[iiiieu]

Rejecting statistics that you don't like while pushing those that you do is a case of selection bias, the extension of which is the infamous "the election was rigged" trope.

It's easy to select a subset of trains to make the statistic of this subset look bad. Take a single train that was canceled. In that subset, the "on time" statistic is 0%. Really bad, eh? Or take one that was on time. 100%. Great, eh? Well, not useful, just like all other anecdotes.

So, yes, sure, we can all pick the subset for which the statistic suits our narrative and then keep disagreeing about the facts. It's not a useful way of debating though and only suits lobbys in pushing their agenda.

(Or we take the facts as they are.)

P.S.: Even the 80.6% that you quote are not "really bad". Ever been in North America? No train traffic at all sounds worse than 80.6% to me. The German train system has its issues, but it's by far not as bad as the public perception seems to believe.

P.P.S.: Check out the German "Staustatistik" some time. Over a thousand every day. Of course nobody who wants to make trains look bad and implicitly push the car as the better alternative talks about this.