Sleuthing thread: Blue firewall in PA

[superrific]

So I feel good about MI and WI. I'm worried about PA. One source of optimism has been this idea of a "firewall" of early ballots. That is, if Dems can get at least 400K more Dem voters to vote early (without knowing who they vote for) than Pubs, that will be a good sign for election day.

But where did this number come from? I had thought it had come from the campaign, but actually I can't find that. I've seen some references on reddit -- and how could anything on reddit be wrong -- to a single data analyst who was pushing this idea and had maybe made it up. Anyway, supposedly the number has increased to 410K and then to 430K. We're going to hit 400K. I don't know if we will hit 430K.

Oh, and then this data analyst said that the firewall was actually what is necessary to get this into tossup territory, which isn't what firewall usually means. I mean, sure, but that means if we miss it, according to this guy, we're not in great shape. And if it's 430K and not 400K . . .

But then there's this: this guy is some amateur. Says he has degrees in aviation management from Purdue and is a data analyst (but for whom and based on what, who knows). So maybe he doesn't know shit. He posts under his real name (one supposes) which is i suppose good; but his twitter handle is @blockedfreq, which does not inspire much confidence. I guess it's better than being catturd, but anyway.

Whoever found references to the blue firewall, where did you find it? Is there any actual legitimacy to it?

 

[lightbluenc]

I'm re-posting what I put int he turnout/data one:

I'm very curious about that "firewall" as well.

And I know we can't correlate 2020 with 2024 since it was a weird time. Buuut I'm going to anyways:

2020 Mail-in ballots returned
Democrats 1,702,484
Republicans 623,404
No Party Affiliation 283,673

So in 2020 the "firewall" was over 1 million + Dems. So why is this 400k number supposed to be so good?

 

[superrific]

Don't look at the 2020 numbers at all. It's like comparing RJ Davis' career point total to Phil Ford's.

This is, by the way, why the PR thing is so important. I've read that PRs in PA (and I think Hispanics in general in the state) tend to vote on election day. So if Trump just gifted Kamala 30K votes from the PR community (from increased turnout or vote switching), that could be pretty important.

 

[Rock]

I'm not very smart but amusingly, Pennsylvania is the state I'm most sure about based off of all the stuff I've seen over the weeks and the fact that Trump is melting down about Pennsylvania right now.

 

[rodoheel]

Bottom line: I think trying to read the tea leaves based on number of early votes by party affiliation is ultimately going to be a pointless endeavor. I do think the "firewall" numbers are essentially made up; it's too hard to dive how party affiliation will translate to vote totals; and it's too hard to compare to prior elections (2016 was basically a lifetime ago in terms of how people vote, and 2020 was an outlier in every respect). I think it's all basically guesswork - mostly blind guesswork - and if anyone ends up being "right" about what they divined from the pre-election-day numbers it will likely just be by random chance.

 

[CFordUNC]

I'm not very smart but amusingly, Pennsylvania is the state I'm most sure about based off of all the stuff I've seen over the weeks and the fact that Trump is melting down about Pennsylvania right now.

Right lol. This times a million. Trump isn’t preemptively melting down annd screaming fraud about any other state.

 

[aGDevil2k]

1. D vs R doesnt indicate as much
2. Even in PA, Rs are mailing in votes more than they did 4 years ago. Trump's whole campaign puts up signs about voting early.

 

[rodoheel]

Right lol. This times a million. Trump isn’t preemptively melting down annd screaming fraud about any other state.

I hope you guys are right, but it's the blue wall state I'm most worried about. I think Trump would be crowing just as much as about a supposed voter fraud investigation in WI or MI if there was one to crow about.

 

[superrific]

Right lol. This times a million. Trump isn’t preemptively melting down annd screaming fraud about any other state.

But there aren't BS stories coming out about other states. And also, the PA Supreme Court has been active and they are appealing one case to SCOTUS. So there's more ammunition there. That said, I obviously have zero visibility into what goes on in the Trump campaign.

I suspect he's more melting down because of the way he shat the bed with the PR community.

 

[uncjhodges]

Josh Smithley seems to be the PA version of Ralston. He originally had 400K as the estimate and then revised that to 500K. This is assuming similar crossover numbers and a decent early vote independent split for Harris

We are not going to get to 500k. maybe 425k. I fear this is ballgame

 

[superrific]

Bottom line: I think trying to read the tea leaves based on number of early votes by party affiliation is ultimately going to be a pointless endeavor. I do think the "firewall" numbers are essentially made up; it's too hard to dive how party affiliation will translate to vote totals; and it's too hard to compare to prior elections (2016 was basically a lifetime ago in terms of how people vote, and 2020 was an outlier in every respect). I think it's all basically guesswork - mostly blind guesswork - and if anyone ends up being "right" about what they divined from the pre-election-day numbers it will likely just be by random chance.

I agree with that. I have some time for pointless endeavors at the moment, as I'm having some trouble getting work done when I've got election anxiety. And while the data might not mean much in the end, it still makes me feel better to be on the good side of the data rather than the bad side. It shouldn't, but it does. At least a little.

I don't think firewall numbers are necessarily inaccurate (esp if the true meaning of firewall is being used), but as you say, they need more context. The campaigns might have that context. Independent analysts, probably not. So that's why I want to know if any of this data has basis in the campaign or if it is just totally invented.

I'm almost positive the Kamala campaign has internal targets for voting turnout in different places and times.

 

[superrific]

Josh Smithley seems to be the PA version of Ralston. He originally had 400K as the estimate and then revised that to 500K. This is assuming similar crossover numbers and a decent early vote independent split for Harris

We are not going to get to 500k. maybe 425k. I fear this is ballgame

Link to Josh Smithley being a PA version of Ralston?

I read that the 400K estimate was not revised to 500K but rather the 500K estimate includes independents. Either way, this is definitely NOT ballgame.

Also, Ralston isn't a forecaster. He reports. Ask him what is going to happen with the election now and he'll say he doesn't know. He's valuable during the counting, when he can focus on specific places and specific numbers and fill in gaps.

 

[aGDevil2k]

Josh Smithley seems to be the PA version of Ralston. He originally had 400K as the estimate and then revised that to 500K. This is assuming similar crossover numbers and a decent early vote independent split for Harris

We are not going to get to 500k. maybe 425k. I fear this is ballgame

So you dont think the GOP has been listening to Trump and voting by mail more than they did before? They finally stopped listening to him on anything and this was it?

 

[uncjhodges]

Link to Josh Smithley being a PA version of Ralston?

I read that the 400K estimate was not revised to 500K but rather the 500K estimate includes independents. Either way, this is definitely NOT ballgame.

Also, Ralston isn't a forecaster. He reports. Ask him what is going to happen with the election now and he'll say he doesn't know. He's valuable during the counting, when he can focus on specific places and specific numbers and fill in gaps.

He has a substack that i just began to follow. Pennsylvania Powered. He’s also on twitter but i’m not on that- think it’s the same name
I also follow @realcarlallen as a statistician and still has Harris the favorite. I’ve felt like he’s been a good follow

 

[HeelInTheOzarks]

I hope you guys are right, but it's the blue wall state I'm most worried about. I think Trump would be crowing just as much as about a supposed voter fraud investigation in WI or MI if there was one to crow about.

Wisconsin will be the closest one again, as it was the last 2 cycles.

 

[superrific]

Josh Smithley seems to be the PA version of Ralston. He originally had 400K as the estimate and then revised that to 500K. This is assuming similar crossover numbers and a decent early vote independent split for Harris

We are not going to get to 500k. maybe 425k. I fear this is ballgame

OK, if this guy Josh Smithley is the source for the numbers and they aren't coming from anywhere else, then we can just throw the whole idea out.

According to Linkedin, Josh Smithley graduated from Purdue in 2022 with a double major in Aviation stuff. Now he lives in Grand Rapids MI. It doesn't appear that he knows anything about PA.

 

[uncjhodges]

Link to Josh Smithley being a PA version of Ralston?

I read that the 400K estimate was not revised to 500K but rather the 500K estimate includes independents. Either way, this is definitely NOT ballgame.

Also, Ralston isn't a forecaster. He reports. Ask him what is going to happen with the election now and he'll say he doesn't know. He's valuable during the counting, when he can focus on specific places and specific numbers and fill in gaps.

Ballgame was too defeatist.

 

[superrific]

Wisconsin will be the closest one again, as it was the last 2 cycles.

And also the one that can be replaced by any other swing state.

PA is the most important one because it takes two to match. If we win NC, almost the only way we can lose is losing PA. I suppose we could lose MI and WI but I think MI is lean blue and if we lose both of those, seems unlikely we win in PA.

MI is semi-important, in that it can't be made up by AZ but can be made up (I think) by either GA or NC. Semi-important compared to PA, is what I mean.

 

[superrific]

Ballgame was too defeatist.

It was. I saw that the guy has 1000 subscribers to his substack. I don't know why we would think this guy has anything to contribute.

 

[rodoheel]

Wisconsin will be the closest one again, as it was the last 2 cycles.

My (completely unsupported, not necessarily rational) belief (guess) is that Walz and good, old-fashioned upper Midwest sensibilities end up leading to a win there, and likely to Kamala doing better there than in PA.