Preparing to Evaluate the Predictions
Making note of some of the predictions and such as we approach November 6th.
As we approach the end of the 2012 cycle (and the start of the 2016 cycle*), it is time to start cataloging the predictions so we can do an assessment after we get the results. This strikes me as an especially useful exercise this year since there has been so much contention on the subject of poll analysis and using said analysis to forecast outcomes.
First we have Dick Morris, who sees a building Romney wave in a column at The Hill:
In the next few days, the battle will move to Pennsylvania (20), Michigan (15), Wisconsin (10) and Minnesota (16). Ahead in Pennsylvania, tied in Michigan and Wisconsin, and slightly behind in Minnesota, these new swing states look to be the battleground.
Or will the Romney momentum grow and wash into formerly safe Democratic territory in New Jersey and Oregon?
Once everyone discovers that the emperor has no clothes (or that Obama has no argument after the negative ads stopped working), the vote shift could be of historic proportions.
Morris does, however, avoid actually making a prediction in the column, he just intimates that Romney might do the impossible or, at least, the highly improbably (like win Oregon). On his web site back in August, however, we get:
If he carries all three of these states and also wins all the others where Obama is now at 50% or less – Iowa, New Mexico, Michigan, Minnesota, Pennsylvania, and New Jersey — he will get 351 electoral votes, a landslide about equal to Obama’s 363 vote tally in 2008.
The strong probability is that Romney does, in fact, carry Florida, Ohio, and Virginia and a share of the other states where Obama is below 50% of the vote.
Of course, he hedges with a rather big “if.” The column is entitled, however “Here Comes The Landslide” on his web site, meaning that it was not just an editor at The Hill who chose it.
Back to the column in regards to the Senate he does provide a more resolute prediction:
The most likely outcome? Eight GOP takeaways and two giveaways for a net gain of six. A 53-47 Senate, just like we have now, only opposite.
Meanwhile, Karl Rove predicts:
Mitt Romney will be declared America’s 45th president. Let’s call it 51%-48%, with Mr. Romney carrying at least 279 Electoral College votes, probably more.
Brendan Nyhan provides the following graph of the stats types’/betting markets’ and probabilities:
For a discussion of probabilities versus polling, see James Joyner’s post from earlier today.
h/t’s to Taegan Goddard and Andrew Sullivan.
*I am only partly joking. As soon as we have a firmish idea about a winner on election night, if not in some of the dead space been results, someone (indeed, numerous someones) will be speculating about 2016.
I’m not sure there are many testable predictions out there. Non-absolute odds cannot be refuted by a single run. Few people are making absolute calls. Most preface their predictions with “likely to” and so on.
An absolute state-by-state list of electoral victories would be testable, but most people making those only expect to come close.
@john personna: Fair enough–“test” really isn’t the best word for this post.
If I were a betting man, which obviously I am, I would predict things accordingly:
EC = Romney. The McCain states from ’08 (obviously) + IN and NC (obviously) + Fla., OH and VA + CO. I don’t think any of those states will fall within recount margins or at least not cognitively sound recount margins. I think Romney also will eke out a very narrow win in NV.
IA is an interesting state. I suspect the margin there will be very close, a la ’04. I don’t think it’ll matter. I try not to pay attention to WI, given that if a state can vote both for Carter and Dukakis then as far as I’m concerned it should be stripped of its voting rights. New Hampshire bores me. I’ve been to New Hampshire.
For the popular vote I’d say 52-47 for Romney. That margin is well within the ballpark of politico-economic history and realities on the ground on Main Street. If we had a real media, instead of the 24/7 national Democrat cheerleading squad with which we’re saddled, we’d be looking at a 1980-style wipeout.
The overall result (Obama wins or Romney wins, and the percentage of the votes they capture) can’t really be construed as a test, no. Depending on how much data can be pulled out of exit polling, one might be able to test some of the underlying assumptions made by the pollsters and modelers (e.g. basic demographics of those who voted, party ID, etc).
But then exit polling has its problems.
@Rob in CT: I have changed the title to “evaluate” as that is more what I am going for here. Indeed, I used the verb “assess” in the post itself.
“As soon as we have a firmish idea”
Jesus, Steven.
@sam: ?
Meaning that as soon as they (the folks on TV) start to think that they have a sense of a winner, even if one is not yet declared, the speculation will start.
Tsar needs a much broader set of friends and acquaintances.
From this article, I think we can conclude that Dick Morris is still an ignorant gasbag with zero credibility. Some things never change.
Dick Morris might believe in Fate.
I am feeling better already, cause Karl has the math, and we all know how it works.
Let’s just try Dick Morris’s and Karl Rove’s predictive powers by ordeal. Now. Just like back in the days of our nation’s founder’s fathers. There’s no need to wait for election day to do this.
@rudderpedals: I just think someone needs to get a restraining order against Karl Rove on behalf of all whiteboards. Every day he’s basically scrawling “I am STUPID” across their poor faces, and they’re powerless to stop his bullying.
@Steven,
While Morris did not make a prediction in that column, he did make one last night on The O’Rielly Factor Morris predicted that Romney in a 5 to 10 point landslide in the popular vote and 300+ in the electoral college.
http://freedomslighthouse.net/2012/11/01/political-consultant-dick-morris-continues-to-predict-a-mitt-romney-landslide-victory-with-more-than-300-electoral-votes-video-103112/
Many here have been predicting Obama to win. So tell us what your predictions on the Electoral College numbers and senate races are and how confident you are in your prediction.
@Wayne: I will be posting a prediction soon.
My prediction is Romney with 292 and senate at 51 Republicans. My confidence level is so so. Understand that my confidence level do tend to be less on average than most but my fairly confident level tend to be right more often than most peoples’ almost certain confident levels.
The projections of Dick Morris and Karl Rove don’t come close to any prediction that I would deem worth “cataloging”. I think everybody knows where those two will stand regardless of the candidates up for election. Their projections are neither significant nor relevant, as their opinions are solely the result of their own political preferences rather than actual polling fact.
@Tsar Nicholas:
You can’t help yourself, can you? Why can’t you write an otherwise reasonable post (which yours was) without tacking on that kind of tinfoil Redstate crap?
Sorry, I didn’t notice this, but I will pretend that it is snark.
I think it will be a tight PV, with Obama winning by a small margin. Electorally, my crazy hunch is that Obama will win Ohio, Virginia, Wisconsin, Iowa, and Colorado. Florida will be super close but will go Romney. Ohio will also be close, but by 11 p.m., Obama will be the projected president.
Never attempt to predict the rank stupidity of the American electorate.
But my anxiety says…
Romney wins 286 electoral votes.
The Senate stays pretty much as is.
Democratic Senators…not being as cynical and self-interested and treasonous as Republicans of the past 4 years…roll over on pretty much everything Romney wants.
The Debt explodes to unforseen levels due to a mathematically impossible economic plan.
Roe v. Wade goes away with the appointment of two litmus tested appointees.
Romney outsources foreign policy to Israel…we attack Iran at their behest…and get involved in another decade or so of un-winnable war.
It’s going to be a lot like Bush/Cheney/Rove all over again…only with a more activist SCOTUS.
Elections have consequences…VOTE.
Morris is the court jester of election punditry. He couldn’t even get the Electoral Vote totals correct for Minnesota (10 not 16) or Michigan (16 not 15).
Plus, Romney ahead in Pennsylvania and tied in Michigan?
Oy. Why does anyone treat this guy as anything other than the toe-sucking freakshow he is?
As far as predix go,
Obama wins with 281 EVs (Ohio plus Nevada, Wisconsin, New Hampshire and Iowa), though CO and VA could also fall into his column. with a slight tailwind. FL would be an “everything goes right” eventuality.
Popular vote margin, somewhere between Gore’s 2000 margin and Bush’s 2004 margin for Obama.
Not too different than the flurry of press releases from McCain in 2008 about the last minute surge that was sweeping him towards the White House. “We are going to win in PA” was a big element of that talk track.
Obama 303 / Romney 235, fairly similar map to 2008. Dems make small gains in both the Senate and House.
I am pretty certain that I won’t believe anything Dick Morris says.
I am not particularly convinced the GOP will take the senate, although they will pick up seats.
I think the house will be solidly in GOP control.
As for presidency-I still think Obama has the electoral advantage here. I think Romney does tons better than McCain and depending on voter turn out (and I think more than anything the real question is just who shows up on Tuesday to vote) may pull off a win, but I still think Obama has the advantage.
I
My predictions:
Obama 332, Romney 206
Popular Vote: Obama 51, Romney 48.
Senate 50 D, 48 R, 2 I who causcus with D (R gains Mont., Neb., ND, D gain Mass. from R, Conn. from I, I gains Maine from R).
House 220 R, 215 D
Dick Morris predicted a big win for McCain. He is always wrong, and often astoundingly so. He gets paid to sing sweet songs to wingnuts.
Rove’s prediction on the eve of the 2008 election was quite accurate, but far more detailed than what he is giving now. His entire prediction rests on Romney winning Ohio because Romney’s campaign manager said he’s going to win Ohio. We’ll see if he predicts something a bit more detailed on Sunday or Monday.
@Geek, Esq.:
Pretty much my prediction as well as far as the presidency… I’m going with the current RCP no toss-up map: 281 to 257 for Obama (giving Romney Florida, Virginia, and Colorado). Obama might get one of those three states, but given recent polling cycles and looking at the electoral history of those states, chances are low he’ll get two, let alone all three of them.
I expect Obama to squeak out a popular vote win, with less than a 2% gap. Possibly below 1% due to the effect that Sandy has on voter turnout.
My predix:
The sun will rise on Weds morning no matter who wins. There will be snow somewhere. Sunshine somewhere else. And rain in still another place. It will be colder this winter than it was this past summer.
Also, there will be absolutely no one who has a clue as to just exactly what Romney is for and what he is against. Except for money, he is for as much as he can stuff into his pockets. That we know.
I don’t think I’d do better than the pros. They lean toward an Obama win. If Romney wins it will be an upset (or Fate).
Alas, Steven, we can evaluate the predictions, but sadly there will never be any negative consequences for being wrong. Dick Morris still gets paid to write columns and you give precious blog space to quote those columns, even though that man has been spectacularly wrong with most of his predictions for the last ten years.
If the US were a true meritocracy, Dick Morris would be pumping gas in New Jersey somewhere today.
I think we’ll see a squeaker in the popular vote with a slightly more solid (but still fairly close) EV win for Obama. Dems hold the senate, GOP holds the House. Basically status quo. Confidence level jittery, because of course there’s still a perfectly reasonable chance of a Romney win + GOP senate control. They’re the underdogs in those categories, but not by a lot. Therefore, I will of course worry.
@James in LA: “Tsar needs a much broader set of friends and acquaintances. ”
No, and this is part of our problem. I could host a champagne and caviar party for every pundit who said ‘I’m wrong’ (and didn’t weasel out of it) without breaking my budget.
The same pundits who’ve been getting it wrong for thirty years will continue until the Big Editor 86’s them, and the internet commentators who were oh-so-triumphant will continue to talk as though they had the right to hold their heads up in public.
@Scott F.:
Nah, he would only be able to find the gas tank about a third of the time.
Let’s face it. Popular vote is at most a consolation prize in our system. Electoral is all.
In his own way, Dick Morris provides as much solace to liberals as Nate Silver does. He’s been wrong so many times that whenever he makes a prediction, we practically expect the opposite to happen, and therefore his consistent bearishness on Obama’s chances itself almost seems like evidence of Obama’s impending victory.
Back in 2011, when Obama’s reelection prospects looked bleaker than they do now, I remember feeling a little invigorated by Morris’s article “No way Obama wins in 2012.” (It’s worth noting that Morris’s other prediction in that article–that 2012 would see runaway inflation–has already been proven false.) I commented at the time that the article constituted the strongest evidence I’d seen that Obama would win in 2012.
Maybe Morris is some kind of weird genius who intuits the future by getting it exactly backwards–the Dyslexic Prophet. At least that’s what I hope.
My prediction: Obama wins the popular vote 50.5-48 (because why not use decimals in a prediction?), and wins 390 electoral votes.
Senate will be 51 Democrats plus 2 independents caucusing with them.
Republicans will hold the house, losing only a handful of seats.
@Kylopod:
That sounds like an interesting premise for a fantasy novel.
@john personna: “An absolute state-by-state list of electoral victories would be testable, but most people making those only expect to come close. ”
IIRC, Nate Silver and others are doing state-by-state predictions, and probably finer ones. With congressional districts, there’d be sufficient sample sizes to evaluate prediction probabilities.