Thread by @RealCarlAllen

1/2 Since 2004, a candidate has polled at or above 49% in a battleground state and gone on to lose that state once. One time. #ElectionTwitter

Never in recent history (curious if ever?) - has a candidate polled over 50% and lost.

@FiveThirtyEight
@RealClearNews

Sidenote before pt 2 (which I think is even crazier than pt 1):

Battleground states are, by definition, supposed to be close.

That doesn't always mean Tossup - it includes many lean blue or lean red states.

HOWEVER.

This is a strong piece of data to show 49%+ is DOMINANT

2/2

Not only has a candidate only ONCE lost a state since 2004 with a polling avg of 49%+, but also

Only TWICE has a candidate with a polling avg of 49%+ underperformed that poll number:
FL12Romney 49.7 --> 49.1
NV04Bush 51 --> 50.47
@FiveThirtyEight @RealClearNews @Redistrict

Say what you will about @RealClearNews polling avgs vs @FiveThirtyEight

49% isn't just strong, it's dominant. 7 days to #Election2020 & RCP has Biden 49%+ in:

PA (49.6%)
WI (49.8%)
MI (50.5%)
538 even higher. Can Trump win without these? #ElectionTwitter
@SeanTrende @Redistrict

Candidates who poll 49%+ win

For @FiveThirtyEight's obsession with "poll error" as they wrongly define it (poll margin minus election margin) - have they done *any* research into candidate's actual poll numbers?

Seems important to understand that the "error" seems to go one way

E.g. just yesterday 538 tweeted this dumb shit:

"In 2 of the past 10 elections, there's been a 5+ point swing between polls now and final result.

* Assume a 50/50 chance such a swing would favor Trump."

Why assume something you can test @NateSilver538?

https://twitter.com/NateSilver538/status/1320792383693488128?s=19

How many times has a candidate polled at 49%+ and lost? Shit, how many times have they even finished below 49%?

@NateSilver538 has this flawed reasoning that I've been pointing out for months - that poll margin isn't the great metric he thinks. Why?
50-45 isn't the same as 45-40

You know what this data tells me?

Candidates who poll well tend to do *better* than their polls

@NateSilver538's flawed "margin" metric says if a candidate polls 50-45 & wins 55-45, "pOLLs OfF bY 5!" Then, for some reason, assumes this "error" is equally likely to go both ways.

Since 2004, candidates polling at 49%+ have OUTPERFORMED their poll number by an avg of 2.4%

In battleground states alone, they've outperformed that number 27/29 times

It gets better. Those 2 underperformances? A whopping 0.6 & 0.53

@RealClearNews @Redistrict #ElectionTwitter

So all of this leads to: can we trust the polls?

If you read them right, yes!

The question "what do polls do?" is so commonly misunderstood that even @FiveThirtyEight gets it wrong.

Seriously, don't feel bad.

Here's a thread breaking it down:

https://twitter.com/RealCarlAllen/status/1316087232806170624?s=19

I messed up the thread because I went to make some dumb joke about paranormal distribution 😂

Let's see if this fixes the thread

https://twitter.com/RealCarlAllen/status/1321343112153829376?s=19

Adding this to the thread bc people who are *really* into data might dig more.

I build shitty sports betting models for fun and only got into #ElectionTwitter because pandemic.

I am confused at the reality that @FiveThirtyEight doesn't do this math.

https://twitter.com/RealCarlAllen/status/1321366577963565058?s=19