First of all, you're correct. For short-term timing, say, in seconds, it might be critical, but, events such as you point out are fairly random, hence can go either way ... hence, will average out over the long term. Normally one would resistor-divide the mains voltage and ac-couple it into a Schmidt-trigger, and enable its interrupt 16.6 ms (not precisely 1/60 second, or in your case, 1/50 second) ms after each count, then disable it immediately once it occurs. This is, after all, a clock ... one for people to look at to tell the time-of-day. The long-term stability is what matters.

One could, of course, use a high-precision time reference, batteries, precision filters, etc. ... but why? It is, after all, a student project. Correctly specifying the hazards will get the student as much credit as designing them out, unless coping with them was a pre-specified requirement.

RE