I don't claim to be expert in the mathematics of probability, but this doesn't sound right to me.
Let us assume an event has a probability of happening once in 1M years. Which means of course a one in 1M chance for any given year. Give it a million total years, and the probability it will happen is 1:1.
In one billion years the probability of it NOT happening is 1 in 1000. In a trillion years 1 in 1M.
But the chance of anything happening, no matter how likely or unlikely, never reaches absolute, no matter how much time is involved. Really, really likely or unlikely does not equal mathematical certainty.
I especially liked the article's blithe assumption, if I was reading it right, that there is a 1 in 10M chance the universe will end this year, and that this means it will most likely come to an end in 10M years from now. That is most definitely not how probability works.
Assuming you are actually referring to a random event which happens, on average, once every million years, that's not right.
If it happens, on average, once every million years, the probability of it happening in any given year is 0.000001.
The probability of it not happening in any given year is then "1 - 0.000001", or 0.999999.
The probability of it not happening in each of two years is then 0.999999 * 0.999999, or 0.999998000001.
The odds of it not happening in a million years is then 0.999999 to the millionth power, or roughly 0.37; in any given million-year interval there's a 37% chance of it not happening at all.
In fact, extending to a 10-million year interval, there's a roughly 0.00004 chance of it not happening in that entire period.