Well, a handful of my favorite writers have followed a similar path. Joe Posnanski wrote for Sports Illustrated (back when that was a thing), The Athletic, various other high-cache publishers. But now he writes primarily (solely?) for Substack, supported by memberships. It seems to be a more sustainable and lucrative model for writers with enough fans.
I highly doubt Silver made the move from NYT and ABC to Substack without considering the financial ramifications of that decision.
Here are my reasons to put more weight in Silver’s model than in my gut:
He nailed the Obama surge way back in 2008 way before anybody else thought he could win that nomination.
He had a much more realistic take on the 2016 race than most other Dem-leaning pundits. When somebody asks a question like the one in the OP, my default answer is “we said the same thing in 2016, and Trump won”.
He correctly predicted very early on that Biden’s campaign was untenable and that in order for Democrats to have a chance he needed to drop out.
So at this point I have no issue with Silver’s current prediction of Trump being a 60% favorite to win. That’s in line with what betting markets are saying, and with my basic reading of the polls. The Electoral College bias is pretty much set at 3-5%, so Harris needs to be polling in that range for it to be 50/50 if you assume that any polling errors are unpredictable in direction. And she’s not - she’s ahead by maybe 1-2% nationally.
If you want to assume that the polls are biased in Trump’s favor, fine, but I don’t know what evidence you can use to support that. If you want to assume the polls are biased in Harris favor, that’s actually a bit more supportable since Trump has over-performed his polls in both previous elections he has been a part of.
And while having the NV/NC/GA/AZ alternative path available is nice for Harris (not relying on PA entirely) there is the unfortunate reality that even if she gets one of GA or NC, she needs either AZ or NV as well, and she’s behind (or tied) in both.
So yeah, the answer to the OP is easy - he wins by winning in PA. Where it is currently tied at best, with multiple recent polls showing Trump ahead there.
Criticism of Silver’s model for being secret in its details, and proprietary, is totally reasonable. I do not think there is a sufficient number of presidential elections for a few points, one way or the other, to be meaningful. But trying to highlight this, by say, rounding to the nearest multiple of five percent, would IMHO make the model less useful, as it would disguise trends.
My problem is with blaming imperfect messengers for numbers that go against a favored candidate. Being under ABC auspices doesn’t make a model better or worse.
The things you said in defense of Silver, however, seem to be more about his personal insights into politics and less about his model. So I don’t know if they justify his current 60%.
538 has Harris back up to 55% this morning, FWIW.
I don’t know about the argument that Harris needs to be so many percent ahead nationally when it can be argued from swing state polls that she is looking OK. That’s an argument from correlation; i.e., in the past, when Democrats won the Electoral College, they tended to be up such a percent nationally.
OK, I will be driving today but will have more to say, I’m sure, after the debate!
Please be specific when using percentages. I’m guessing the “55%” is the modelled chance of winning the electoral college. And your “percent ahead” is referring to a polling results.
There is some statistical (or at least quantitative) basis for putting an approximate number on this (with the usual caveats that each election is unique, etc., etc.). This time around, according to several experts, it’s basically:
Harris up 2% nationally? Trump slightly favored to win the electoral college (though he might not, of course).
Harris up 3% nationally? Truly a crapshoot. (And, this is the situation most likely to result in yet another “popular win, electoral loss” for the Democrat – something Silver, for one, currently pegs as 1-in-5 chance).
Harris up 4% nationally? Harris favored to win the electoral college as well. (Trump could still win the EC, but it would be quite unlikely – something like a 1-in-10 chance. But that’s still a chance!). More than one expert has said this is when Harris supporters can start to “breathe a little easier.” Not take the foot off the gas – just breathe easier.
Yes, but each state delegation gets one vote. Even if the Dems take the House they won’t win a majority of state delegations.
ETA: I guess there is some possibility of “non-faithful” Reps abstaining or somehow flipping GOP-held states to neutral or Harris. But that has to be a <1% chance.
Haha, fair enough. But that’s like answering “How can the Bears beat Houston?” with “Score more points than they do.”
I’m really interested in the strategies (such as they are) and voter mindsets that can enable Trump to win the states he needs, when any sane analysis would have him losing 99% of the vote. But of course that’s been the topic of innumerable threads and posts over the last eight years.
Well, here’s one small (but probably real, and maybe – just maybe – crucial) factor that combines Jas09’s list with what you seek:
No longer having someone from Scranton, PA on the ticket.
Seriously – Uncle Joe did better than he otherwise “should” have (the “should” based on other voting data) in PA in 2020 partly because of his overperformance in the greater Scranton area.
So there’s that. The “mindset” is “he’s one of us, and she isn’t.”