Card Counting in Blackjack: How It Actually Works

How to count cards, in four steps

1. 1. Assign a value to every card

   Give each card one of three values as it appears. Low cards — two through six — count plus one. Middle cards, seven through nine, count zero and are ignored. Tens, picture cards and aces count minus one. These are the Hi-Lo tags, and they are the whole vocabulary of the system.

2. 2. Keep a running count

   Start at zero on a fresh shoe and add each card's value as it is dealt, keeping one number in your head. Do not restart it between hands; it runs until the cards are shuffled.

3. 3. Convert it to a true count

   Divide the running count by the number of decks you estimate still remain to be dealt. This is the step that makes the number mean something: plus six spread through five decks is a faint tilt, while plus six in one remaining deck is a strong one. The running count alone cannot tell those apart.

4. 4. Bet with the edge

   Raise your bet when the true count is high and drop to your minimum when it is not. This is where the advantage is actually taken — counting produces nothing on its own, and the money is made by having more on the table when the remaining cards favour you than when they do not.

Why high cards favour the player

A deck rich in tens and aces is better for the player than for the dealer, and the reason is in the rules rather than in luck. The dealer has no choices: on a stiff hand they are required to draw, so when the small cards that would have saved them are gone, they bust more often. A player holding the same hand can simply stand.

Two more rules point the same way. A blackjack pays the player more than an equal hand pays the house, and blackjacks arrive more often when tens and aces are concentrated. And the player can double down or split when the situation is good, putting extra money out exactly when the deck is helping — options the dealer never has.

Every one of those advantages grows as the remaining cards tilt high, and every one shrinks as they tilt low. That is the whole mechanism a count is measuring: not which card is coming next, but whether the balance of what is left favours the side of the table that gets to make decisions.

What counting actually is

Card counting is not memorizing the deck and it is not a trick. It is one running tally that tracks whether the cards remaining in the shoe are rich or poor in tens and aces. When the remainder is rich in big cards, the player's chances improve — blackjacks pay more than they cost, doubles connect more often, the dealer busts more. Counting simply tells you when that is true, so you can bet more then and less otherwise.

That is the entire idea. Everything else — systems, conversions, deviations, camouflage — is engineering around that one fact.

The Hi-Lo values

Hi-Lo assigns every card one of three values: 2 through 6 count +1, 7 through 9 count 0, and tens, faces, and aces count −1. As each card leaves the shoe you add its value to a running count. Low cards leaving is good for you — they were the dealer's friends — so the count rises. Big cards leaving is bad for you, so it falls.

A positive running count means the shoe that remains is richer in big cards than normal. That is the signal.

From running count to decision

In a multi-deck game the raw running count overstates itself — a +6 with five decks left is thinner than a +6 with two decks left. So you convert it: divide the running count by the number of decks you estimate remain. The result, the true count, is the number that actually drives your bets. Our guide on the true count walks the whole conversion, including the deck estimation that makes it honest.

How you actually learn it

You learn to count the way you learn an instrument: a small skill drilled until it stops costing attention. Deal a deck one card at a time and keep the tally; a counted-out deck ends at exactly zero, so the deck grades you honestly every run. Then add a clock. Then add pairs. Then add noise. The free countdown drill on this site runs that exact ladder and times you, which is the part self-dealt cards can't do.

What the count is actually measuring

A count is not a prediction and it is not a memory trick. It is a running estimate of one thing: whether the cards still to come are richer or poorer in tens and aces than a fresh deck would be. That is the only question, and everything else in this subject is machinery for answering it while a game is moving.

Tens and aces help a player more than they help a dealer, for reasons that have nothing to do with luck. A player who holds a stiff hand can choose to stand; a dealer holding one is required by the rules to draw and is therefore more likely to bust when the small cards are gone. A player can double down on a good starting hand and the dealer cannot. Blackjacks pay more to a player than an equal hand pays to the house. Every one of those advantages gets larger as the remaining cards tilt high.

So the count is a bet-sizing instrument first and a playing instrument second. When the remaining cards favour you, you put out more money. When they do not, you put out the minimum and wait. The counting itself produces nothing on its own — the bet sizing does.

Why one number is enough

It is tempting to think a better system would track more: which specific cards are gone, how many of each rank remain, what the dealer is likely holding. Systems like that exist and they are measurably harder to run without error at the speed a real table moves.

The cost of complexity is not theoretical. A system you can run perfectly for six hours beats a more precise system you run with occasional mistakes, because a counting error does not merely cost you the difference between two estimates — it can invert your bet at exactly the moment the bet matters most. The edge available here is small enough that a handful of large mistakes across a session can erase it entirely.

This is why the plainest balanced system remains the standard for people who play for their own money rather than write about it. It is not the most powerful. It is the most powerful thing most people can actually execute, hour after hour, without the accuracy quietly decaying.

The order of learning, and why getting it backwards is expensive

Basic strategy comes first, and not as a formality. Counting adds a small edge on top of whatever you are already doing, so a counter playing imperfect basic strategy is adding a fraction of a percent on top of a much larger leak — going backwards while feeling sophisticated.

The sequence that works is: play the chart until it is automatic, then learn to keep a running count without losing the chart, then learn to convert that count for the cards remaining, then learn to size bets from it. Each step assumes the one before it has stopped requiring attention. Attempting them together is the most common way people spend months and arrive nowhere.

Questions

Do you have to be a genius to count cards?

No. Hi-Lo is adding one, subtracting one, and dividing small numbers. The hard part is doing it fast, quietly, for hours — that is training, not talent.

How long does it take to learn?

Counting a deck down cleanly takes most people days. Doing it at casino speed under distraction takes weeks of honest drilling. There is no shortcut that survives a real table.