He's always been very open about the models they use (though not the parameters of the secret sauce) as well as the errors they've made, which says a lot to me. And while they were off, I do think they've been earnest in figuring out what the limitations of their model were: as they did during the campaign as well. They've also been very clear in pointing out the difficulty in making predictions of singular events. Others may pay lip service to this, or say nothing all.

You could also read this as him being weasels from the get go. From what I've seen, that's not warranted in his case.

A last point I always have to remind myself of: human psychology just doesn't do stats well. That's why we have these tools. Kinda like quantum physics or relativity. We are smart enough to figure out the math, but the results—or our interpretations of them— don't always jive with our intuitions.