Thank you.
That people refuse to accept the posters' *unstated* assumptions (fully random shuffles even though physical shuffles are not, a randomized starting deck setup) is a sign that the posters have failed to relay their assumptions in advance. To me, it's unproductive to complain about how the posters don't understand math. Simply refine your point.
It's really, really important to consider both the theoretical statistics and the practicality of implementation to properly frame the issue. You need to both recognize that 52! is absurdly enormous and recognize that people can't shuffle perfectly. Anyone just looking at the numbers or just looking at shuffles is missing half the problem.
Why do I say this?
Ironically, the folks saying it is *impossible* are functionally ignoring the incomprehensibility of infinity. As the number of cards in the deck approaches infinity, the actual probability of a repeat (with perfect randomization of shuffles) approaches zero. 52! is a very very very long way from infinity.
Thus, this *cosmically unlikely* event is *comparatively probable* at 52 cards. To make a repeat shuffle physically impossible you need an infinite deck.
Sure you can be *pretty goddamn certain* that a randomly ordered deck of 52 cards will occur in a configuration that has never happened before and will never happen again, but that is not the same thing as mathematical impossibility.
You don't need an infinite deck to make a repeat sequence "physically" impossible. You only need an infinite deck to make the probability go to zero. In order to make a repeat sequence "physically" impossible as human beings with 52 card decks, you just need an infinitesimally small time window to run your test in. Given how large 52! is, the span of all human existence happens to be a plenty small enough window.
"But the probability is still not zero!"
But there are also a lot of other things that this 'mental experiment' assumes have a probability of zero that actually have a non-zero probability in the real world. For example: the probability that someone running the experiment makes a mistake, or the probability that they're lying, or the probability that the computer they're using for record keeping gets
struck by a cosmic ray and modifies the history of shuffles you've been tracking...
...or most likely: the probability that you have calculated the probability of the event wrong because you left out some external factor.
When I say it's physically impossible, I mean that the probability of something interfering with the experiment dominates the probability of the event 'naturally' occurring.
This is where the importance of the errors inherent in physical shuffling come in.
If you can prove with absolute certainty that nobody is lying, and can provide
irrefutable evidence that two identical sequences actually happened, then the rest of us have to decide what physically happened: either you really got that cosmically lucky, or there is a flaw in your shuffling method. And 1/52! is so absurdly small that the probability of your shuffling method being flawed absolutely dominates it beyond a shadow of a doubt.
I posted the story about the
perfect bridge deal earlier: a shuffle result that can never happen, but it did. But obviously it did because the probability of that event occurring from a freshly opened deck is significantly higher than the probability of that event occurring from an arbitrarily mixed deck. That's an example of what I'm getting at: if one of these 'cosmically unlikely' events happen, you can say with absolute certainty that
something else is going on.
In closing - obviously this argument is semantic. If I put it into explicit terms, 'impossible' means 'if someone claims it happens, the rest of us should assume they're wrong.' And from the perspective of the human experience, that definition is
functionally the same as 'the probability of that event occurring is zero,' even though 1/52! is obviously > 0.