have fun
2 Likes
I don’t understand polling nearly as much as many of you. Is there any reason an 87/13 polling prediction is any different than an 87/13 all in preflop?
You are using the word fun pretty fast and loose! 
I think this is a borderline philosophical question. Most would say no, but i suspect the answer is yes.
2 Likes
Biden wins FL and he’s 99% to win.
Biden wins PA and he’s 97% to win.
If Trump wins both he’s 93% to win.
2 Likes
If you believe that it really is 87/13, then no there is no reason to assume there is any difference.
1 Like
I don’t think we needed 538 to tell us that.
When you’re all in preflop, you can’t call on the justice department to stop the dealing after you hit a lucky flop.
14 Likes
I don’t think it’s as clear as that. If the 18% that have voted are made up of people that have been decided for months then I don’t think Trump needs an extra shift from that poll.
So the knocks at 538 aren’t that the 13% might hit but that the real math should have been closer?
Doesn’t the 13% assume there is built in error in the polls?
There’s no difference in terms of how often the outcomes occur. But there is a difference in the type of uncertainty.
All-in preflop is randomness uncertainty.
Polling error is information uncertainty. If you lost you didn’t get ‘unlucky’. You were always losing, but you just didn’t know it.
16 Likes
538 tracker back to 88-12. Thank fucking god. I have never been so upset about a half point of equity before.
7 Likes
New IBD national pole: Biden + 3
3 Likes
Here’s the early voting so far in NC compared to 2016. Not showing any signs of slowing down so far.
https://twitter.com/OldNorthStPol/status/1318879834811105281?s=19
I look at it as with all-in preflop we know (with almost 100% certainty) that if we deal it an infinite amount of times we’ll get the X/Y results approximately as often as we should.
But here there’s an extra element of uncertainty around the 83/13 itself in the 538 sense. If we ran this election an infinite number of times (Dear God…) we’d almost certainly find that the probability at this point isn’t actually 83/13.
4 Likes
New YouGov national pole: Biden + 7
1 Like
1.912 million votes in Georgia already. Almost half the total 2016 vote.
2 Likes
New USC national pole: Biden +11/ Biden +13
7 Likes
Right that makes sense.
That goes to my question about the model. Is the 13% just a representation of everything going Trump’s way or does it include uncertainty in the model itself?
I think that’s a good way to put it.
Maybe a better ‘all-in’ pre-flop analogy would be – You get all-in preflop as an 87/13 favorite. But now I remove all the cards in the deck except 8. If you were betting on the outcome, it’s still 87/13 because we have no information about what cards we removed. In reality, if we ran the board with just these 8 cards over and over we’d understand that the actual probabilities are much different than 87/13 and actually very little randomness.
9 Likes