Here’s what I’m seeing at 538 (updated at 8:06am):
[ul]
[li]Arizona[/li]Romney: 84%
Santorum: 16%
Gingrich: 0%
[li]Michigan[/li]Romney: 51%
Santorum: 49%
Gingrich: 0%
[li]Georgia[/li]Romney: 34%
Santorum: 32%
Gingrich: 32%
[/ul]
Mitt’s ad barrage must be paying off.
Well, Mitt’s ad barrage and publicity on what Santorum actually, you know, says and believes.
What puzzles me about FiveThirtyEight’s current projections is that the projected vote percentages don’t match the projected likely winners. For instance, for Michigan, he projects Santorum to get 35.5% of the vote to Romney’s 35.4%, but gives Romney a 51% chance of victory to Santorum’s 49%. And in Georgia, he projects Gingrich with 27.2% of the vote to Romney’s 27.1%, but gives Romney a 34% chance of victory to Gingrich’s 32% chance. In other words, in both states, he’s saying that the most likely outcome is one which puts Romney in second place, but that Romney has a slightly better chance of winning than his nearest competitor.
Now, statistically, I know there are ways that this could happen, if the distributions of possible totals are complicated enough, but I don’t think he’s given any indication that his primary models are sufficiently complicated to justify that. I think his model just uses Gaussian errors on the most-likely values, and the model doesn’t really include enough data (nothing but polling numbers) to construct a very complicated error model. It’d be nice to see a post from him explaining what’s going on there.
Lots of info here including fancy words like logistic regression and quantile regression.
From what I can grok the two outputs (win probability and projected vote totals) are pretty much independently calculated from the data.
Updated at 1:20 AM:
[ul]
[li]Arizona[/li]Romney: 39.7%
Santorum: 31.9%
Gingrich: 17.4%
[li]Michigan[/li]Romney: 36%
Santorum: 37.4% <-------The Amazing Shrinking Man
Paul: 12.2%
[li]Georgia[/li]Romney: 27.1%
Santorum: 26.8%
Gingrich: 27.2% <------What!!!
[li]Ohio[/li]Romney: 26.5%
Santorum: 45.1%
Gingrich: 14.8%
[li]Oklahoma[/li]Romney: 26.9%
Santorum: 37.3%
Gingrich: 23.3%[/ul]
Shifting sand…
‘Buddy’ Roemer announces end to GOP presidential campaign
[/scrapes off bumper sticker/]
It all ends tomorrow, 23 Feb, according to the release.
Perhaps he will release all his delagates to Ginggrinch?
Gary Johson is still hanging in there - oh yeah, he’s a Libertarian now - er -again - um - I think.
Crane
What delegates? Am I being whooshed?
Yes, I had to scrape off the Gary Johnson bumper sticker too!
538 has a big lead for Romney in Arizona and Virginia, big leads for Santorum in Oklahoma and Ohio and basically a tie in Michigan and Georgia.
With his money and ground operation you would back Romney to squeak ahead in the two states where he is tied (though I do remain a bit skeptical he will win Georgia). That would him give him wins in Arizona and Michigan with huge momentum for Super Tuesdsay. Ohio hasn’t been heavily polled and I am sure Romney will invest a lot of money and attention there so he could possibly turn things around there also. If he wins Arizona, Michigan, Georgia, Ohio and Virginia, the race is probably over after Super Tuesday.
However if Santorum manges to hang on in Ohio, win Oklahoma and run close in Michigan and Georgia, he will still have a fighting shot especially if Newt gets knocked out after Super Tuesday. And it’s not inconceivable that Santorum could pull off Michigan and Georgia which would almost make him the front runner. Michigan is huge because a Santorum win there would likely cause enough social conservatives to bolt from Newt and give him Georgia as well.
That is why the last debate was so huge but it appears that Santorum couldn’t do the job. He probably needs Romney to blunder in the next few days to stop his momentum in Michigan.
Romney has established a clear lead in Michigan though he has fallen back in Georgia. Santorum’s best hope probably is to come a strong second in Michigan and persuade enough Gingrich voters to switch to him as the only candidate who can stop Romney. If he can win Georgia, Ohio and Oklahoma and knock Newt out he is still in decent shape. However Romney has a knack of grinding out wins from behind in big states like Florida and now probably Michigan and he will certainly try again in Ohio.
Romney wins by saturating the problem with money. If need be he can do it with his own pocket change. I do not believe that will gain him electability.
Crane
Newt has opened up a reasonable lead in Georgia so he should survive at least for some time. Meanwhile Ohio has narrowed with a 7 point Santorum lead in the latest poll. This is the most important race on Super Tuesday and hopefully we will get regular polls from now on. While Romney has recovered in Michigan, his momentum has stalled and he is still only up by just two points. It’s not inconceivable that Santorum could pull off an upset though probably Romney’s ground game will win the day. Of course in terms of delegates a narrow win in Michigan won’t mean much.
Michigan is now so close that it looks reasonably likely that Democratic cross-over voters could give Santorum the win. I don’t know whether to laugh or cry. Probably laugh.
Apparently even Santorum is robocallingDemocrats to vote for him. I am not sure whether GOP primary voters will like this particularly if he ekes out a narrow win. Still Michigan is going right to the wire and clearly each campaign will strain for every advantage. A narrow popular win may not matter that much for the delegate breakup which as per a Nate Silver postis quite complicated but it’s probably going to matter a lot for momentum for Super Tuesday.
Speaking of which there is a poll out for Tennessee which shows a 20 point Santorum lead. I don’t think even Romney’s attack machine will be good enough to claw him back over there and in Oklahoma and he has a massive lead in Virginia so Ohio may be the only race which is up for grabs though of course it’s a massive prize.
Is Santorum on the Virginia ballot? I thought that was just Romney and Paul.
Correct. He means that Romney has a big lead in VA. So Super Tuesday comes down to Romney winning a non-competitive VA primary and Santorum winning the Southern states, with only Ohio being interesting.
Of course, one could also argue that the nominee to win Ohio is pretty much exactly the candidate the GOP should want in the general, since that seems to me to be very high on the “tipping point” list.
Ah yes, I had actually forgotten. I was wondering why Romney had such a massive lead in Virginia and it’s of course because his main competitors aren’t on the ballot. That makes Ohio all the more important. Romney really doesn’t want to come out of Super Tuesday with just a Virginia win.
So, if Newt drops out, Santorum may be resurgent. But, now Palin is dropping hints. Could there be one more clown in the car?
Crane
No. Palin’s a narcissistic, yet lazy know-nothing. All she wants is for people to continue talking about her.
At this point, I doubt many die-hard, true-believing, Tea Party loving conservatives take Palin seriously anymore.
ETA: I changed my mind. Palin does know something; she knows how to stretch her 15 minutes farther than almost anyone.