It's that time of week again, I did a few changes to my methodology a little bit I'll get into, but otherwise, just to present the data.
Presidential
Well ####.
What do I mean by that? Well, look at the popular vote share. I don't focus on popular vote as electoral college matters, but we just saw a one point swing in the national environment overnight. And before people think "well its right wing polls", no I tried removing the right wing polls from the averages for the heck of it, and it made no difference. The numbers really are that bad. And many pollsters are now saying the polls they're coming out with are their "final polls", so while I expect a lot of polls over the next week, some pollsters are basically locking themselves in...with these numbers. This does not bode well for Harris at all.
On the electoral college level. Not has changed. Minor fluctuations. The election is currently 44-56 in Trump's favor, with a 226-312 map. Not really great, but hey, Trump had odds like this in 2016, and anything could happen. However, Trump had a trend behind him in 2016 showing Clinton falling apart and now Trump has...a trend showing Harris falling apart. Well ####. That's why I keep saying that. Not good. Not good at all.
Anyway, I did do a couple methodology changes here. There are states that have been kind of a thorn in my side with my current chart. What do I mean by that? Well, they got like 1-2 polls for them that RCP records, they're weirdo outlier polls that are probably wrong, which we get from a lot of borderline swing states, and RCP just aint updating the data, despite new polls available. Idk if it's because RCP doesnt deem them trustworthy, they don't care, but I got OCD and much like that one green spot on cell's perfect cell games arena, I wanna get rid of them. So yeah. Let's discuss the states and why I decided to shift over to 538's data.
Maine- Okay, RCP has two polls. a D+9 and a D+17 poll. For ME2, this corresponds to a R+7 and D+5 poll. Does anyone really believe maine is going to be 13 ahead? Or Maine CD2 being a tossup? NO, OF COURSE NOT! Heck, ME2 was one of the polls I got wrong in 2020. It was like D+3.7 or something and ended up going like R+7. I expect that to happen this time. So switching to 538's data with more polls (but still including the two above polls), I get D+11 for Maine proper, and R+5 for Maine 2. I still think it will go about 2 points more red than that, D+9 for maine proper and R+7 for Maine 2 seem right, but still, close enough. I'll let the polling data take it from here.
Oregon- Likewise, Oregon is NOT going D+5. It's pretty much a safe state and had some really outlier poll with nothing to offset it. 538 has another poll with it going D+12, which sounds more right, so I'm letting the average of D+8.5 do its magic. Again in both these cases, I'm not throwing out polls i disagree with, polls are polls, outliers happen, but when RCP isn't adding new polls to mitigate the effect of such outliers it's gonna mess up my forecast.
Iowa- Same thing but on the R side. Like ME2, I think it's gonna go R+7. Instead, RCP has ONE POLL, which has it going R+4. 538 has 2 polls, with the one being the above R+4 poll, but also an R+7 poll, so yeah, once again, averaging that out.
Alaska- RCP doesn't even have data on Alaska, but it was relevant on the Biden map but 538 does, so I used 538's data for that for a while.
I'm not changing anything else over to 538. Quite frankly, I trust RCP more than 538 other than that due to their track record and history. I've always used RCP. And quite frankly other than those examples and maybe the popular vote, the two don't have greatly differing results anyway. No really, they don't. 538 would give me a 47-53 spread rather than a 44-56 one and maybe the rust belt states are like half a point more blue, but that's about it. The two sites are converging.
Also, looking 538's actual polls that make their forecast different, they seem to have a lot more churn, only keeping the newest 5 or so on their page. And a lot of the polls are questionable. Like you'll have D+1, R+1, D+2, a tie, and then the 5th is from some pollster RCP doesn't track because they're probably garbage, and they're like D+8. THat's a huge reason why their forecasts are more left than RCP. otherwise they're using similar data and converging on RCP's data mostly. I just like how 538 sometimes keeps track of less followed states a little better. RCP tends to struggle with that. So I'm gonna use 538 to enhance my forecast where RCP drops the ball, but I'm still primarily going by RCP's numbers.
Senate
We're still having a lot of tightening up in the rust belt. Ohio remains a tossup, wisconsin is now a tossup. Pennsylvania is trending toward a tossup but is still lean D. Nebraska is a tossup in the Osborn race. And while I still formally predict a 50-50 split, the odds of the GOP being able to just control the senate have been strong for a while and are getting stronger. I'll discuss this more in the simulation section since it's readily apparent in the simulator.
House
Once again, "well ####". the generic congressional vote dropped quite a bit, dropping over a point in like 2 days, and now the GOP is ahead. Things really are falling apart here and I feel like I'm one of those violin players on the deck of the titanic as it's sinking. I now have the GOP having a 60% chance of winning the house outright, and if I had to guess the remaining tossups, the florida district would likely go GOP and the other, idk, maybe R at this point due to the republican leaning generic congressional vote. So that's functionally a 215-220 house controlled by republicans. Narrow, but yeah, they're probably gonna take it. Let's see how the simulations do.
Simulations
Given the presidential simulation is not affected by the national popular vote, we're going by the electoral college, and yeah, we got 561 Trump wins, 423 Harris wins, and 16 ties. The median outcome is a 266-272 republican win. The most likely tipping point is obviously PA at 35%, but Michigan, Wisconsin, and North Carolina remain popular as well.
For the senate, we got 843 republican wins, 18 democratic wins, and 139 ties. The simulator favors the republicans, ESPECIALLY as there are 3 tossup races the dems absolutely need to win just to get a 50-50 split. The simulator often has them failing to do this since each race is its own coin flip. The median outcome is 48-52 republicans. Nebraska is the real tipping point, but Wisconsin, Ohio, and Pennsylvania are all likely to do so too. Funnily enough, the dems winning outright are so low all 3 tipping points for them are below 1%. Texas is the most likely one, followed by Florida, and yeah, it really does look like we're losing Tester's seat in Montana.
For the house, again, doing the old simple simulator with 100 outcomes, not 1000, and no dashbord of fancy data, but the GOP wins 78 of those, the dems 22.
Conclusion
And yeah, it looks like we're probably gonna get a GOP trifecta....with a fascist on the republican ticket. And a republican supreme court.
This is NOT good. Really, ICEBERG, STRAIGHT AHEAD, TURN THIS SHIP AROUND!
This is scary. This is a nightmare scenario for the country. I mean I can live with a republican congress, but Trump? Trump's dangerous. We gotta take SOMETHING and obstruct his agenda. This is really bad. I'm sounding the alarm bells as hard as I can on this. We are NOT facing a good outcome in november here. This is a slow moving train wreck for the dems, it really is. If you're dem leaning, just get out there and vote. if you care about keeping america free, VOTE! We need every vote.
No comments:
Post a Comment