, and most tribal gaming jurisdictions plus a growing list of states with regulated online gambling (PA, MI, NY, CT, WV, etc.). Internationally: UK, Malta, Gibraltar, most of EU, Macau, Singapore, Australia (with varying rules and restrictions). The legal landscape changes frequently — check current local regulations before assuming. Online legal status often differs from physical casino status.

Key rules: don't touch the cards (dealt face-up) or the chips after betting starts. Hand signals only — verbal calls don't count for the cameras. Tip the dealer occasionally with chips. Don't criticize others' play (it doesn't affect your math anyway). Don't use a phone at the table. Tip cocktail waitresses $1-2 per drink. Most pit bosses are friendly; talk to them if you have questions. Casinos are professional; act professional.

Conventions vary but $5-25 across a typical session is standard for moderate players. Some players tip with chips after a winning hand ("placing one for the dealer" or "a hand for the dealer"), which the dealer plays alongside yours. Heavy tippers tip more; minimum-bet players tip less. Tipping isn't obligated but is appreciated, and dealers remember regulars. Tip more for friendly dealers, less for distant ones.

Las Vegas has the volume and variety; Atlantic City has the legal counter protection from Uston v. Resorts. For 3:2 games at moderate stakes, downtown Vegas and Reno often have better rules than the Strip's lower-min tables (which have largely converted to 6:5). Macau has the highest stakes globally. For pure rule quality at low stakes, regional US casinos sometimes have better rules than tourist destinations. The "best city" depends what you're optimizing for.

In regulated jurisdictions (Nevada Gaming Control Board, NJ Division of Gaming Enforcement, UKGC, etc.), it's essentially nonexistent — the regulatory and legal consequences are too severe. In offshore unregulated venues, there have been documented historical cases. The other risk is unintentional dealer error (mistakes happen at busy tables) which usually self-corrects but occasionally favors the house if unnoticed. Watch your cards and your chips.

Nobody invented blackjack as we know it — it evolved from older European card games over centuries. The earliest documented reference is in a 1601-1602 novella by Miguel de Cervantes mentioning "veintiuna" (twenty-one). Various 21-target games existed across France, Spain, and Italy through the 1700s. The American form solidified in casinos through the 19th and early 20th centuries. The "natural" 21 bonus paying extra is a 20th-century American addition.

MIT-trained mathematician who, in 1962, published Beat the Dealer — the book that proved blackjack could be beaten through card counting. The book sold 700,000+ copies and forced the casino industry to restructure itself (multi-deck shoes, surveillance, eventual 6:5 payouts). Thorp went on to apply the same probabilistic thinking to financial markets, founding Princeton/Newport Partners in 1969 — one of the first systematic hedge funds. Foundational figure in both gambling and quantitative finance.

Yes, but very different from the Bringing Down the House book and 21 movie. The team operated from 1979 through the early 2000s, structured as a hedge fund (Strategic Investments). Real founders: Bill Kaplan (Harvard) and JP Massar (MIT). Conservative earnings estimates from alumni: "several million dollars" over 20+ years, far less than the inflated "tens of millions" of the book. The team's actual demographics were substantially Asian-American, which the film whitewashed.

Mostly no. 21 (2008) heavily fictionalized the MIT Team story — invented romantic subplot, fabricated dramatic confrontations, whitewashed casting, compressed decade-plus history into a school year. Rain Man (1988) treats counting somewhat accurately for the era. The Hangover (2009) shows fast cinematic counting that doesn't resemble real play. Most blackjack in film is dramatic shorthand, not documentary. Watch for entertainment, not education.

No documented evidence. The claim circulates in low-quality gambling content, sometimes paired with a fabricated Einstein quote about being unable to beat the dealer "without cheating." Einstein died in 1955 — seven years before Thorp's 1962 proof that the game could be beaten. There's no record of Einstein discussing blackjack in his correspondence or biography. The Einstein blackjack anecdote is invented.

Standard guidance: 40-50 times the table minimum as a session bankroll. At a $25 table, that's $1,000-$1,250 for a comfortable session. This bankroll absorbs normal variance without busting out. Bringing less means you may go broke from variance even with perfect strategy. Bringing way more without discipline can lead to chasing losses. Many players also set a "win cap" — leaving the table after reaching a specific profit.

Bankroll management is the practice of having enough money allocated to gambling to survive normal variance without going broke. The math underneath: with a 0.5% house edge and 1.15-unit standard deviation per hand, swings of ±30 units over 100 hands are normal. If your bankroll is only 50 units, you have meaningful risk of busting from variance alone. Professional counters typically use 1,000+ unit bankrolls. Recreational players need at least 40-100 units per session.

Possible but rare and hard. Requires: counting skill mastered at casino-table speed, 6+ figure bankroll, willingness to relocate between casinos (since you'll get backed off), tolerance for major variance (months of losses are normal), and acceptance that the math has been getting harder as casino countermeasures evolve. Most who try fail; most who succeed earn modest middle-class incomes after expenses. The MIT Team peak era of organized counting is largely over.

No. Martingale (doubling after each loss) doesn't change the house edge — the math is brutal. With even a $10 starting bet, after 7 consecutive losses you need $1,280 for the next bet just to recover the $10. Most casinos have table maximums that cap this at exactly the point you'd need to break through. Progressive betting systems sound clever but provably cannot overcome a negative-expectation game. Skip them.

Don Johnson's $15 million run across three Atlantic City casinos in 2011 is the largest publicly-documented single stretch. He wasn't counting — he negotiated favorable rule sets (including 20% loss rebates) that pushed his expected value into positive territory. Within six months all three casinos cancelled his terms. The story is documented in Mark Bowden's April 2012 Atlantic article. Individual hand wins of $1M+ exist but are usually high-roller suite play.