I wonder a couple of things.

First, do the projections you’re looking at make conventional assumptions about turnout based on recent elections? Given the huge of disparity of turnout for the respective candidates’ rallies (and to some degree the primaries) that could be an essential mistake.

Second, do your projections consider a possible “Bradley effect,” where people lie even to pollsters about their actual voting behavior when they fear their choice might be considered politically incorrect? I know that pundits on the left call this possibility phoney. However, as you surely know, a common tactic of pundits on both sides is to offer such “analysis” in an effort to make it come true.

You talk about your data as if it reveals an objective picture of reality. Maybe it does. But every once in a while comes a time when the edifice of heretofore reliable assumptions about how the world works has become invalid, but the wonks, pollsters, pontificators and statisticians just don’t know it yet.

Based on the course of the election so far, this feels like it could be one of those times, and to dismiss it so off-handed feels like hubris.