Blackjack

Q: What is the actual house edge in blackjack?

Between 0.4% and 2% , depending on rules and player skill. A typical six-deck, 3:2 blackjack, dealer stands on soft 17 (S17), double after split (DAS), surrender available — house edge is about 0.40% with perfect basic strategy. Remove surrender → 0.49%. Change to 6:5 payout instead of 3:2 → 1.45%. Add a continuous shuffler → 1.5%. A casual player who doesn't use basic strategy typically faces a 2-3% house edge. The 'blackjack is 0.5%' figure assumes both favorable rules and strict basic-strategy play.

Q: Why does the dealer have an edge if both reach 21?

The dealer's edge comes from order of play . The player acts first. If the player busts (exceeds 21), the player loses immediately — even if the dealer subsequently also busts. This single rule contributes roughly 8% raw house edge before all the player-favorable rules (3:2 naturals, doubling, splitting, surrender) claw most of it back. Without those rebates, blackjack would be unplayable for the house — it would lose every hand. The 0.5% house edge is what remains after the tug-of-war between 'player acts first' and 'player has flexibility.'

Q: Single deck vs six deck — which is better?

Single deck has a lower theoretical house edge (about 0.15% with same rules) because removing cards has more impact on remaining distribution. However, modern single-deck tables almost universally pay 6:5 on naturals (raising house edge to ~1.4%), and most have shallow penetration (dealer reshuffles after 50% of deck dealt, killing card-counting value). The result: today's 3:2 six-deck game is usually mathematically better than today's 6:5 single-deck. Verify both rules — payout AND penetration — before choosing.

/ˈblækdʒæk/ · Twenty-One · Vingt-et-Un

Blackjack table with cards and dealer — the canonical game of 21 — *Image: Pixabay Content License.*

I. Origin — From Vingt-et-Un to the Vegas Strip

The earliest documented predecessor of blackjack is Vingt-et-Un (French for "twenty-one"), played in Parisian gambling houses around 1700. The game arrived in North America with French colonists, settling in New Orleans by the early 1800s. When American casinos began offering bonus payouts to entice players — particularly a 10:1 bonus for a hand of the Ace of Spades plus a black Jack (clubs or spades) — the moniker "blackjack" stuck, even though that specific bonus disappeared within a decade.

The 1956 Baldwin-Cantey-Maisel-McDermott paper in the Journal of the American Statistical Association — "The Optimum Strategy in Blackjack" — produced the first computer-derived basic strategy. Six years later, MIT professor Edward O. Thorp published Beat the Dealer, demonstrating that card counting could tilt the long-run edge to the player. Las Vegas casinos responded within months — first by changing rules, then by training pit personnel to identify counters.

II. The Rules in Plain English

Each player and the dealer receive two cards. Player cards are face-up; the dealer's first is face-up, second face-down.
Card values: 2-10 face value; J/Q/K = 10; Ace = 1 or 11 (player choice).
"Blackjack" / "natural" = Ace + 10-value card on the first two cards — pays 3:2 (or 6:5 on degraded games).
Player options: Hit (take another card), Stand (no more cards), Double (one more card and double the bet), Split (separate a pair into two hands), Surrender (forfeit half the bet to fold, where offered).
Dealer plays last according to fixed rule: usually "hit on 16 or less, stand on 17+." Variants include H17 (hit soft 17) which adds ~0.22% house edge.
If player exceeds 21 ("busts"), player loses immediately. If dealer busts, all remaining players win.
Otherwise, closest to 21 wins. Push (tie) returns the bet.

III. The Math of Basic Strategy

Basic strategy is the mathematically optimal action for every possible combination of player two-card hand vs dealer up-card, assuming infinite-deck shuffle. It is a single 17-row by 10-column matrix.

Key non-obvious rules from the chart:

Always split Aces and 8s. Aces because two 11s become two strong starting hands; 8s because 16 vs dealer 10 is the worst hand in the game.
Never split 10s. Two 10s is 20 — already a near-certain winner.
Always double 11 vs dealer 2-10. Most powerful doubling spot in the game.
Stand on 12 vs dealer 4-6. Counterintuitive: you'd rather hope dealer busts than risk busting yourself.
Hit soft 17 (A-6). Soft hands cannot bust on a single card; never stand passively on soft 17.

The result of strict basic strategy under standard rules (6 decks, S17, DAS, surrender): house edge ≈ 0.4-0.5%. Under H17 + no surrender + 6:5 naturals: house edge ≈ 1.85%. The same player, same skill — but a fourfold change in long-run cost.

IV. Card Counting — The Math Behind the Legend

Card counting strategy — Hi-Lo system applied to blackjack — *Image: Pixabay Content License.*

Card counting works because the cards already dealt change the probability distribution of what remains. A deck heavy in tens and aces favors the player (more naturals, better double-downs, dealer busts more often on stiff hands). A deck heavy in low cards (2-6) favors the house.

The Hi-Lo count is the most-taught system:

2, 3, 4, 5, 6 = +1 (count up)
7, 8, 9 = 0 (neutral)
10, J, Q, K, A = −1 (count down)

Running count is divided by the number of decks remaining to give "True Count." Each True Count of +1 shifts the house edge by about 0.5% — so at True Count +4 the player has a ~1.5% edge.

Bet sizing: minimum bet at neutral counts; raise bets proportionally to advantage. A typical 1-8 bet spread converts a 1.5% advantage into about $25/hour profit on a $25 table — modest, but real, and the only mathematically certain positive-expectation game on a casino floor.

V. The MIT Team and the Modern Era

From the late 1970s through the 1990s, multiple MIT student teams — led at various times by Bill Kaplan, J.P. Massar, and Jeff Ma — pooled bankrolls (often $1-5M), trained members in Hi-Lo and Wonging (back-counting only during favorable counts), and toured high-limit rooms in Las Vegas, Atlantic City, Foxwoods, and Macau. Their cumulative profits ran into the tens of millions. The teams were dramatized in the 2008 film 21 and the 2002 book Bringing Down the House.

By the 2000s, casinos had countered with:

Continuous Shuffling Machines (CSMs): each hand returns to the shoe, eliminating count tracking
6:5 payouts on naturals: erodes the ~2.3% natural-frequency advantage
Trans Lux databases: shared blacklists of identified counters across operators
Facial-recognition surveillance: instant identification on entry

Modern profitable card counters typically operate in teams, in jurisdictions with weaker information-sharing (regional casinos, Caribbean cruises, smaller Asian markets), and accept frequent table back-offs as a cost of business.

VI. Rule Variants and Their Cost

Rule Change	House-Edge Impact
Naturals pay 6:5 (vs 3:2)	+1.39%
Dealer hits soft 17 (vs stands)	+0.22%
No double after split	+0.14%
No surrender	+0.07%
European no-hole-card rule	+0.11%
Single deck (vs 6 decks, all else equal)	−0.48%
Continuous shuffler (vs hand shuffle)	+0.014% (mild — but kills counting)

VII. Side Bets — Where the Real Edge Hides

Most modern blackjack tables include optional side bets (Perfect Pairs, 21+3, Lucky Lucky, Royal Match, Insurance). All carry house edges 4-12 times higher than the main bet:

Insurance: ~7.5% house edge (skip unless counting)
Perfect Pairs: 4-11% depending on payout structure
21+3: 3.5-7.2% (lowest when based on poker hand of player two cards + dealer up-card)
Lucky Lucky: 2.7-8% depending on shoe size and pay table

Side bets are profit centers: a $5 side bet alongside a $25 main bet doubles the casino's hourly win on that seat. The mathematically optimal player ignores all side bets except as part of a verified counting strategy.

VIII. Online and Live-Dealer Blackjack

Playing cards face-up — ace king queen of the deck — *Image: Pixabay Content License.*

Online blackjack splits into two universes:

RNG blackjack — each hand uses a fresh shuffle, mathematically equivalent to playing against an infinite deck with no opportunity to count. House edge is exactly the basic-strategy table value. Faster play (~250 hands/hour vs 70 live), higher hourly expected loss.
Live-dealer blackjack — streamed from studios in Riga, Manila, San José (Costa Rica), or Bucharest. Same rules as physical tables, often with deeper penetration (75-80% of shoe before reshuffle). Card counting is theoretically possible but team play coordination is harder; chat-only communication and 12-second decision timer constrain the advantage.

IX. Common Misconceptions

❌ "The goal is to get as close to 21 as possible." The goal is to beat the dealer. Standing on 12 vs dealer 6 is correct — even though 12 is "far" from 21.
❌ "Always take insurance if you have blackjack." Even-money insurance pays the same as 3:2 only if dealer hits blackjack 1/3 of the time. Actual probability is 4/13 ≈ 30.8%. Insurance loses on average.
❌ "A bad player at the table cost me the hand." Statistically false. Over 1,000 hands the variance is symmetric — bad play hurts the dealer equally often.
❌ "I'm due for a winning streak." Each hand is independent. The gambler's fallacy is the most expensive mental error in blackjack.

X. FAQ · Sources · Responsibility

What is the actual house edge in blackjack?

Between 0.4% and 2%, depending on rules and player skill. A typical six-deck, 3:2 blackjack, dealer stands on soft 17 (S17), double after split (DAS), surrender available — house edge is about 0.40% with perfect basic strategy. Remove surrender → 0.49%. Change to 6:5 payout instead of 3:2 → 1.45%. Add a continuous shuffler → 1.5%. A casual player who doesn't use basic strategy typically faces a 2-3% house edge. The 'blackjack is 0.5%' figure assumes both favorable rules and strict basic-strategy play.

Why is 6:5 blackjack so much worse than 3:2?

On a 3:2 game, a natural blackjack (21 on first two cards) pays $15 on a $10 bet. On a 6:5 game, the same hand pays only $12. This shaves $3 per natural, and naturals occur once every ~21 hands. Over 100 hands of $25 bets, you lose about $35 to the payout change alone — equivalent to a 1.4% rise in house edge. The 6:5 trick is the single most profitable rule change casinos have made in the last two decades. Las Vegas Strip 6:5 blackjack market share rose from 5% in 2004 to over 60% by 2022 (UNLV Center for Gaming Research).

Does card counting actually work in 2026?

Yes, mathematically — but the practical hurdles have grown. Card counting flips the house edge by ~1-2% in favor of skilled players when the deck is rich in tens and aces. The Hi-Lo system (introduced by Harvey Dubner, 1963; refined by Stanford Wong) is still the most-taught. However, modern casinos counter with: continuous shuffling machines (which destroy count value), 6:5 payouts (which eliminate the natural advantage), deep table monitoring (Bet Spread > 4:1 triggers attention), and back-off authority (casinos can refuse service in most jurisdictions). Profitable counting in 2026 typically requires a team (Spotter + Big Player) and tolerance for being barred.

What's 'basic strategy' and where do I get the chart?

Basic strategy is the mathematically optimal play (hit / stand / double / split / surrender) for every possible combination of player hand vs dealer upcard, derived by solving the game tree assuming an infinite-deck shuffle. The free, definitive chart is at Wizard of Odds (wizardofodds.com) — Michael Shackleford updates it for each rule variant. Memorizing one chart (six-deck, S17, DAS) takes 4-6 hours and reduces your house edge from ~2% (casual play) to ~0.4-0.5%. A laminated pocket card is permitted at most US casinos as long as it doesn't slow play.

Why does the dealer have an edge if both reach 21?

The dealer's edge comes from order of play. The player acts first. If the player busts (exceeds 21), the player loses immediately — even if the dealer subsequently also busts. This single rule contributes roughly 8% raw house edge before all the player-favorable rules (3:2 naturals, doubling, splitting, surrender) claw most of it back. Without those rebates, blackjack would be unplayable for the house — it would lose every hand. The 0.5% house edge is what remains after the tug-of-war between 'player acts first' and 'player has flexibility.'

Single deck vs six deck — which is better?

Single deck has a lower theoretical house edge (about 0.15% with same rules) because removing cards has more impact on remaining distribution. However, modern single-deck tables almost universally pay 6:5 on naturals (raising house edge to ~1.4%), and most have shallow penetration (dealer reshuffles after 50% of deck dealt, killing card-counting value). The result: today's 3:2 six-deck game is usually mathematically better than today's 6:5 single-deck. Verify both rules — payout AND penetration — before choosing.

Sources

Baldwin, Cantey, Maisel & McDermott (1956), The Optimum Strategy in Blackjack, JASA, vol. 51
Edward O. Thorp (1962, rev. 1966), Beat the Dealer: A Winning Strategy for the Game of Twenty-One
Stanford Wong (1994), Professional Blackjack, Pi Yee Press
Michael Shackleford ("Wizard of Odds"), House Edge by Rule Variant, wizardofodds.com
UNLV Center for Gaming Research, Las Vegas Strip Blackjack Payout Trends 2004-2024

Responsible play: This article is mathematical analysis, not gambling advice. The house edge in blackjack is the lowest in the casino, but the variance is real — a single session can easily deviate hundreds of percent from expected value. If gambling stops being entertainment, contact the National Council on Problem Gambling: 1-800-522-4700.