Saturday, October 08, 2016

election projection

There are a lot of forecasts for this presidential election. There's 538 (currently showing an 81.6% chance of a Clinton victory), the Upshot (83% for Clinton), the DailyKos (90% for Clinton), the Huffington Post (84.7% for Clinton), Predict Wise (87% for Clinton), the Princeton Election Consortium (93% for Clinton), the Cook Political Report (lean towards Clinton), Rothenberg & Gonzales (leans towards Clinton), and Sabato's Crystal Ball (Clinton victory projected). They all use different methodologies, but they are all analyzing polling data, and sometimes factoring in other "fundamentals" that have a record of predicting the outcome of presidential elections. Some are sponsored by organizations that have a partisan slant (Kos is overly on the left, Huffington Post is viewed as a site that leans left, Sabato has a reputation of having a slight conservative slant) and some don't seem to. There are a lot of places if you want to find a mathematical model that tries to predict the outcome of next month's presidential election. And all of them, every single one, says that a Clinton victory is the likely outcome, and all the models that try to calculate a percentage of victory are predicting very high odds of her victory.

And then there's Scott Adams, yes the guy who makes Dilbert. He's giving a 98% chance of a Trump victory. What polls is he factoring in and how is he counting the electoral votes to reach that conclusion? He's not. He has just pulled the 98% number out of his ass. He's not "unskewing" the polls, he is ignoring the polls altogether and giving his own gut feeling. Maybe Adams is on to something.

Maybe all of the polls are completely off base, or there is some other way of Trump reaching 270 electoral votes (or I guess just 269, with the GOP in control of the House that is all he would need) that only Adams is privy to. But I suspect not. The models could all be wrong. But to convince anyone but the wingnuts, he needs to give me some reason that isn't just an argument about what he imagines will appeal to the American voter. He needs to explain a plausible path for Trump to reach 269.