Pick up a deck of cards. Any deck. Shuffle it. Not a fancy riffle shuffle, just the messy overhand thing you do when nobody's watching. Done? Good. You are now almost certainly holding an arrangement of 52 cards that has never existed before in the history of the universe, and will never exist again.
That sounds like the kind of thing someone says to sound clever at a party. It isn't. It's a provable mathematical fact, and the numbers behind it are so grotesque they stop making sense about three sentences in.
The number
A standard deck has 52 cards. The number of ways to arrange 52 distinct objects is 52 factorial, written 52!. That means 52 x 51 x 50 x 49, all the way down to 1.
The result is approximately 8.07 x 10^67.
Written out in full, it's 80,658,175,170,943,878,571,660,636,856,403,766,975,289,505,440,883,277,824,000,000,000,000. That number is so large that it doesn't fit into any physical analogy cleanly, but let's try a few anyway.
The observable universe contains roughly 10^80 atoms. So 52! is smaller than the number of atoms, which at first sounds reassuring. But consider what it means in practice. There are about 10^19 grains of sand on Earth. If every grain of sand were itself an entire planet covered in sand, and every grain on all those sand-planets were itself a planet covered in sand, you'd still be nowhere close to 52!.
Time doesn't help either. The universe is about 13.8 billion years old. In seconds, that's roughly 4.3 x 10^17. If you'd been shuffling a deck once per second since the Big Bang, you'd have completed about 4.3 x 10^17 shuffles. That's not even a rounding error against 10^67. You could run a billion universes in parallel, each shuffling a billion decks per second for the entire age of the cosmos, and the total number of shuffles performed would be about 10^44. Still twenty-three orders of magnitude short.
Why repeats don't happen
This is the part people push back on. Surely, they say, with billions of people playing cards across centuries, someone must have dealt the same hand twice. And the answer is: almost certainly not. Not once. Not ever.
The number of shuffles that have ever been performed by all humans in all of history is, generously, somewhere in the low trillions. Call it 10^13 if you're being extremely generous. The birthday paradox tells us that collisions become likely when you've sampled roughly the square root of the total space. The square root of 8 x 10^67 is about 9 x 10^33. So you'd need roughly 10^33 shuffles before you'd even have a coin-flip chance of a single repeat. Humanity has produced maybe 10^13.
We are not close. We are so far from close that "close" is meaningless.
So what if it happened?
Suppose it did. Suppose you shuffled a deck and, by some cosmic accident, produced an ordering identical to one that came up in a poker game in 1847, or in a child's game of Snap in Osaka in 2003. What would it mean?
Nothing and everything.
Nothing, because you'd never know. There's no registry of previous shuffles. No one wrote down the order of that deck in 1847. The event would happen in perfect silence, witnessed by nobody, recorded nowhere. Two identical arrangements separated by decades or centuries, and neither one aware of the other.
But if you could somehow verify it, the implications would be genuinely strange. You'd have witnessed an event with odds so extreme that it makes winning the lottery look like a certainty. The probability of matching a specific previous shuffle is 1 in 8 x 10^67. The odds of winning the UK National Lottery are about 1 in 45 million. You could win the lottery every single week for a thousand years and that streak would still be trillions of times more likely than one repeated shuffle.
At that point you'd be justified in questioning whether pure chance was actually responsible. Not because magic exists, but because when something that improbable happens, the rational response is to suspect your model is wrong. Maybe the shuffle wasn't truly random. Maybe the cards were in a similar starting configuration both times and the shuffling was insufficiently thorough. (This is, in fact, the most likely explanation for any apparent repeat. Most people don't shuffle well enough to fully randomise a deck. Mathematicians have shown you need about seven good riffle shuffles to approach a uniform distribution. Most people do three or four sloppy ones.)
The birthday paradox won't save you
People who've heard of the birthday paradox sometimes think it changes the story. In a room of just 23 people, there's a 50% chance two share a birthday. Doesn't that mean repeats become surprisingly likely with decks too?
It does make repeats more likely than you'd naively expect. But "more likely" is doing a lot of heavy lifting here. The birthday paradox works because there are only 365 possible birthdays. The square root of 365 is about 19, which is why 23 people is enough to make collisions probable.
With cards, you'd need the square root of 8 x 10^67 people all shuffling simultaneously. That's about 10^33 shuffles. Written out: a thousand billion billion billion. Even if every atom on Earth could shuffle a deck, you'd have roughly 10^50 shufflers, and you'd need them running for billions of years to approach collision territory.
The birthday paradox is real. It just doesn't rescue you when the number of "birthdays" is 10^67.
What the number actually feels like
I think the reason 52! is hard to process isn't that it's big. Lots of numbers are big. It's that it emerges from something so small and ordinary. Fifty-two cards. You can hold them in one hand. You can buy them for two quid at a corner shop. Children play with them.
And yet the combinatorial space they generate is larger than anything in the physical universe that humans regularly interact with. Not larger than the number of atoms. Smaller than that, actually. But larger than the number of stars (about 10^24), larger than the number of grains of sand on every beach on Earth, larger than the number of seconds that have elapsed since the beginning of time by a factor of 10^50.
It's a reminder that combinatorics doesn't care about your intuitions. Linear growth, even exponential growth, we can sort of feel our way through. Factorial growth leaves intuition behind almost immediately. 10! is 3.6 million. 20! is 2.4 x 10^18. By 52! you're in a region of number space where physical analogies become performative rather than informative. I can tell you it's more than the grains of sand, but that comparison doesn't actually help you feel how big it is. Nothing does.
Every shuffle is an orphan
There's something quietly strange about this. Every time you've played cards, the specific order of that deck was a one-off. The particular sequence of hearts and clubs and diamonds and spades that sat in front of you during that boring game of Rummy at your nan's house on Boxing Day 1998 was, in all likelihood, a configuration that had never occurred before and will never occur again, anywhere, for the remaining lifespan of the universe.
You didn't notice because there was no reason to notice. The order of a shuffled deck isn't meaningful. No one writes it down. It exists for a few minutes between shuffles and then it's gone, replaced by another unique arrangement that also goes unrecorded.
Billions of these one-time arrangements get created and destroyed every day. Each one is technically a miracle of probability, in the sense that its specific occurrence was vanishingly unlikely. But because there are so many possible arrangements, one of them has to come up every time. The miracle is guaranteed. It's just never the same miracle twice.