For one, "low" is rather relative, it's rather subjective on what value "low" is. But given the usual term of "low chance", "(normal) chance" and "high chance", it's fair enough to say that anything on the lower 1/3 is "low chance". 4/12 is in fact 1/3 and 5/13 only slightly above..

Secondly, there exist a low chance that a low chance event happens frequently.
For example, statistically speaking, there is a 0.65% chance that a 10% event will occur 5 or more times out of 13. And I'm sure even you would agree that 10% would been properly labeled low chance.
So really, small sample size never says anything conclusive. You really have to gather a sample of at least 100 attacks to really tell if it is, in fact, low chance (or rather, what the likely chance is, as "low chance" is subjective)

Well, there is a 0.95% chance to only hit 4 or less times out of 12 if the failure rate is 30%.

So even if it was on that extreme end, it doesn't prove anything. My point stands that it takes at least 100 attack to prove anything.

I don't watch videos posted on this forum much due to.... reasons...