I'm sure this thread will be bump material no matter what happens.
Also, my understanding is that this model actually started predicting in 1996 (per here).
He got it wrong in 2000, fixed it up a bit, and this is going to be the 6th prediction? Doesn't sound like a big sample size.
But hey, I think it's still early to make a call on this. Definitely a weird electoral year, and a lot of things that can change from here to november.