For political junkies like me, Nate Silver's FiveThirtyEight.com web site is a treasure trove of election prediction delights. It's beautifully designed, statistically sophisticated, and a good way to learn statistical principles without actually taking a course.
In this post I'm going to share how Silver looks upon some key House, Senate, and Governor races, naturally with a special focus on Oregon races. First, though, here's some tweets that Silver shared today:
So the lean/likely zone for Nate Silver is between a 60%-95% chance of a candidate winning. And the toss-up zone is when no candidate exceeds a 60% chance of winning. Keep in mind that "chance of winning" is different from the predicted percent of the vote a candidate will receive.
With these definitions out of the way, on to the races!
I'm using the middle "Classic" 538.com forecast, rather than the "Lite" or "Deluxe" forecast (the latter includes experts' ratings, while "Classic" is billed as I'll take the polls, plus all the "fundamentals": fundraising, past voting in the state, historical trends, and more.)
Also, I made screenshots of the pop-ups on 538.com, which explains the bleed-through of what lay beneath the pop-up.
Oregon governor. Silver gives Democrat Kate Brown a 85% chance of winning, which warrants a "Likely D" rating. Yet I've seen newspaper stories calling the race a toss-up. Maybe by their definition, but this isn't how 538.com views the Oregon governor race.
Florida governor. It's interesting that Democrat Andrew Gillum has almost as high a chance of winning, 78%, as Kate Brown, which puts the Florida Governor's race also in the "Likely D" category.
UPDATE: Somehow I forgot to include the Georgia and Wisconsin governor races. Here they are:
Georgia governor. This is a really tight race. I don't understand why Kemp has a 58% chance of winning, given that his forecasted vote share is just slightly higher than Abrams'. But I'm sure Silver has some very good reasons for making this prediction.
Control of the House. Silver gives Democrats a 85% chance of winning control of the House of Representatives. That's high, but I try to keep in mind that Hillary Clinton had a 71% chance of beating Trump, and we know how that race turned out. It'll be interesting to see how close Democrats come to gaining 39 House seats, which is Silver's predicted average gain.
Schrader's House race. Democrat Kurt Schrader has exactly the same chance of holding onto his Oregon 5th House seat as the Democrats have of taking control of the House: 85%.
Walden's House race. In the Oregon House district where Greg Walden is running for re-election, his Democratic opponent has been doing a lot of traveling around the far reaches of eastern Oregon. But her chances of winning are very slim: 1%. The three other Oregon house races are all rated as solid D, with DeFazio, Blumenauer, and Bonamici all given at least a 99% chance of winning. So Schrader's race is by far the most competitive in Oregon.
Rohrabacher House race. My daughter and her family live in Dana Rohrabacher's Orange County, California House district. She's been working to defeat him, so she'll be pleased to learn that his Democratic challenger has a 64% chance of beating him. Orange County used to be solidly Republican, so this shows how the Blue Wave may become a reality.
Control of the Senate. As is well known, Democratic control of the Senate is unlikely, since many more Democratic senators are up for re-election (many in states Trump won), a result of how well Dems did back in 2012. So the 17% chance of Democrats winning control of the Senate is about the same as the 15% chance Republicans have of maintaining control of the House.
North Dakota Senate race. Listening to satellite radio as I drove around today, I heard a pundit say that Democrat Heidi Heitkamp could lose her Senate race by 20 points. But Nate Silver gives her a 32% chance of winning, and has her forecasted vote share at 48%. So this seems to be an instance of people giving a candidate essentially a zero chance of winning when actually the race is much tighter.
Nevada Senate race. Likewise, the prevailing "wisdom" (note the scare quotes) seems to be that Dean Heller is on his way to keeping his Senate seat in a state that Clinton won. Silver, though, has the race as a toss-up, since Heller has only a 59% chance of winning.
Florida Senate race. Roughly mirroring Andrew Gillum's chance of winning the Governor race, fellow Democrat Bill Nelson has a 72% chance of beating current Florida Governor Rick Scott in the Senate race. This surprises me, as I thought Scott was quite popular. But it seems a combination of Democratic enthusiasm and "red tide" worries in the state are buoying Nelson.
Texas Senate race. Beto O'Rourke is getting a lot of Democratic love nationally, but his chance of beating Ted Cruz is slim at 21%. Could happen, though. A 1 in 5 chance is a lot greater than a 1 in 100 chance, as is the case with 4 of 5 Oregon House races.
Arizona Senate race. Democrat Sinema has a 64% chance of beating Republican McSally in this race for an open Senate seat. So that's good news for Democrats, since nearly everything has to go right for Dems in Senate races for them to have a chance of taking control of the Senate.
I'll end by observing that I trust Nate Silver's methodology a lot more than I trust my progressive mind -- which dearly hopes that against all odds, somehow the Democrats take control of the Senate.
I believe that Silver is correct with his 17% chance of this happening.
But I can understand why some consider that the Democrats have a greater chance, since mere "clicking around" the 538.com Senate map can produce a different outcome (which I'd call by a non-technical term: wishful thinking.)
Here's what I mean:
I zoom around and see that in Missouri Democrat McCaskill has a 58% chance of holding onto her seat in a red state. In Indiana Democrat Donnelly has a 68% chance of winning. In West Virginia Democrat Manchin has a 89% chance of winning. In Montana Democrat Tester has a 85% chance of winning.
Then I fantasize that Democrat Heitkamp beats the odds with her 32% chance of winning, and holds onto her North Dakota seat. Plus, Democrat Sinema wins the open seat in Arizona, and Democrat Rosen beats incumbent Heller in Nevada.
Bingo! Seemingly this means Democrats pick up two Senate seats and they have a 51-49 control of the Senate.
The flaw in this is my wishful non-statistical thinking. Silver runs computer simulations. I simply run what I want to have happen in my own mind. It could happen, because Silver says there's a non-trivial 17% chance Democrats can take control of the Senate.
It's just a lot less likely than the "everything goes right for Democrats" scenario I can fashion in my left-leaning mind.