Are the polls wrong?

 So, I regularly get into arguments with people who whenever I bring up my analysis of the 2024 election and Biden's chances, they think I'm wrong. They say the polls arent everything, the polls have been off before, and they will be off again, and that they can't predict things this early. 

I mean, polls are snapshots in time. I admit things can change, the real question is will they? And can we predict these changes? And the answer to that second question is largely no. People think the polls are magically gonna shift in Biden's favor like the 2022 mid terms did and that remains to be seen. I dont think that we should count on that. I do admit such a shift is possible, but we have no idea what it will look like or how big it will be.

Some people also think trump will definitely be convicted, this remains to be seen. he could be, but once again, wouldn't count on it.

My basic approach is to go by the numbers based on what they are right now, and react to changes when they happen. if things even out by november, then good. I hope they do, I hope these early predictions are wrong, but ive been following elections like this since 2008, and the only election that had a surprise upset was 2016. 

Speaking of which, are polls WRONG?

Yes and no. Polls come with a certain amount of error. Ideally 3 points is the margin of error, and given both candidates could be 3 points off, we could be up to 6 points off within a level of confidence of 95%.

I tend to expand this to a 4 point margin of error, hence my 8 point buffer. I expect polls to be 95% accurate within 8 points. I do this because only the best polls are within 3 points. Many weaker polls have 4 or even 5 point margins of error. And while averages should minimize that, I think given the regular 6-8 point swings we've seen out of the rust belt in 2016 and 2020, my instincts to go with 4 points margin of error works out. 

All in all, I would expect my results to be 95% accurate, or 97.5% accurate in a one tailed scenario like this (given the error on the other side of the bell curve just leads to the same candidate winning), within 8 points or so.

Given how we're closing in on a 88-91% chance of Biden losing, that's not good. I mean, this is roughly what happened in 2008. I remember following that and being like "Jesus, Obama literally has like a 95% chance", and that never changed. 2012, it was a little narrower but Obama comfortably won. 2016, I mean, the trend is very much on this blog. I started out saying Trump was screwed and like 6 points down and he DID manage to narrow it, but that was the only case we actually had an upset. And that seemed to be a perfect storm of stuff all coming together at once to JUST let trump NARROWLY win. As I said back in 2016, by election day, it was almost 50-50. CLinton had a 56% chance, 272-266 was the most likely outcome. i didnt predict the rust belt shift, but quite frankly, no one did, and nothing about it was TOO unreasonable given the room for error any prediction has. 

All I can do is follow the data. From there, we have a range of scenarios that can happen. Heck I've run ranges of scenarios with random number generators before. Back when Biden had a roughly 30% chance of winning by my prediction, he won 10/40 of the outcomes. More extreme results lead to fewer breakthrough outcomes, and we're in a position where despite a 12% chance of Biden winning, I wouldn't expect him to win many, if any outcomes by random number generator. Of course in reality, trends are more accurate than each state being an independent trial, but still, you get the idea. 

I will follow the data until the data changes, then I will follow that data. Right now the data tells me Biden is toast. It can change, but barring systemic polling bias or some earth shattering event shifting the race a good 6-8 points like a Trump conviction or someone dying, I'm not seeing massive changes here. And keep in mind those changes could flip either way. A trump conviction could shift the race 7 points toward Biden with a return to his 2020 baseline. On the flip side, Biden having a heart attack and dropping dead could lead to a 4-5 point shift as the party scrambles to nominate a less popular replacement like Kamala Harris. That would all but guarantee a trump win by blowout margins. Anything can happen. But that doesnt mean anything will. I trust the data that exists, until that data changes. Because that's all I can reasonably do. 

Anyone who makes a prediction NOT based on current data is just talking out you know where, that's how I see it. They're engaging in wishful thinking based on their own biases and proclivities, not based on data. I see people all the time who still think Biden is gonna win in a blowout. I tell them to look at polls and then they get huffy because i dared break their illusion. Well...sorry, but reality is reality whether you like it or not. Do you think i LIKE trump winning? Hell no. I DEFINITELY do not want this. But...I see the writing on the wall, I see the trends, I see the data, and I'm like "yeah unless something changes this IS gonna happen, we should be worried and try to address this". 

But it seems like, once again, bluemaga just wants to bury their heads in the sand, just like in 2016. Let them learn the hard way. I'll be happy to say i told them so.