Reading the Lie of the Green (and Polls)

<![CDATA[Reading the lie of the green is a golf term that has to do with helping you increase your score by having to putt less. If you've seen any tournament on television or even in person, golfers have all sorts of mannerisms they display when sizing up a putt. Sometimes, they squat and stare. Other times, they walk in a large circle all the way around the ball and the hole to see it from all 360 degrees. You've seen golfers hold up their putter like a pendulum to help measure what slope there is, if any. In politics, especially when media gets involved, there's reading the lie of the green, too.Hugh Hewitt, for the nearly 25 years I've been producer of the Hugh Hewitt Show, has preached for media to disclose who they vote for and abandon the ridiculous veil of neutrality when it comes to how they cover politics. Most Americans are pretty smart about this, and can usually sniff out where a reporter or journalist sits on the ideological spectrum by the questions they ask, the way they frame those questions, and whether they become a debate combatant instead of an interviewer, depending upon the partisan leanings of the subject of the interview. Dana Bash all but wore a powder blue hat with a white D in a circle Sunday morning trying to wage rhetorical combat with GOP vice-presidential nominee J.D. Vance. Hugh's point is to disclose all that up front so that the viewer or reader can correct for the lie of the green. When it comes to polling, there's also a lie of the green that needs to be done, which we're going to unpack for you here. Since it is down to 50 days and change until the election, you're about to experience a tsunami of polling, both national and in battleground states, and with that, the analysis of what it all means from people that have desired outcomes for the election, whether those desired outcomes are stated or not. Real Clear Politics has been the standard-bearer when it comes to compiling all the polling data out there. And part of what makes them so good is they offer the ability to compare the track records of the pollsters as well as the ability to compare where the race is today compared to prior election cycles. This is especially useful, in this case, because Donald Trump was the Republican nominee in the last two elections of 2020 and 2016, so the comparisons, while not perfect, make it a lot easier to extrapolate where things really sit today. Let's start with the polling industry as a whole. RCP tracks 18 different polling outfits that do multi-state polling. And of those 18, 14 of them over the course of the last decade have missed the actual results in favor of the Democrat candidate. Four missed in favor of the Republican. So if 78% of the polling consistently reports polling that overstates the final margin for Democrats, how much is that average miss? 4.6%. The least amount of error on behalf of the Democratic candidate is 2.2%, and the biggest miss is 6.1%, but the sweet spot for missing is just a touch over four-and-a-half. As of Sunday, the RCP average of averages shows Kamala Harris with a 1.7% lead nationally. That's down from 1.9% before the debate, by the way. So what does that mean? Using 2020 as a baseline example, roughly 165 million Americans cast a ballot for someone for president. A lot for Trump, more for Biden, and almost 10 million for minor parties and independents. 1.7% of that 165 million is a 2.8 million vote lead nationally for Kamala Harris, giving the polling out there the benefit of the doubt that they're accurate. Now again, using 2020 data as the baseline, take a look at how California voted. It's as deep blue of a state as there is in the Union, and they voted overwhelmingly for Joe Biden. The final margin, once they finally finished counting a couple months after Biden was sworn in (I exaggerate, but not by much), was 63-42. In raw vote totals, Biden netted 5.1 million more votes just in the Golden State than Donald Trump. Going back to the RCP average as of Sunday, if Kamala allegedly has a 2.8 million lead nationally, and would have roughly the same results out of California, you could actually make the case that if you were to take California out of the equation, Donald Trump is actually leading the national vote in the remaining 49 states by a little under 3 million votes. Now we get to the poll that was released over the weekend by AtlasIntel, the most reliable pollster based upon prediction versus reality over the last two elections - 2020 and 2022. In the former, AtlasIntel was only off by 2.2%. In the case of the latter, they slipped a little bit, missing by 2.3%. AtlasIntel by far has the best track record this decade, earning 2.7 out of 3 stars at out of 538's ranking system. They just released their new national numbers, entirely surveyed post-debate. ]]>