How to Win at Blackjack

Updated August 10, 2026

There is no way to win at blackjack with certainty, because it is a gambling game. Two things genuinely move the needle. Playing perfect basic strategy cuts the house edge to about 0.44%, and counting cards can add 0.5% to 1.5% on your side. Even then you will win about 42% of hands, lose 49%, and push the remaining 9%.

Test yourself in the basic strategy trainerA hand comes up, you make the call, and you find out immediately whether the chart agrees. Weighted toward the hands people actually get wrong.

Open the trainer →

Is there a guaranteed trick?

No, and the ones people sell you are betting systems. Martingale and its relatives do not change any probability in the game. They rearrange when you lose, concentrating many small wins and one enormous loss, and the table maximum guarantees the enormous loss eventually arrives.

The only two levers that touch the actual math are how you play each hand and how much you bet when the shoe is in your favor. Everything else is decoration.

1. Play basic strategy, exactly

Learn the chart until you never have to think. It takes the house edge to roughly 0.44% on a good six-deck game, and that is the entire ceiling of what is available without counting.

Exactly matters. Basic strategy is not a set of guidelines that you round off when a hand feels different. Every cell was computed, and the cells that feel wrong are usually the ones carrying the most value.

2. Count cards

Counting tells you when the remaining shoe is rich in tens, which is when you bet more. The realistic gain is 0.5% to 1.5% depending on the rules, the penetration, and how wide you can spread your bets.

That is the only method that turns a negative game positive, and it takes real work. Start at how to count cards, or try it in the live counter first.

What are the odds, really?

• You win about 42% of hands.
• You lose about 49%.
• About 9% push.

Losing more hands than you win and still only giving up half a percent looks contradictory until you notice which hands pay extra. A blackjack returns 1.5 times your bet, and doubles and splits put more money out precisely when you are ahead. Those close the gap almost entirely.

They also explain why 6:5 hurts so much and why refusing to double is so expensive. Both attack the hands that were supposed to make up the difference.

Which mistakes cost the most?

"Play the chart" is easy to say. The interesting question is what each specific bad habit costs, so we computed it: a six-deck game, dealer stands on soft 17, double after split, late surrender, blackjack pays 3:2, with a player who does everything right except the one thing named.

Computed by exact expected-value recursion over the shoe. Two standard simplifications: dealer probabilities come from the shoe with the upcard removed, and splits are treated as two independent hands without resplitting. Both are worth a few hundredths of a percent. Method in how we calculate.

That table gives you the order to fix things in. Fear of busting is the most expensive habit in blackjack. Standing on every hard 12 through 16 costs you more than twelve times what unnecessary insurance costs, and roughly eight times the entire house edge of a correctly played game.

Put the extremes side by side. A player who draws to 17 and never doubles or splits faces about 5.7%. A player who follows the chart faces 0.44%. That is a factor of thirteen, and the price of admission is a few hours of study. For scale, even well-executed card counting only adds 0.5% to 1.5%.

What does not work

• Betting systems. Martingale, Paroli, D'Alembert. They move losses around and eventually meet the table maximum. No sequence of bet sizes changes the edge on the next hand.
• Hunting for a "hot" table. Cards have no memory in the sense people mean. They have memory only in the sense of what remains in the shoe, and the only way to read that is to count.
• Copying the good player at third base. Their decisions do not change your probabilities, however it feels when they take the dealer's bust card.
• Insurance "just in case." Without a count it loses on every single offer. See insurance.
• Blaming the player who hit a 15. Over many hands, a bad player at your table costs you nothing at all.

How long until the math shows up?

This is the part nobody enjoys. A 1% edge means an expected profit of $10 on $1,000 wagered, and the swing on any single evening is many times that.

Losing sessions in a row is normal with a real edge. Thousands of hands and a bankroll that can absorb the swings are what separate a mathematical advantage from money in your pocket, and most people underestimate both by an order of magnitude.

Penetration decides whether counting is worth attempting at all. If the cut card sits at the halfway mark, the deep part of the shoe, where counting earns, never gets dealt. Details in how penetration changes your edge.

Keep reading

By Martin Kosek · every number on this site is computed by us, not copied.