Hi, I'm Dan Pramak and welcome to Axios Recap, presented by Facebook.

Today is Monday, October 26th.

Stocks are way down, COVID infections are way up, and we're focused on the art and business

of political polling. We are now just eight days away from the election, and it's not just the candidates whose

futures will be determined by the results.

It also could be make or break for political pollsters, four years after wrongly predicting

a Hillary Clinton presidency.

Here's how veteran Republican pollster Frank Luntz summed it up recently to Fox News.

If Donald Trump surprises people, if Joe Biden's had a five or six point lead, and then Donald

Trump wins, my profession is done.

Two things to know. First, the pollsters' national miss on Hillary Clinton wasn't as big

as a lot of us remember it being.

As 538, editor-in-chief Nate Silver later noted, the final national result was within the

normal rate of accuracy.

The trouble was really more in certain state polling, particularly some of those so-called

blue wall states, like Michigan and Wisconsin, Pennsylvania, where the

mistakes were more pronounced and more consequential.

Two, many pollsters believe they learn their lessons from 2016, so have reworked their

modeling accordingly, particularly things like re-waiting on educational attainment.

In other words, they believe that the Biden leads

they're projecting now are more likely to hold. The bottom line is that Biden is the frontrunner

because the polls say he's the frontrunner, and the pollsters may now need Biden to win

...