Gary and I disagree a lot when it comes to politics but after reading his article yesterday about ‘Media Polls’ I thought, Amen.

A New York Times poll ‘predicted disaster for Biden and Democrats in big states a year from now’ – but, as Gary pointed out, two days later, Democrats swept elections in ‘big states: Ohio, Kentucky, Virginia, and Pennsylvania.’

Next, Gary asked a hard question: Why are media polls so wrong so often?

His answer was straightforward: ‘They’re done on the cheap. Good polls cost a lot. Getting them right is hard. And the media wants as many clicks as possible, at as little cost as possible.’

To take a poll in Iowa, or New Hampshire, alone can cost $40,000. A national poll costs twice that. Media polls cost less – and miss subtleties buried in reams of polling data.

Gary once told me: “In a campaign you’re like a frog on a log floating down a river watching the banks roll by telling yourself I’ve really got this log rolling – but it’s not you it’s the river.”

I worked with a pollster named Arthur Finkelstein for years – his genius was staring at poll numbers and seeing those rivers. The first time Reagan ran for President pointing to numbers Arthur showed Tom Ellis and me a river of people who would elect Reagan.

Media stories about polls talk about Biden’s down, Trump’s up – but media polls seldom mention a word about rivers.