[mistersquid]

Predicting election outcomes based on polling (including exit polls) is very different than alleging massive voter fraud.

In the case of Clinton losing the election, the predictions which were based on statistical evidence did not match the actual outcome. Even so, an unlikely outcome is understood as a possibility within the domain of predictive statistical modeling and the wholesale dismissal of the utility of statistical modeling is anti-science at best.

On the other hand, asserting voter fraud without providing a single shred of evidence (regardless of claims to have said evidence) is quite different from evidence-based statistical modeling. So people who "thought Hillary was going to win in a landslide" could actually have quite a bit to say about alleging voter fraud without producing evidence. For example, such people could say that Clinton won the popular vote by a margin of nearly 3 million votes.

To turn your statement on its head, it is in fact the people who allege voter fraud without evidence who have nothing to say.

EDIT: spelling, reduce number of single-sentence paragraphs

[masonic]

  predictions which were based on statistical evidence

Polls aren't evidence. In this case, they were barely data.

There are polls which are designed to gauge opinion, and there are polls which are designed to steer opinion.

  evidence-based statistical modeling

Given that the model failed, it apparently wasn't evidence-based.