The tightest published-RTP game in the casino — 0.44% house edge on the reference rules,
an order of magnitude better than any slot. The catch is that every rule change carries a
price and every strategy deviation leaks edge. Below: the rule-set ledger (what each
variation costs you), the compressed basic-strategy chart, the side-bet edges, and why
card counting doesn’t work online.
The reference payout. Any table not paying 3:2 is a downgrade — check the felt before you sit.
Blackjack pays 6:5
+1.39%
SKIP
Adds ~1.4% to the house edge on a single rule change. The single worst rule variant in mainstream blackjack.
Blackjack pays 1:1
+2.27%
SKIP
A destroyed table. Some live-lobby "Free Bet Blackjack" builds pay 1:1 — that is not blackjack, that is a keno bet in card form.
Dealer stands on all 17s (S17)
0.00%
BASELINE
The player-friendly baseline used in most published RTPs. Ask before you sit.
Dealer hits soft 17 (H17)
+0.22%
NEUTRAL
Standard on most Vegas Strip and online live tables. Small edge gift to the house.
Double after split (DAS) allowed
-0.14%
POS
Lets you press splits into 10-value hands. Small but always in your favour.
Late surrender allowed
-0.07%
POS
Late surrender against a 10 (hard 16) and against an ace (hard 15/16) saves half your bet in the worst spots.
Re-split aces allowed
-0.08%
POS
Turns split aces into a proper multi-hand situation instead of a one-card ceiling.
Single deck
-0.48%
POS
Massive edge reduction versus 6-deck. Rarely offered online — when it is, expect H17 and a 6:5 payout to claw it back.
Two decks
-0.19%
POS
Solid improvement over 6-deck. Most audited live tables ship 6 or 8 decks — 2-deck is a niche.
Six decks
0.00%
BASELINE
The reference deck count. Most audited live builds run six decks with a cut card.
Eight decks
+0.02%
NEUTRAL
Trivially worse than six decks. Almost every Evolution live table runs eight.
Continuous shuffling machine (CSM)
+0.06%
NEUTRAL
Reshuffles every round — kills any residual card-counting signal. Small edge gift to the house.
Insurance / even-money offered
0.00%
SKIP
The side bet itself carries a ~7% edge and does not change base-hand math. Always decline unless you are counting a specific shoe.
Baseline is the reference table used by every published blackjack RTP: six decks, dealer stands
on all 17s, blackjack pays 3:2, double any two cards, double after split allowed, no surrender.
That build runs a 0.44% house edge. Stack the player-positive rules (single deck, DAS, late
surrender, re-split aces) and the theoretical edge drops below 0.20%; stack the house-positive
ones (H17, 6:5, CSM, no DAS) and you can push it past 1.9%. Every rule delta is additive.
Chart is for the reference rule set — six decks, S17, 3:2, DAS, late surrender. On H17 tables
a handful of borderline hands shift (A,8 vs 6 doubles, A,7 vs 2 doubles). If the table doesn’t
offer surrender, every R cell falls back to hit. Memorise this chart before you sit and every hand
converges on 99.56% RTP over volume — the tightest sustainable expected value in the casino.
[ variant edge ledger · every named online blackjack build ]
Variant
House edge (base)
Note
Classic Blackjack (S17, 3:2, 6-deck)
0.44%
The reference build. Evolution's "Blackjack A/B/C" tables, most Playtech classic tables, and Pragmatic Play Live default here. What every RTP figure is quoted against.
Classic (H17, 3:2, 6-8 deck)
0.66%
The most common live-lobby default. Dealer hits soft 17 — a small but persistent edge gift to the house.
Blackjack Switch (Playtech)
0.17%
Play two hands, swap the top cards. Dealer 22 pushes standing hands. The swap mechanic makes this the lowest-edge branded blackjack build online.
Spanish 21 (Real Time Gaming, others)
0.40%
Played with a Spanish shoe (all tens removed, 48-card deck). Bonuses on 21 compensate for the missing tens.
Free Bet Blackjack (Evolution)
0.98%
Studio covers your doubles on 9/10/11 and non-ace splits. Sounds like a gift; dealer 22 pushes all standing hands and eats the value.
Infinite Blackjack (Evolution)
0.51%
Unlimited seats — one dealer hand vs an unlimited number of players. Base rules match S17 6-deck; side bets always offered.
Power Blackjack (Evolution)
0.66%
8-deck, all 9s and 10s removed. Player doubles up to quadruple. Dealer 22 pushes. The 9/10 removal drives the edge back to H17-classic levels.
Lightning Blackjack (Evolution)
1.14%
Random 2×-25× multipliers on winning hands after a chip fee. High variance, notably worse base edge.
Double Exposure
0.69%
Both dealer cards face up. Ties lose (except blackjack), blackjack pays 1:1. The dealer information cost is paid back at settlement.
Blackjack Party (Playtech / Evolution)
0.72%
H17 8-deck with side bets and a chatty host format. Base math sits at a typical live H17 table; the side bets do the damage.
Six-to-Five branded tables
1.83%
Any table paying 6:5 on blackjack — regardless of other rules. Anywhere from Vegas to Evolution promotional lobbies. Do not sit.
Edges quoted at basic strategy on the variant’s house rules. Live-lobby "Free Bet" and
"Lightning" builds trade a novelty mechanic (studio-covered doubles, hit multipliers) for a
structural push rule (dealer 22 tie) that eats the value. The three lowest-edge builds are all
3:2 tables with structural rule advantages — Switch (top-card swap), Spanish 21 (bonus payouts),
Classic S17 6-deck (the audited default). If you cannot see the specific rule set on the info
panel, assume the worst plausible variant.
[ side-bet ledger · the "extra options" ]
Side bet
Typical pays
House edge
Note
Insurance
2 : 1
7.4%
A €5 side bet the dealer has blackjack. At a fresh shoe, 4 of 13 ranks give her the ace-plus-ten. The bet needs 33.3% frequency to break even and pays at 30.8%. Always decline.
Perfect Pairs
5/12/25 : 1
4.1%
Pays on mixed pair (5:1), coloured pair (12:1), perfect pair same suit (25:1). Adjusted paytables at some studios shift this 3-8%.
21+3
5-100 : 1
3.2%
Your two cards plus dealer up-card make a poker hand (flush, straight, three of a kind, straight flush, suited trips). Paytable-dependent — the low-cap versions edge higher.
Lucky Ladies
4-1000 : 1
24.7%
Bet on a total of 20. Queen of hearts variants push top prize to 1000:1. The high-cap versions carry the worst edge on the felt.
Bet Behind (Evolution)
1 : 1
0.5%
You wager on another live player's hand. If they play basic strategy well, the edge is near the base table. If they don't — you inherit their leaks.
Free Bet double / split
push on 22
1.0%
Casino covers doubles on 9/10/11 and non-ace splits, but pays 1:1 on blackjack AND pushes dealer 22 against your standing hand. Reads generous — the 22 push does all the work.
Every branded blackjack side bet is a losing proposition at its posted paytable. Insurance and
Perfect Pairs are the cheapest of the lineup at 4-7%; Lucky Ladies at 24.7% is one of the worst
bets on any casino table. Bet Behind is the only side bet that can approach neutral EV — the
catch is that you inherit the strategy of the seated player, and most seated players are not
running the strategy chart.
Card counting online — the honest verdict
Card counting works by tracking the ratio of high cards (10s and aces) to low cards (2-6) still
to be dealt. A shoe rich in high cards favours the player — more blackjacks, more double-down
hits, more dealer busts. A counter raises their bets when the true count is positive and drops
them back when it swings negative. Executed cleanly against a permissive live game, that’s
good for +0.5% to +1.5% expected value depending on rules, penetration and betting spread.
Online blackjack closes both doors. The RNG tables reshuffle from a fresh 52 × n deck between
every hand, so no residual count exists to carry forward — mathematically unbeatable by counting.
Live-dealer tables use a physical shoe, but every audited studio (Evolution, Playtech, Pragmatic
Play Live) cuts at roughly 50% penetration — the cut card moves to shuffle the shoe after four
of eight decks are dealt. A counter needs 65-75% penetration and a 1:8 minimum bet spread to
overcome a typical +0.5% base edge. The studios move the cut card exactly to prevent that.
Every live table also cameras your bet-sizing, and dramatic bet spreads are flagged for review.
The one exception is a promotional deep-shoe live event with a stated 75%+ cut, and even those
are watched. Treat online blackjack as a −0.44% game played on basic strategy, and the FAQ below
explains why insurance never breaks that ceiling.
Advantage-play myths that don’t work
SKIP
Card counting the RNG shoe
The pitch: Count Hi-Lo, raise bets at +2 true count
The reality: RNG builds reshuffle every hand from a fresh 52 × n deck. The count you carry from the previous hand is discarded before the next deal — zero signal, zero edge. Every audited RNG blackjack table is unbeatable by counting.
SKIP
Card counting live-dealer shoes
The pitch: Count the 6/8-deck live shoe until reshuffle
The reality: Live shoes reshuffle at ~50% penetration (roughly 4/8 decks dealt) and every studio cameras your bet-sizing. A +5% counter needs 65-75% penetration to overcome the base edge — and the studios move the cut card exactly to close that door. Playable in theory, unplayable in practice.
SKIP
Progression betting (Martingale, Oscar's Grind)
The pitch: Chase losses with doubled bets, retreat after a win
The reality: Base hand EV doesn't change with bet size. All progression does is convert small losses into occasional catastrophic ones. Every operator flags aggressive progressions for bonus-clearance limits.
SKIP
Shuffle tracking
The pitch: Track slug positions across the shuffle
The reality: Applies to old-style shuffle procedures with visible clumps. Every audited live studio (Evolution, Playtech, Pragmatic) runs machine-shuffled shoes with random cut. No trackable slugs.
SKIP
Insurance as a hedge on your own blackjack
The pitch: Take even-money on your own natural — "sure win"
The reality: A natural at 3:2 returns +1.5 units. Even-money returns +1.0. You are surrendering a third of your winning ticket for the certainty of a smaller win. Long-run: −4% on every insured natural.
The playable version of "advantage" at online blackjack is compliance with the strategy chart on
a 3:2 S17 table with DAS. That converts to 99.56% RTP over volume, which is roughly 40× tighter
than an average slot. It is not a positive-expectation game; it is the least negative expectation
in the casino. See the bankroll management guide for
session sizing on negative-EV games.
Demo mode uses the same audited RNG as the real-money build — same rules, same house edge, no
cash at risk. Use demos to memorise the strategy chart before you sit at a real table. See the
demo-mode guide for how demo differs from real-money
play, and the basic strategy glossary for the term
breakdown.
FAQ · online blackjack
Which blackjack variant has the best odds?
Blackjack Switch at 0.17% is the lowest-edge branded build online (allowed thanks to the "swap top cards" mechanic). For non-gimmick blackjack, classic S17 6-deck with 3:2 payout and double-after-split at 0.44% is the reference. Avoid anything paying 6:5 (adds ~1.4%) or advertised as "Free Bet" (dealer 22 push eats the value).
What is the house edge on blackjack with basic strategy?
On the reference rules — six decks, dealer stands on all 17s, blackjack pays 3:2, double after split allowed, no surrender — the edge is 0.44%. That translates to an RTP of 99.56%, the highest of any mainstream casino game and roughly 200× tighter than an average slot. Deviate from the strategy chart on any hand and you leak edge one hand at a time.
Should I ever take insurance?
No — not even on your own blackjack. The insurance side bet pays 2:1 that the dealer holds a 10 under the ace. She holds one 30.8% of the time, needs 33.3% to break even, so insurance costs 7.4% per unit wagered. Taking even-money on your own natural converts a +1.5-unit result into +1.0 — you're surrendering 33% of the ticket for certainty.
Does card counting work at online blackjack?
RNG tables: no — the shoe is reshuffled every hand, so there is no residual count to carry between hands. Live-dealer tables: theoretically yes on a 6- or 8-deck shoe, but studios cut at ~50% penetration (roughly 4/8 decks dealt) and every studio actively flags bet-sizing. A +5% counter needs 65-75% penetration to overcome the base edge; the studios move the cut card exactly to close that door. In practice, live-dealer counting is not profitable.
Why does the strategy chart tell me to surrender hard 16 vs 10?
Hard 16 vs a 10 up-card wins only about 23% of hands under any decision — hit or stand both lose you ~54% of a unit on average. Surrender loses you half a unit. Half a unit beats 54% of a unit, so surrender is the least-bad play. It only exists on tables that offer late surrender; on tables without it, the chart falls back to hit.
What is dealer soft 17 (H17 vs S17)?
Soft 17 is any hand totalling 17 with an ace counted as 11 — most commonly ace-6, but also ace-2-4, ace-3-3, ace-2-2-2. On S17 tables, the dealer stands. On H17 tables, she draws — which is worse for the player because it gives her extra pathways to 18-21. H17 costs the player 0.22% relative to S17. Read the felt before sitting.
Do blackjack side bets ever have positive expected value?
No. Every published side bet — Perfect Pairs, 21+3, Lucky Ladies, Match the Dealer, Bust It — carries a 3-25% house edge. The house cannot afford to expose itself on high-limit longshot pays without a heavy edge. Bet Behind (wagering on another player's hand) is the only exception, and it inherits their strategy quality — you need a good player at the table to make it neutral.
Does blackjack count toward bonus wagering?
Barely. Most operators credit blackjack at 5-10% toward wagering because a competent player runs so close to breakeven that treating it as full-contribution wager would let players clear bonuses at near-zero expected loss. On a €4,000 wagering requirement, that's €40,000-€80,000 of table turnover — clear bonuses on slots and play blackjack with a clean balance. See the wagering-requirements guide for the math.