Every card counting system in use today descends from one idea: the cards already dealt change the odds on the cards still to come. That sounds obvious now. For most of blackjack's history it was not, and the person who proved it with a computer and then told everyone was a mathematician named Edward O. Thorp.
The story is worth knowing for a practical reason, not just a historical one. Understanding where the count came from explains what a count is actually measuring, and that makes the arithmetic much easier to hold onto than a memorized table of tag values.
Before the count: four men and a pile of desk calculators
The first serious mathematical work on blackjack was not about counting at all. It was about how to play a single hand correctly, which nobody had rigorously established.
In 1956, Roger Baldwin, Wilbert Cantey, Herbert Maisel and James McDermott published a paper titled "The Optimum Strategy in Blackjack" in the Journal of the American Statistical Association. Working on desk calculators, they computed the correct play for player totals against dealer upcards and produced what we would now recognize as basic strategy. They followed the paper with a book, "Playing Blackjack to Win", in 1957.
It was correct, careful, and largely unread outside academic circles. A statistics journal is not where a casino player looks for advice, and the book did not reach a mass audience. The mathematics of blackjack had been settled and almost nobody knew.
Their contribution still matters more than its readership suggests, because it established the baseline. You cannot ask how the count changes the right play until somebody has established what the right play is.
The IBM 704 and the insight that followed
Edward O. Thorp, a mathematician, picked up that work and asked a further question. Basic strategy assumes a full, freshly shuffled deck. But cards are dealt from a deck without replacement, so as a shoe is played, the composition of what remains changes continuously. What happens to the odds when it does?
Using an IBM 704 mainframe, Thorp extended the earlier calculations to show that removing cards from the deck shifts the odds in measurable ways, and, decisively, that a player who tracked those changes could hold an edge over the house at certain moments.
That last part was the break with everything before it. Basic strategy is a defensive achievement: it reduces the house edge to a small number and leaves it there. Thorp showed the number was not fixed. It moved as the deck was consumed, and it sometimes moved past zero.
The mechanism, in modern language, is that high cards favor the player and low cards favor the dealer. When a lot of low cards have already been dealt, the remaining deck is rich in tens and aces, which produces more player blackjacks, more successful doubles, and more dealer breaks on stiff hands. The full explanation of how counting works covers that logic in detail.
21 TrainerLearn the systems that grew out of this work: Hi-Lo, KO, Omega II and Zen, with drills for each. Free basic strategy trainer on the App Store.
Get the app
Beat the Dealer, 1962
Thorp published "Beat the Dealer" in 1962. It presented the Ten Count, the first winning system published for a general audience, and it made card counting widely known.
The Ten Count tracked the balance between ten-value cards and everything else remaining in the deck. When tens were plentiful relative to the rest, conditions favored the player, and the system told you to bet more and adjust certain playing decisions. The specific mechanics have long since been superseded, but the shape of the method has not: observe what has been dealt, infer what remains, size your bets to the conditions, and adjust play at the margins.
What made the book different from the 1956 paper was not the mathematics but the audience. It was written for people who actually played, and it presented a system a determined reader could learn and use. Card counting stopped being a result in a journal and became a thing people did.
Casinos responded with countermeasures over the years that followed, in the general directions you would expect: changes to how games are dealt and shuffled, and closer attention to players who vary their bets. That response is ongoing rather than a single event, and the modern shape of it includes devices like continuous shuffling machines, which keep deck composition near neutral and make counting ineffective.
From the Ten Count to the counts people use now
Thorp's system worked, but it asked a lot of the player. Tracking a ratio between two categories of card, and converting that ratio into a decision, is heavy arithmetic to run continuously at table speed while also playing hands and looking normal.
The systems that followed simplified it. Modern point counts assign each rank a small integer tag and ask you to maintain a single running total. Nothing about the underlying mathematics changed. The bookkeeping got dramatically lighter.
Hi-Lo is the canonical example and still the standard first system. Cards 2 through 6 count plus one, 7 through 9 count zero, and tens through aces count minus one. That is the entire tag set, and it is balanced, so a full deck counted down returns to zero. The true count conversion, dividing the running count by the decks remaining, adjusts for how much deck is left to be dealt.
Later systems traded simplicity for precision. KO removed the true count conversion by unbalancing the tags. Omega II and Zen use level two tags, distinguishing more finely between card values at the cost of harder arithmetic. All of them are variations on the same insight from the IBM 704 runs.
What the history actually teaches
Three things carry over from this history into anything you might practice today.
- Basic strategy came first, and still does. The 1956 work had to exist before Thorp's could. The same ordering applies to a learner: a count is a set of adjustments to correct play, and it is worthless layered over incorrect play.
- The count is a proxy, not a rule. The tags are a compressed estimate of how the remaining deck differs from a fresh one. Remembering that keeps you from treating a count as a promise about the next card.
- Simplification is a legitimate goal. The reason Hi-Lo displaced the Ten Count is that a system you can actually run under pressure beats a more precise system you cannot. That trade-off is worth remembering when choosing a system to learn.
The honest part
Counting done in your head is legal in the United States. It is observation and arithmetic applied to cards that are dealt face up in front of you, and no law prohibits paying attention. Our article on whether card counting is illegal goes through the specifics, including where jurisdictions differ and how electronic devices are treated very differently from mental counting.
Legal is not the same as welcome. Casinos are private businesses. They may refuse service, change how they deal to a player they are watching, or ban someone they suspect of counting, and they do. Nothing about the legality of the skill obliges anyone to let you use it.
And no technique removes risk. What Thorp demonstrated was a small edge under favorable conditions, not a way to win reliably. The swings around that edge are large, losing stretches are routine, and modern game conditions are less generous than the ones he studied. Counting is best understood as a demanding skill with an interesting history, which is exactly how we treat it in the app: practice it at home with virtual chips, where the only thing at stake is your accuracy.
Practice counting with virtual chipsSingle card and multi-card drills, true count conversion, bet spreads and index training. No real money anywhere in the app.
Download
Frequently asked questions
Who invented card counting?
Edward O. Thorp, a mathematician, developed and published the first card counting system that reached a general audience. His book Beat the Dealer, published in 1962, presented the Ten Count and showed that tracking which cards had already been dealt could give a player an edge. He built on earlier work by Baldwin, Cantey, Maisel and McDermott, who had computed the first rigorous basic strategy in 1956.
What was the first blackjack basic strategy?
Roger Baldwin, Wilbert Cantey, Herbert Maisel and James McDermott published The Optimum Strategy in Blackjack in the Journal of the American Statistical Association in 1956, computing it on desk calculators. They followed it with a book, Playing Blackjack to Win, in 1957. It was rigorous and correct, but it did not reach a mass audience the way Thorp's work later would.
What is the Ten Count in Beat the Dealer?
The Ten Count was the system Thorp presented in Beat the Dealer, tracking the ratio of ten-value cards to other cards remaining in the deck. When tens were plentiful relative to the rest, the player raised bets and adjusted play. It was the first winning system published for a general audience. Modern point counts later simplified the same idea into a running count with small integer tags.
Is card counting illegal?
Counting done in your head is legal in the United States. It is observation and arithmetic, not cheating, and no federal or state law prohibits thinking about cards you can see. Casinos are private businesses, though, and may refuse service, change how they deal, or ban players they suspect of counting. Electronic counting devices are treated very differently and are illegal in some jurisdictions.
Are modern card counting systems still based on Thorp's work?
Essentially every modern count is a descendant of Thorp's core insight that removing cards from the deck shifts the odds. Point counts such as Hi-Lo, KO, Omega II and Zen replaced his ratio approach with small integer tags added into a single running count, which is much easier to maintain at table speed. The underlying mathematics of deck composition is unchanged.