The whole overdue thing is also a logically fallacy.
If something has an x% chance of happening per year and it hasn't happened for many years it has an x% change of happening next year, not some greater percentage.
That's not generally the case when dealing with things involving plate tectonics, which often have a build up of pressure over time (tens, hundreds, thousands of years). The tension can slowly build up at a plate boundary before snapping to relieve the tension.
In general with phenomena like this it makes less sense to start from the assumption that the root cause for something is just a random number generator that the Earth's programmer decided to hook up to add variability.
Rather, things in the real world have actual causes, and if they typically happen on the order of 100,000 years apart, there's probably a time-need reason for that.
(There are plenty of phenomena that are just statistics, of course.)
Also, even if it would be possible to compute that an earthquake/eruption is overdue I've yet to see it done correctly. Obviously I don't know where the person I'm replying to got their statistic. But if you Google something like "Yellowstone overdue" the results use very poor logic, often just extrapolating from three data points millions of years ago.
The odds of some natural event happening aren't always independent from year to year. If a volcano eruption is caused by a buildup of pressure past a certain threshold, and pressure is actively building in an area, then going longer than expected without an eruption absolutely would increase the chances for upcoming years.
The gambler's fallacy. It's landed on black for the last 20 rolls, it must land on red soon! No, it will land on red with 50% probability, just like the last 20 times.
If something has an x% chance of happening per year and it hasn't happened for many years it has an x% change of happening next year, not some greater percentage.