180Gambler > The Games

Card games and shuffling

Field: Game mechanics. This page was last modified on 17 August 2026.

Cards are the one common gambling mechanism with a memory. A die does not know what it rolled last time, but a deck dealt without replacement changes composition with every card that leaves it. That single difference produces conditional probability, makes strategy meaningful, and turns the quality of the shuffle into a genuine question rather than a formality.

Dependence between draws

The probability that the first card dealt from a full deck is an ace is 4/52. If it was an ace, the probability that the next card is also an ace is 3/51, because both the count of aces and the size of the deck have changed. If it was not, the probability is 4/51. This is conditional probability, and it is the reason the same question has different answers at different points in a deal.

The consequence for games is large. In a dice game, no information accumulates. In a dealt game, every exposed card is information about what remains, and a player who tracks it holds a slightly different picture of the deck from one who does not. Whether that picture is worth anything depends on the rules: it is worthless if the deck is reshuffled after every hand, and it is not worthless if a substantial part of the deck is dealt before shuffling.

Composition and strategy

In blackjack, the composition of the remaining deck changes the value of the decisions available, because the dealer must follow a fixed rule while the player may choose. That asymmetry is the source of the game's unusually low edge under correct play, and it is why published blackjack figures are always conditional on a strategy and a rule set: the number of decks, whether the dealer draws on a soft seventeen, whether doubling after splitting is allowed, and the payout on a natural all move it.

The mathematical study of that idea entered public circulation in 1962 with the publication of Beat the Dealer, which set out a counting method derived from computed deck compositions. The industry's response, more decks, earlier shuffles, continuous shuffling machines and cutting a portion of the deck out of play, is itself an application of the same mathematics in the opposite direction: each measure reduces the information a player can accumulate before the deck is reset.

What a shuffle has to achieve

A deck of fifty-two cards has 52! possible orderings, a number of about sixty-eight digits. A shuffle is adequate when the ordering it produces is, for practical purposes, no more likely to be one arrangement than another. A single riffle is very far from that: it preserves long ascending runs from the previous order, and those runs are detectable and exploitable.

The question of how many riffles are enough was settled quantitatively in 1992 by an analysis of the riffle shuffle as a mathematical process. The result, widely cited since, is that the deck's distance from random falls slowly for the first few shuffles and then drops sharply, with about seven riffles bringing a fifty-two-card deck close to uniform. The practical form of that finding is the rule of thumb used at tables everywhere: several riffles with a strip and a cut, not one.

Cuts, strips and machines

A cut alone does not randomise anything. It rotates the deck, which changes where the sequence begins but not the sequence itself, so a cut after a single riffle leaves the runs intact. Its purpose is procedural: it removes the dealer's knowledge of the top card and makes a stacked deck harder to deliver.

Automatic shuffling machines address the same problem with mechanism instead of technique, and they are tested for the same property: that the distribution of orderings they produce is close enough to uniform to be statistically indistinguishable from it. That testing is the card-game equivalent of the random number generator testing described in the regulation article.

Where the edge lives

Card games take their margin in several different ways. Some pay slightly short odds on a winning hand, as in the standard mechanism. Some take a commission on a particular winning bet. Some, like poker played between players, take a fixed rake from the pot and leave the players to compete with each other. The last of these is structurally different from every other game on this site: the operator is not a party to the wager at all, and the mathematics of the contest belongs to the players.