Election update 7/12/24

Ok, so we should already have a good idea of what my prediction is gonna look like based on stuff I've been sharing all week. I did make an updated chart with the latest data though. 

So, the current tipping point is Nevada at 5.2% Trump. That gives Biden a 9.7% chance of winning, and Trump a 90.3% chance. Hence why I keep saying Biden has a "10% chance" (it rounds to 10%). 

This isn't good. Based on previous election cycles, it's not out of the question to overperform by 2-3% with some level of consistency, but 5 is kind of beyond the pale. It has a 10% chance of happening in theory, but this is pretty much "screwed" territory. Biden said yesterday that he'd get out if he had "no chance of winning." I won't say he has "no chance", but what counts as "no chance"? As you can tell by the chart you can be down by 12 and still have a 0.1% chance. So what's "no chance"? 10%? 5%? 1%? 0.1%, we should set some standard here.

I personally think if you're below the 16% chance that comes with a 4 point deficit, pack it up, you're probably not gonna win. A 25-30% chance is arguably still in "turn it around" territory, but again, despite the percentages existing on a bell curve, well, evaluating my own past predictions, they dont look like a bell, but like a top hat. Anything <4 points you can have around a 25-30% probability based on observed results from past elections if you're on the losing side of it, but >4 points and basically you're screwed. 

Now, the real measure here should be, can Harris or someone else do better? I admit, we don't have great data on the state level as most of it is old. Back in November when I looked at it, Harris seemed WORSE. However, as we can see now, Harris is seeming better. She has a 0.8% advantage over Biden nationally. if applied to state level polling, we could say she could bring that 5.2% deficit down to a 4.4% one. That would put her at a hypothetical 13.6% chance. Not much better, but it is better. As I keep saying when you're down this far you're kinda screwed. It's possible my "top hat" distribution of actual observed results is just a bell curve with too small of a sample size to conclude anything substantial. Either way, 68% of results happen within one SD/MOE of 0, and given it only matters one way, again, if you're on that tail end of the distribution with only 16% of results, you're not doing very well. And until you can get up into the the main part of the bell curve, you're not likely to win.

I would say, functionally, that a <16% chance is functionally no chance, and Biden should at least be in talks about passing the torch. 

As for the electoral map, I mean, we've seen it before. I'll post it again.

Based on this, we can say we can honestly say the 226-312 Trump result is still a thing.

And btw, it could be worse. The problem is, a lot of states don't have consistent polling data. Most polls look at the 7 in the Trump column. Wisconsin, Michigan, Georgia, Nevada, Pennsylvania, Arizona, North Carolina. 

But keep in mind, we have very little idea what's going on in Nebraska CD2. If it shifted 2 points to the right with the rest of the nation, it could be in the Trump column. We also need to keep an eye on Maine, Virginia, Minnesota, and New Hampshire. Those COULD be flipping but because they're not treated as swing states by many, the data is largely unnknown. Heck keep in mind, that one internal openlabs poll that leaked had Trump winning NEW MEXICO. So this could actually be worse than it looks. I'm kind of limited by the data. 

Btw, I also decided to present the third party data this time. And I'll show you exactly why I don't very often.

Does Biden REALLY have a 22.7% chance there? No. It's just that most old data has Biden still at his pre debate averages and that's what makes up the sample of data here. Honestly, the results seem slightly worse on the whole than the 2 way data, but not in a way that gives us a different picture or actually tells us anything.

Now, before I move on from the presidential election, I would like to post the trend chart again so we can see how this race has been evolving:

 We are SCREWED. Biden has NEVER, AT ANY POINT, really come close to winning here. And now the bottom really fell out on him. I do think we are kinda seeing the inflection point where things should start somewhat returning to the mean, but given the "mean" is something barely inside "so I have a chance" territory, that's not good. 

Btw, I did make 2016 and 2020 charts of this graph, let's look at them for comparison.

To be fair to Biden, he might have moments getting closer to election day where something like this can happen. In 2016, we had that september spike when Hillary was seen crawling to a car opening up questions about HER health, and of course, as last minute voters decided in 2016, the results evened up a lot, and I actually had Clinton at a 56% chance on election day. So is a turnaround possible for Biden? Sure. Would I count on it? No.

In 2020, it was never close, but to be fair, I culled the polling averages to remove pro Trump polls that I deemed unscientific. Massive mistake, it actually ended up being much closer than we thought.

So...that said, can Biden still win? In theory, if he overperforms like Trump did in 2016 and 2020...against Trump, but keep in mind, it's equally probable it goes the other way and VA/MN/NH/ME fall to Trump too. 

That's the thing about probability. The above predictions are essentially the median prediction within the realm of possibilities. Most outcomes are pro Trump, with 50% of them being more pro Trump than what I have above. 50% of them are LESS pro Trump, with Biden winning 10%. I had Trump winning at 11.5% in 2020 (although again my forecast was erroneous, had I used existing methodology his odds would be 62%), and 44% in 2016 (meaning his unexpected win wasn't that improbable at all). 

With that said, let's move on to the senate. 

More significant changes here as we got more polling. And now we're losing Montana. And now the most likely outcome is 49D-51R. I got the GOP at a 64.6% chance they win the senate now, a 7.7% chance the dems win, and that leaves us with a 27.7% chance of a tie. That's not good. But, it could be worse. Most of the seats in the D column are states that the dems are losing in the presidential race, so whatever people think about Biden, it doesn't necessarily translate to dems nationally. Still, we could be at risk of losing all three branches of government just in time for a dictator to take over.

Anyway, here's the senate map.

Again, just a bad year for democrats. They're getting slaughtered. It's actually amazing they're doing as well as they currently are. 

With all that being said, that's my election prediction this week. CAN Biden win? Sure. Things can chance. WILL he? I kinda doubt it. He would need to significantly even the odds before election day or there would need to be significant polling error for things to realistically flip here. It can happen but right now, 10% chance. Harris could raise that to 14%, not great, but...better. But yeah, these are dark days for the democrats. Under normal circumstances I'd have schadenfreude as their malevolence and incompetence brought them to this point. They made their own bed and now they have to sleep in it, but still, given Trump is as dangerous as he is, I just feel very frustrated by the state of things, and fear what a second Trump term will look like given SCOTUS, congress, and his advisors who are all yes men would likely enable his worse tendencies and possibly cause a coup. So...we're screwed. Nothing to celebrate here. Just how I see it.