So...I've been noticing that some people are getting increasingly hostile about election predictions. I know I share my predictions on several subreddits at this point because it's October and everyone else is doing it so might as well show off my work on the subject. But then I get ripped on as of late, especially by democrats, because my elections...show Trump winning. I'm told I'm using biased sources (RCP), that they're trash, I'm only cope posting because I want Trump to win, blah blah blah, so I wanna really set the record straight.
I don't want Trump to win. I can't say I'm huge on Harris, but I am left wing aligned all things considered, and while I don't really LIKE the democratic party, I do want them to win. HOWEVER, and here's the thing that these people don't understand: election predictions aren't about cheerleading for your side. They're about the data. They're about the science. It's about the process of arriving at the conclusions you arrive at. I use RCP because I've ALWAYS used RCP. I've been doing prediction since 2008 and back then RCP was the only polling aggregator site just about. 538 didn't come onto the scene until 2012, and I dont remember his work being hyped until he basically got a perfect prediction. In 2016, he got hyped again for "being the closest". Well, given the entire media industrial complex minus the conservative media sites were cheerleading for Clinton, they got it wrong. It was easy to go for clinton. The polls favored her. However, leading up to the election, the polls shifted, and by election day, I considered the race a tossup. And a tossup it was. Silver is famed for getting a 70% probability for clinton and acknowledging that Trump had a chance. Again, I had 56%. You can go back to 2016 posts and read my prediction. I literally posted it live here. In 2020, I followed Silver's lead in removing conservative pollsters accused of poll bombing the averages. it made my prediction more off than if i just let it ride. Silver made the same error, got roughly the same prediction, and let's face it, we both got clowned for it. So...in 2024...with all of the accusations of biased pollsters flooding the averages, I say, let it ride. Because I tried to remove that stuff last time, and then Trump overperformed. This time, I'm just reporting the data and what happens happens.
It's not about cheerleading your side. It's about getting it right. It's about being honest and reporting the data even if it doesn't go in your favor. Quite frankly, I'll be honest with you. America deserves better than both of these jokers. I dont like Harris, I don't like Trump. In 2016 and 2020 I felt the same about Biden and Clinton. In those elections I protest voted and went green. This time I went harris, primarily to save democracy. I don't even like Harris. She's going too far in the new dem direction and any previous praise I gave her can effectively be retracted because she's just becoming a generic new dem clone politician. I held my nose for her. Doesn't mean I want trump to win, I actually really am scared about that, but I have little to no reason to cheerlead Harris on at this point. My vote was transactional. I ain't really invested in her presidency, that's it. If anything as her numbers tank I've become more critical, as have most lefties as they recognize that her centrist pivot is landing like a wet fart.
But yeah. My predictions have always been about the data. If my side is winning. I report that. If they're not, I report that. I don't fluff the data or try to lie about who I think is gonna win. If there's anything I learned, it's that getting high on your own BS and thinking you're ahead and the data that doesn't conform with your perspective is wrong is just gonna lead to pain on election day as reality comes crashing down and we realize that the data was correct after all. And it almost always is...within reason. Not saying it's always dead on, but again, if you dont trust polls, learn what error is and understand that everything is about statistics and probabilities. And even if my data shows trump ahead, it's a REALLY REALLY close election and at 45-55, anyone can win.
Keep in mind, anything less than 60-40 is what i consider a "statistical tie". Like...flip a coin. That's basically it. anything in the 61-39 to 84-16 range is in the "lean" category. Whoever is winning has an obvious edge but it's still very possible things flip the other way due to normal polling error. 84-16 to 98-2 is "likely"...basically I'd expect the side ahead to win decisively, although as we saw in 2020....being that far up is no guarantee, especially if you use questionable methods to get that conclusion, like removing polls affiliated with the political party you don't like. And anything 98-2 or better is basically...safe. Like statistically, don't expect it to flip. it probably won't.
But yeah. That's basically it. And right now at 45-55, it's like who knows. I give trump a slight edge, and I do believe the momentum is on his side going into these final weeks of the campaign, but it's not a statistically significant one by any margin. It's virtually not an edge at all. A single point of polling error in either direction is likely to decide the election either way. I mean, those probabilities mean something. And when it's 45-55, that means it literally can go either way at a near 50-50 level. It means if i run my simulator, in theory, i should get 11 trump outcomes and 9 harris ones for every 20 simulations i run (currently it's closer to 10-10 as it seems to slightly favor harris for some reason, still haven't figured that out). And yeah. So...please, people getting wrapped up in the election prediction stuff where they can't keep their partisanship out of it, take a chill pill and remove that stick from you know where. There are places where you can cheerlead for your own side. Prediction subs and forums shouldn't be one of them. Let's keep those academic and based more about the science and not about the feels. My predictions aren't meant to hurt liberals' feelings. They're just data and math that I applied to said data that give me probabilities. This isn't even my first rodeo. This is the 5th election i've actively tried to predict in my life. I got 3/4 so far right, and the 4th one is the one no one got right, so...yeah.
No comments:
Post a Comment