The implication made is that the chances that it happened are small. This has no bearing on whether it actually did or didn't happen, as either way it has already occurred. So my claim that it is useless to use statistics to predict the past stands. In other words, how likely it is that something happened is not proof that it didn't happen.
Yes it is.
This isn't a case of the odds against a winning lotter ticket being proof that someone didn't just win the lottery.
It's a case of the odds against the likelyhood that the winning ticket was formed by someone spraying black paint into the air, and it happend to land on a random piece of paper, which was exaclty like a lottery ticket, and the droplets of black ink exactly formed the winning number, in the correct font, etc. etc. etc.