in a series (one single, continuing series!) of coin tosses, the probability that you get a run of heads at least half as long as the overall length of the series (eg ttththtHHHHHHH) is always >0, but it is not guaranteed to happen, no matter how many chances you give it.
… any event for which you don’t change the epsilon such that the sum becomes a convergent series. Or any process with a Markov property. Or any event with a fixed epsilon >0.
That should cover round about any relevant event.
(and also you mis-apply the Law of large Numbers here)
Law of Large Numbers states that sum of a large amount of i.i.d variables approaches its mathematical expectation. Roughly speaking, “big samples reliably reveal properties of population”.
It doesn’t state that “everything can happen in large samples”.
… any event for which you don’t change the epsilon such that the sum becomes a convergent series. Or any process with a Markov property. Or any event with a fixed epsilon >0.
That should cover round about any relevant event.
Explain.
Law of Large Numbers states that sum of a large amount of i.i.d variables approaches its mathematical expectation. Roughly speaking, “big samples reliably reveal properties of population”.
It doesn’t state that “everything can happen in large samples”.
Thanks. Memory is more fragile than thought, wrong folder. Updated.