For what it's worth...

...here is my prediction of how the presidential election will go.

Of the swing states, I'm predicting that Harris wins PA, WI, AZ, NC, and NV, but loses GA and MI (also that Harris will get that one electoral vote from NE and Trump will get that one from ME). I am least secure about my prediction for NV. If I had made that map 10 minutes earlier, I might have colored that state red. But even without NV, I predict that Harris wins.

How confident am I about the above prediction? Not very. More than in prior years, this year feels more like a guess than a reasoned conclusion. The main reason for that is that I think polling is fundamentally broken right now. Now that everyone uses cell phones with caller ID and we have a well-entrenched culture of not answering phone calls from numbers we don't know, response rates are abysmally low. To compensate, pollsters have used expanded the small adjustments they used to make to assure it was a representative sample into huge adjustments. That change magnifies the effects of their assumptions to such an extent that the polls are more about the assumptions than actual data. In other word, I no longer think that polls are an accurate snapshot of candidate's support. Instead they are more about the biases and group-think within the polling industry. And yet, what else do I have to go on? I can rely on the pollsters biases or my own. I don't think either are likely to be accurate. And yet, this map is a hybrid of both of those unreliable things.

I should add, I created my first draft of this map before the Selzer poll came out showing Harris in a surprising lead in Iowa. I get that Selzer has a better track record than other polls. And I understand that in 2016 and 2020, her poll was one of the few (or maybe the only) one that detected determinative shifts in the electorate. But Selzer's polls still have the same low-response issue that all the other polls have. And two prior good calls is too low of a success rate to definitively establish her poll as more predictive than the others. The N is too low. It could just be two lucky guesses.

However, if Selzer was on to something, then the whole map would look different. Not just a blue Iowa. If Iowa is going for Harris because of a surge of women voters motivated by reproductive rights, then other red states on my map will be blue too. I would love it if Harris did win by a total blowout. That may be the only scenario that saves the country from Trump's post-election shenanigans posing a serious threat to democracy. But I'm not willing to rewrite my prediction on the basis of a single poll, even if it is the famed Selzer poll. It is still just a poll with a low response rate.