The Economist: Forecasting the US elections

[The Other Castro]

Weird that 538 still has Emerson rated A-.

[GeorgiaModerate]

https://twitter.com/gelliottmorris/status/1290857816786382849

[wbrocks67]

https://twitter.com/gelliottmorris/status/1290857816786382849

See. This is what I'm talking about. I'm all for making things better and more accurate, but it's funny how every time there is a "change" among these people, it comes after some type of criticism, and somehow comes at the cost of Biden/Ds? 2016 really did a number on these people. They refuse to believe Biden could actually be up as much as the fundamentals and polls say he is.

[wbrocks67]

Weird that 538 still has Emerson rated A-.

yeah, that caught my eye the other day too

[Del Tachi]

I'm curious what the "uncertainty" is in this race. Other than candidates dying, I really don't see what's so uncertain here. This election cycle has been incredibly consistent.

That we're still 89 days away from Election Day?  Polls in August are generally no more predictive those from June; polls only become greatly predictive in the last 45 days or so

[Sadader]

Silver’s complete lack of respect for Morris is pathetic. Especially given the hints we’ve gotten at Nate’s model, e.g.

https://mobile.twitter.com/conorsen/status/1287903534739054596

[ProgressiveModerate]

At this point, all these forecasts are jokes. 538 has taken way too long, and the Economist's seems way too confident in their model and uses a lot of baseless speculation in their model (i.e. trying to bump up Trump's chances despite Biden leading him by 8-10 points nationally)

[DaleCooper]

I'm curious what the "uncertainty" is in this race. Other than candidates dying, I really don't see what's so uncertain here. This election cycle has been incredibly consistent.

That we're still 89 days away from Election Day?  Polls in August are generally no more predictive those from June; polls only become greatly predictive in the last 45 days or so

Then why even have the model? Isn't the purpose of the model to forecast the election as it stands now based on the data that's available?

[GeorgiaModerate]

The new model is up (same URL: https://projects.economist.com/us-2020-forecast/president).

Biden's probability of winning the EV went from 91% to 88%, and of winning the PV from 99% to 97%.  But his median EV went from 347 to 351.  Also, Texas has now moved into their Tossup category.  Basically, his position looks slightly stronger in the tweaked model but with the percentages somewhat lower to account for uncertainty (which I think is fair, given the amount of time until the election).

ETA: more details:

We fixed a series of medium-sized bugs in the model that collectively caused too little uncertainty in our forecast. The new model better represents the lower confidence we have both in our fundamentals-based “prior” prediction, and in the polls when we are further out from election day. Statistically, the change is three-fold. First, we increased the band of uncertainty around the prior for every point in the election year, creating a minimum standard deviation of three percentage points on the Democratic vote share nationally. The previous model had a minimum closer to one-and-a-half points. Second, we modestly increased the amount of movement our model allows in the polls across the whole cycle. The average national vote share now moves with a standard deviation of five percentage points over the course of the entire election—up from about 4.5 in the previous model. Finally, we performed various minor tweaks to how we judge the correlations between the states, how we update state-level trends with national data and how we calculate the bias in pollsters that don’t weight their data to match the partisan composition of the electorate. In sum, these changes caused Joe Biden’s chances of winning the popular vote and electoral college to each decrease by roughly three percentage points.

[Del Tachi]

I'm curious what the "uncertainty" is in this race. Other than candidates dying, I really don't see what's so uncertain here. This election cycle has been incredibly consistent.

That we're still 89 days away from Election Day?  Polls in August are generally no more predictive those from June; polls only become greatly predictive in the last 45 days or so

Then why even have the model? Isn't the purpose of the model to forecast the election as it stands now based on the data that's available?

Every model represents a range of possible outcomes based on what went it to it (i.e., polls, fundamentals, etc.)  If you understand forecasting as a range of outcomes rather than a single prediction, then you understand why uncertainty exists in these models.   

[wbrocks67]

Is there a reason why Trump gains 1% in the "election day" vote compared to the current day vote?

[GeorgiaModerate]

Is there a reason why Trump gains 1% in the "election day" vote compared to the current day vote?

Their model assumes that day-to-day changes will tend toward the fundamentals-based prior, as opposed to a random walk.  That is, Biden is currently polling slightly better than the prior would suggest, so the model thinks it will revert back toward it.

[ProgressiveModerate]

Basically, there's just more uncertainty in the model. That's definately a welcomed change

[Beefalow and the Consumer]

The new model is up (same URL: https://projects.economist.com/us-2020-forecast/president).

Biden's probability of winning the EV went from 91% to 88%, and of winning the PV from 99% to 97%.  But his median EV went from 347 to 351.  Also, Texas has now moved into their Tossup category.  Basically, his position looks slightly stronger in the tweaked model but with the percentages somewhat lower to account for uncertainty (which I think is fair, given the amount of time until the election).

ETA: more details:

We fixed a series of medium-sized bugs in the model that collectively caused too little uncertainty in our forecast. The new model better represents the lower confidence we have both in our fundamentals-based “prior” prediction, and in the polls when we are further out from election day. Statistically, the change is three-fold. First, we increased the band of uncertainty around the prior for every point in the election year, creating a minimum standard deviation of three percentage points on the Democratic vote share nationally. The previous model had a minimum closer to one-and-a-half points. Second, we modestly increased the amount of movement our model allows in the polls across the whole cycle. The average national vote share now moves with a standard deviation of five percentage points over the course of the entire election—up from about 4.5 in the previous model. Finally, we performed various minor tweaks to how we judge the correlations between the states, how we update state-level trends with national data and how we calculate the bias in pollsters that don’t weight their data to match the partisan composition of the electorate. In sum, these changes caused Joe Biden’s chances of winning the popular vote and electoral college to each decrease by roughly three percentage points.

inb4 "they just want to make the race look more exciting"

[Gass3268]

Even with the updates, it’s back to 91%:

https://twitter.com/gelliottmorris/status/1291710743642025984?s=21

[Amenhotep Bakari-Sellers]

COVID 19 expanded in CA, FL, TX and AZ. McSally was doing better, she cut it down to 2 to 4 points, but Kelly got back his lead

[wbrocks67]

Yeah, it's bonkers to me how we can get Trump +5 in SC, Trump +7 in KY, and Trump +7 in KS results and pundits like the Nates and Harry really can even *try* and say "oh this is tightening!!!"

[ultraviolet]

Nate Silver is making himself look more and more like a jerk. If you’re not going to do your job at least let others do theirs in peace.

[ProgressiveModerate]

Nate Silver is making himself look more and more like a jerk. If you’re not going to do your job at least let others do theirs in peace.

Silver isn't entitled to make a model for anyone technically, but it is disappointing that it hasn't been released yet since we're so used to it. As a top pundit an expert in this area, he has the right to criticize the Economist's forecast the same way any of us can; I personally think it has too little uncertainty and relies too heavily on polling personally (I.E. PA being to the right of WI, UT being >99.9%" R), but overall you do have to give them credit for making the forecasts in the first place.

[Ferguson97]

Nate Silver is making himself look more and more like a jerk. If you’re not going to do your job at least let others do theirs in peace.

Silver isn't entitled to make a model for anyone technically, but it is disappointing that it hasn't been released yet since we're so used to it. As a top pundit an expert in this area, he has the right to criticize the Economist's forecast the same way any of us can; I personally think it has too little uncertainty and relies too heavily on polling personally (I.E. PA being to the right of WI, UT being >99.9%" R), but overall you do have to give them credit for making the forecasts in the first place.

I don’t understand this “he’s not entitled to make us one”, he not only promised it but it’s also literally his job

[GeorgiaModerate]

Elliott finally responds to Nate Silver.  Worth reading the whole thread.

https://twitter.com/gelliottmorris/status/1293243570149888005

[GeorgiaModerate]

https://twitter.com/gelliottmorris/status/1294028137026789378

[GeorgiaModerate]

https://twitter.com/gelliottmorris/status/1295019699609251842

[new_patomic]

https://twitter.com/NateSilver538/status/1296490992166752257?s=20

The world's most statistically petty drama may finally be coming to a close.

[Penn_Quaker_Girl]

https://twitter.com/NateSilver538/status/1296490992166752257?s=20

The world's most statistically petty drama may finally be coming to a close.

Live view of Nate Silver's compound: