Beating Online Blackjack: An Honest Look at Strategy, Counting and the House Edge

Beating online blackjack, in the sense of making money over the long run, isn't something the maths supports in a typical RNG game. What you can do is play at the lowest cost the rules allow. Perfect basic strategy cuts the house edge to its minimum: in our simulation of 10 million hands per rule set, that minimum was about 0.47% for a good 6-deck game and 2.13% for a poor one. Card counting, the one method with a real mathematical basis, needs conditions that games reshuffled every hand don't provide.

This page goes through what does and doesn't move the numbers, using our simulation data rather than promises.

What "beating" blackjack would actually mean

The house edge is the share of every bet the game keeps on average over a very large number of hands. To beat blackjack, you'd need to push that average below zero, so the game paid you more than you bet over time.

There are only three levers that touch the average:

1. Your decisions on each hand (hit, stand, double, split).
2. The rules of the table.
3. Information about the cards left to be dealt, which is what card counting uses.

Bet sizing patterns, streak tracking and "systems" aren't on that list, because they don't change the expected value of any individual hand.

Lever 1: basic strategy minimises the edge

Basic strategy is the set of decisions with the best expected return for every hand against every dealer upcard. It's the best you can do without information about the remaining cards. We simulated a player who follows it perfectly:

Every number in that table is positive. Perfect play gets you to the floor; it doesn't take you through it. Any mistake raises your edge above these figures. Our basic strategy trainer is the quickest way to stop making them, and the blackjack calculator settles any hand you're unsure of.

These are Monte Carlo estimates with a standard error of about 0.036 percentage points. Our rule-to-rule gaps are close to the Wizard of Odds rule-variation figures, within about two standard errors of the difference.

Lever 2: the rules matter more than any trick

Choosing the table is the biggest saving available. The gap between our best and worst rule sets is 2.13 − 0.47 = 1.66 percentage points.

The single biggest rule is the blackjack payout. Switching from 3:2 to 6:5 added 1.36 points in our simulation; the Wizard of Odds puts it at 1.39. Our 6 to 5 blackjack analysis breaks this down. After that come the dealer hitting soft 17 (+0.18 in our sim, 0.22 per the Wizard of Odds) and no doubling after splits (+0.06, 0.14 per the Wizard of Odds).

Lever 3: card counting, and why it fails online

Card counting is the only approach with a serious mathematical case for shifting the edge. Edward Thorp's Beat the Dealer (1962) showed that tracking the cards could overcome the house edge, and UC Irvine's library exhibit on his work describes it as a mathematical proof. The condition that matters is that cards dealt earlier stay out of play until the next shuffle.

Counting works like this. As cards leave the shoe, the mix of what's left changes. A shoe rich in 10s and aces favours the player; one rich in small cards favours the dealer. Counters keep a running tally, divide it by the decks remaining to get a "true count", and, as the Wizard of Odds explains, use that to size bets and adjust some plays.

Now apply that to an RNG game that reshuffles before every hand:

• Every hand is dealt from a complete, freshly shuffled deck.
• So the mix of cards is identical at the start of every hand.
• So there's no information to track, the count is always zero, and every hand has exactly the basic strategy edge.

A continuous shuffle, where used cards go back in immediately, has the same effect: cards never stay out long enough for the mix to drift. This isn't a statistic we're quoting; it follows directly from how counting works. If a game reshuffles every hand, counting is pointless however good you are at it.

Live dealer tables are different in principle because they deal from a physical shoe. Whether counting is worth anything there depends on how far into the shoe the dealer goes before reshuffling, which you often can't see in advance, and on whether the operator tolerates it. Our card counting resources cover the method, and our card counting practice guide explains how to drill it and where it's legal.

Why it can feel like you beat it

Blackjack's swings are large compared with its edge. We can estimate the typical swing per hand from our own simulation, because the standard error of an average shrinks with the square root of the number of hands:

Over 100 hands the expected loss is under half a bet, but the typical spread is more than 11 bets. Plenty of sessions finish ahead. As hands pile up, the expected loss grows in proportion to the number of hands while the spread grows only with its square root, so the edge takes over. A winning week is luck showing through, not proof of a method.

What doesn't work

• Betting systems. Martingale, D'Alembert and similar systems change bet sizes, not the edge on any hand; the Wizard of Odds' analysis of Martingale makes the same point. In our roulette test of 50,000 sessions per system, Martingale lost 2.77 USDT per 100 wagered against 2.71 for flat betting. In blackjack, doubles and splits make doubling systems harder still, as our page on the Martingale system in blackjack explains.
• Waiting for a "due" win. In a game reshuffled every hand, past results have no effect on the next deal. Believing otherwise is the gambler's fallacy, which Britannica describes as falsely assuming that independent plays depend on each other.
• Bots, signals and predictors. If the deck is shuffled fresh each hand, there's nothing to predict.
• Unchecked side bets. We haven't analysed side bets, so we quote no figures, but they're separate wagers with their own edge and don't improve the main game.

The best way to play blackjack online

If beating the game isn't on the table, the goal is to pay as little as possible for the entertainment:

1. Choose the rules first: 3:2 payout, dealer stands on soft 17, double after split. Our guide to online blackjack for real money has a full table checklist.
2. Play perfect basic strategy. Keep our strategy card open.
3. Bet flat. Same stake every hand, sized so a bad run doesn't hurt.
4. Slow down. Online games let you play very fast, and cost scales with total amount bet.
5. Set limits in advance: a loss limit and a time limit, decided before the first hand.

Bankroll in USDT

A stablecoin bankroll keeps the maths clear: your results reflect the cards, not the coin's price. Using the swing figure above, a 2 USDT flat bettor playing 1,000 hands should expect a loss of about 9.40 USDT at good rules, with a typical spread of around ±72 USDT (2 × 1.14 × √1,000). Size your bankroll for the spread, not the average. Splits and doubles can also put up to four stakes on a single round.

For a game where the edge is published as a single number and each round can be verified, compare our crypto dice guide.

Common mistakes

• Reading a winning session as evidence of a winning method.
• Trying to count cards in a game that reshuffles every hand.
• Spending effort on betting systems instead of on choosing better rules.
• Playing a 6:5 table because it's convenient.
• Increasing stakes after losses to "get back to even".

More guides are on the blackjack hub.