1/ One thing I don& #39;t ask myself enough: How does Trump win?

Let& #39;s leave aside ways he could cheat. They& #39;re quite possible, but too unpredictable to contemplate here.

Every state where he& #39;s winning in the polls—TX, AK, etc.—gets him 163 EVs. Precious little, but a start.
2/ Let& #39;s say things regress to the mean a bit. Biden does horribly in the other two debates, or gets hit with a scandal, and state polls all shift 3 points toward Trump.

If polls are accurate, he gets IA, OH, GA, NC. Still only at 219 EVs.

Hmm. Now let& #39;s look at polling error.
3/ If polls shift 3 pts to Trump between now & E-Day, he needs a uniform polling error of ~3.8 points to pick up FL, AZ, NV and either WI or PA and pass 270.

But IRL, poll error isn& #39;t uniform. What& #39;s plausible?

He& #39;s only a tiny error away from AZ and FL. That gets him to 259.
4/ Now we pick 1-2 states where polling error is larger, like WI& #39;s 6.7pt error in 2016.

If polls move 3pts his way, MN/VA/CO are likely still out of reach. But 4-5pt errors in NV/WI or MI or PA get him past 270.

And that& #39;s how Trump wins via poll shift & poll error alone.
Notice, this is not an easy path. It assumes that multiple factors go Trump& #39;s way in the next few weeks.

But although it& #39;s unlikely, Trump *can* win even without any cheating whatsoever.

/fin
You can follow @OtherSideReflec.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: