Number of benchmarks in critical difference plot

I tried to understand how plotting critical difference in fuzzbench works and I see that you use [number of benchmarks](https://github.com/google/fuzzbench/blob/master/analysis/plotting.py#L479) when generating a plot. From what I understood, if N is meant to be the number of datasets then I think this should instead be the _number of trials_ in our case?

For example given two experiments (trials=10 and trials=100) having the same ranking, should the experiment with 100 trials have smaller critical difference?

Trials = 10

Trials = 100