How to Win at Blackjack: What Every Decision Actually Costs

The hand every blackjack player agonises over is a hard 16 against the dealer’s 10. Hit it and you probably bust; stand and you probably lose anyway. It comes up in 1.82% of hands, more often than almost any other decision worth arguing about, and players will hold up a table debating it.

It is worth 0.06 of a point. Getting it wrong costs you about six hundredths of one percent of your bet. The same 16 against a dealer’s 7 — a card nobody worries about — costs 6.06 points if you play it wrong, roughly 100 times more.

That gap is what this page is about. Below is the full basic-strategy chart, the price of each common mistake in the order that actually matters, and what the three usual ways of playing are really worth. Every number was solved from the rules rather than copied, and each one was checked a second way before it went on the page.

What actually separates a good player from a bad one

Blackjack has a house edge you cannot remove. Played perfectly under common rules — the dealer stands on soft 17, you may double on any two cards, you may double after splitting, blackjack pays 3:2 — the house keeps 0.5096% of every bet in the long run. That is the floor. No system gets under it, and anyone selling you one is selling you something else.

What varies enormously is how far above that floor you play. Here are the three styles you see at every table, priced:

Two things stand out. The first is that giving up doubling and splitting is the single most expensive habit a careful player can have. A player who hits and stands correctly but never doubles and never splits is playing at 2.42% — nearly five times the house edge of correct play — while feeling like they are playing well. Every decision they make is right except the two that put more money on the table when the odds are good.

The second is that refusing to bust is worse than simply copying the dealer. Standing on every hard 12 or more feels safe, and it is the most expensive of the three at 6.08%. Playing by the dealer’s own rules — draw to 17, stand on 17 — is 5.67%. The dealer’s rulebook beats the instinct to protect your hand, because the hands you save from busting are mostly hands that were going to lose anyway.

So the order of priority is settled before you learn a single chart cell: double and split when you should, then worry about the close decisions.

The basic strategy chart

This is solved output, not a chart copied from somewhere. Every cell is the play with the highest expected value given your hand and the dealer’s upcard. It was checked cell by cell against the published multi-deck chart for these rules, and all 260 cells agree.

Read down the left for your hand and across the top for the dealer’s upcard. H = hit, S = stand, D = double if you can (otherwise hit), P = split.

Hard totals — no ace, or an ace that has to count as 1

Anything below 8 is always a hit, and 17 or more is always a stand. The interesting band is 12 to 16, where you stand against a dealer 2 to 6 and hit against 7 or higher. The reason is not your hand — it is that a dealer showing 2 to 6 is the one likely to bust, so you stop drawing and let them.

Soft totals — an ace still counting as 11

A soft hand cannot bust on the next card, which is why the chart is so much more aggressive here. Note A-7, the row that costs players the most money: an 18 that looks perfectly respectable. You stand on it against a 2, 7 or 8, double it against 3 through 6, and hit it against a 9, 10 or ace. Standing on soft 18 against a 9 is the fourth most expensive mistake in the game.

Pairs

Two rules here never change and are worth memorising on their own: always split aces and eights, never split tens or fives. Splitting aces turns one poor hand into two hands starting on the best card in the deck. Splitting eights breaks up 16, the worst total you can hold. Splitting tens breaks up 20, which is already close to the best hand at the table, and splitting fives breaks up a 10 that you should be doubling.

These charts assume the common rule set named above. If your table pays 6:5 on blackjack instead of 3:2, no chart will save you — that single rule costs more than twice the entire house edge of the 3:2 game, and the right response is to find another table.

Which mistakes actually cost you

Not all errors are equal, and the ones players argue about are mostly not the ones that matter. Each cost below is the expected value you give up, in points of your bet, every time you play that hand wrongly. Beside it is how often that exact hand appears, so you can see which are worth learning first.

The table rewards reading in both directions. A-A against a 6 tops it at 48.14 points — hitting a pair of aces instead of splitting them is the most expensive single decision on the list, because you are turning the strongest start in the game into a soft 12. It is also rare, at 0.05% of hands, so it costs you seldom but hurts every time.

At the other end, hard 16 against a 10 costs 0.06 points and is the hand everyone deliberates over. Hitting and standing are almost exactly as bad as each other, because the hand is already close to lost whichever way you go. Do whichever you like and spend the attention somewhere it earns something.

Where it earns something is the middle of the table: hard 11 against a 10 turns up in 1.46% of hands and doubling rather than hitting is worth 6.02 points every time, and hard 16 against a 7, where standing feels natural and costs 6.06 points. Those two are common and cheap to learn.

The advice that does not work

Card counting is the strategy most often recommended and the least useful online. In software blackjack the deck is shuffled before every hand, so there is nothing to count — the running count is reset before you can act on it. Live-dealer tables use real cards, but shoes are shuffled frequently and bet-sizing patterns are easy to spot, so it is not a plan for an online bankroll either.

Betting systems have the same problem in a different shape. Doubling your stake after a loss does not change the odds of the next hand; it changes how quickly you reach the table limit or the end of your money. The house edge applies to every bet regardless of what the previous bet did.

What does work is unglamorous: learn the chart, always double and split when it says so, pick a table paying 3:2, and choose stakes small enough that a normal losing run does not end your session. That is the whole edge available to a player, and it takes the house edge from 6.08% at its worst down to 0.5096%.

None of it makes blackjack a way to earn money. Correct play means losing more slowly, not winning. Set a budget before you sit down, treat it as the price of the entertainment, and stop when it is gone.