Has a Shuffled Deck Ever Been Repeated?

Almost certainly not. No two fairly shuffled 52-card decks are known to have ever matched, anywhere, in the roughly one thousand years humans have been shuffling cards. ShuffleArchive exists to test that claim with data instead of repeating it as folklore: every shuffle we record is compared against every shuffle before it, and across the audited era 1 archive of 12,300,339 recorded shuffles, the count of matches is zero.

This post explains why that is the expected result, what it would take to see a repeat, and what makes our archive the only dataset that can answer the question empirically.

Why a repeat is so hard to find

A standard deck holds 52 cards, so the number of possible orderings is 52 factorial: 52 × 51 × 50 and so on down to 1. Written the usual way that is 8.07 × 10⁶⁷, which slides off the brain. Written out in full, it does not.

Orderings of 52 cards, written out

80,658,175,170,943,878,571,660,636,856,403,766,975,289,505,440,883,277,824,000,000,000,000

Shuffles in the audited era 1 archive

12,300,339

Numbers this large defeat intuition, so the honest way to compare them is on a logarithmic axis, where every step to the right multiplies everything before it by ten.

Even the most generous estimate of every shuffle ever performed by every human in history, every casino, every kitchen table, every bridge club, tops out around 10²⁰. That is 47 orders of magnitude short of the number of possible decks. The 52 factorial page walks through these magnitudes in more detail.

The birthday problem, applied honestly

"Rare" is doing a lot of work in most tellings of this story, so here are the actual numbers. The right tool is the birthday problem: the chance of a repeat among N random shuffles is governed by the number of pairs, which is N × (N minus 1) / 2, divided by 52!.

For the audited era 1 archive, N is 12,300,339. That gives 75,649,163,607,291 distinct pairs of shuffles, about 7.6 × 10¹³. Checking all of them against 8.07 × 10⁶⁷ possible orderings gives an expected number of matching pairs of about 9.4 × 10⁻⁵⁵. Zero matches is not a surprise. Finding one would have been the single most surprising empirical result in the history of probability.

Scale up to all of humanity and the answer barely moves. Take the 10²⁰ upper bound for every shuffle ever dealt by human hands. The expected number of matching pairs among all of them is about 6.2 × 10⁻²⁹. In plain terms: it is overwhelmingly likely that no two fair riffle shuffles in human history have ever produced the same deck.

For a coin-flip chance of seeing one repeat, you need roughly 1.06 × 10³⁴ shuffles. Every ant on Earth shuffling once per second since the Big Bang gets you to about 8.7 × 10³³. Almost, but not quite.

One caveat, because precision matters here: all of this assumes fair shuffles. Brand-new decks leave the factory in the same order, and a deck riffled only once or twice retains most of that structure, so identical near-sorted decks certainly have occurred. The claim is about well-shuffled decks. Our shuffles run 7 to 13 riffle passes of the Gilbert-Shannon-Reeds model, past the seven-shuffle threshold where a deck becomes effectively random.

Nobody had actually checked, so we did

Here is the strange part of this famous fact: until this project, nobody had a dataset to test it against. The claim was always derived, never measured, because measuring it requires recording millions of shuffles and comparing every one against every other.

That is what ShuffleArchive does. From April to September 2025, our original engine recorded 12,300,339 simulated riffle shuffles. In 2026 we re-encoded that archive and ran the first all-pairs duplicate audit in the project's history: every one of those 75.6 trillion pairs, checked exactly. The result was zero duplicates, zero parse errors, and zero hash collisions. The archive has kept growing since, past 12.6 million, and every new shuffle, whether a visitor clicks the button or our engine runs a batch, is checked against the full archive before it gets a number. The engineering behind that check is its own story.

So the empirical answer, as far as it can be given by the only project we know of that records and cross-checks shuffles at this scale: more than 12.3 million audited shuffles, more than 75 trillion pairs compared, zero repeats.

What happens on the day it happens

The archive is built for the event it will almost certainly never see. If a new shuffle's 52-card order ever matches an archived one exactly, byte for byte and not merely by hash, the system records the match permanently, and an alert fires from inside the database itself so the evidence cannot be lost even if every other layer fails. The methodology page documents the full path, because a claim like "we would know" deserves receipts.

Until then, the homepage headline stays at zero matches, and it is not a hardcoded string. It is a live count of a database table with no rows in it.

Test it yourself

Run a shuffle. Your deck gets checked against every shuffle in the archive before it is admitted. Even against an archive ten times the audited era 1 size, the odds of matching any recorded shuffle stay below one in 10⁵⁹. You will receive a permanent archive number and a deck order that, almost certainly, has never existed before and will never exist again.