Nate Silver Analyzes A 3-Way Delaware Senate Race
Genius statistics guru Nate Silver analyzes the numbers for a possible Castle write-in bid.
I tried to model the three-way race and came up with the following: Mr. Coons with 37 percent of the vote, Mr. Castle with 34 percent, and Ms. O’Donnell with 29 percent.
That, obviously, would imply a winnable race for Mr. Castle — although it assumes that he is fairly serious about his write-in bid. Mr. Castle is 71 years old and has not faced too many competitive elections in recent years, and he took his time to enter the Senate race last year after considering retiring from politics entirely, so this is far from a sure thing.
One dynamic that may be more favorable to Mr. Castle is that it seems conceivable he could win with as little as 35 percent of the vote. It seems probable to me that Ms. O’Donnell’s support is going to hover in the area of 30 percent — not a lot higher, not a lot lower — which was about the percentage of voters who had a favorable impression of her in the Public Policy Polling survey. That would leave Mr. Castle and Mr. Coons to fight over the remaining 70 percent of the electorate, making half that total — 35 percent — the magic number.
Yes, I’m patting myself on the back for getting a similar number for Coons but I think Castle’s number is still overestimated. Plus there’s still 1-2% for 3rd party candidates. Based on this analysis, I think Castle doesn’t take the plunge. He’d have to fight HARD and he’d have to walk a pretty fine line to keep his R-I-D coalition together. It’s certainly a much different race than the one he was planning to run. It’s definitely no coronation. One thing it does do, however, is change the race from Lean D to Toss Up.
What do you think? Does Castle take the plunge?
Tags: Chris Coons, Christine O'Donnell, DE-Sen, Mike Castle
so at this point, Castle still loses. He jsut loses to a legitamate human, AND forever marks Chrissy as someone who lost to a write-in…..
i dunno. i think he should man up and retire to Greenville.
If Castle gets a smaller number than suggested here, who gets those voters?
What if Castle just runs a pro-forma campaign not to win but to tip the plurality to one side or the other.
And how do you even calculate whether someone is likely to actually write in a candidate?
I think that’s a good question anon – I think voters will be a lot flakier about saying they’ll vote Castle and then actually doing it.
If Castle’s plan is to make a Republican win more likely then this is a good way. If Castle is actually looking for some kind of redemption then this is a bad plan. Not only does he have a chance of coming in 2nd place, he could also come in 3rd and he could give the seat to the unqualified person who beat him. I think there’s more downside than upside for him.
so who are the people/players that are going to lose out when Castle is gone? How many people drown in the wake he leaves behind?
like when he goes away how many staffers and suck ups that have made a living off of him their entire lives now have to forage for new connections?
They are going to be the ones to pressure him to do the write in.
dv asks a good question. Does he do it out of some sort of loyalty to his supporters, staff, etc? This had to play into his decision to run for the Senate seat in the first place. The rumors say Jane is pushing him to retire and his health is not as robust as it is portrayed. Has he satisfied himself that he’s done his best for his team?
There’s only one thing I know about this race, and it is that I don’t know anything. Grasping at the only actionable conclusion I can reach, it’s all about GOTV for Chris’s ID’d supporters. We’ve got to get them to the polls. If you can take a vacation day on November 2nd do it and show up to work the GOTV effort.
There are entire committees of Dr’s, Dentist’s, etc that all would meet for this guy. I’m sure they have no idea what to do
When polling comes out showing Mike Castle won’t win and he says “never mind” won’t he look like a chump?
Color me unimpressed by the analysis because it suggests that in a head to head matchup, Castle would get more than 60% of the vote, which strikes me as extremely unlikely.
I think if you simply assumed that the 30,000 who voted for O’Donnell in the primary also vote for O’Donnell in the general means Castle would be in serious trouble. But then factor in that O’Donnell’s name is on the ballot, and Castle’s is not, and I think it’s game over.
I agree with you Publius that this is more of a best-case scenario for Castle. I think all you could say is that is it hypothetically possible for Castle to eke out a win while mostly improbable.
I think O’Donnell’s team might hope for Castle to jump in, since he pulls support from Coons and makes a win for her more possible.
The only scenario under which O’Donnell has a chance to win is if Castle jumps in. Voters are gonna have to vote on who is most electable based on the polls–Castle or Coons–if they want to avoid the disaster of an O’Donnell election.
I sincerely hope he doesn’t run. Can you imagine his last political act as one facilitating the election of a lunatic? Would sure put the lie to this premature ‘selfless public service’ crap I’ve been reading for over a week and a half.
The people he served are the fat-cats urging him to jump in. The ones who don’t want their Bush tax cuts for the greedy to expire.
“…if they want to avoid the disaster of an O’Donnell election.”
That’s a classic example of the “political junky” fallacy. It assumes that voters make rational decisions and are moved to vote based on an analysis of various candidates qualifications and positions.
I think that it is safe to say that voters are fucking morons.
Interesting how in this race the visceral reaction is against O’Donnell for her social conservative positions.
Nate,
My calculation of your numbers give Coons the win with 43%.
What am I doing wrong?
I’m not Nate but I got for Coons: 0.67*45 + 0.30*20 + 0.03*35 = 37.20%. Perhaps you were adding the numbers down the column instead of across the row?
My guess that if he gets in Castle will find that people find it a lot rougher as a write-in candidate. With a chance to keep the seat and fear of COD, the Dems will stay home. Castle could pull some I’s and R’s who are loyal to him but he risks a second embarassment. He’ll come in third. He won’t get in the race.
Don’t you have to be able to spell to vote for a write-in candidate?
Ralph Nader. Had no chance to win. He ran anyway. +4 years GW. *sigh*
If we assume that those numbers are correct, my guess is that puts Castle w/i a margin of error of winning. Bad news if he seriously is itching to go to the Senate. He might go for it if he is getting similar numbers (not to mention that many establishment GOPers in DE are probably, unofficially, pressuring him to run).
The other bad news is those numbers indicate that a Castle write-in candidacy clearly hurts Coons more than it does COD.
I have been hoping for a blow out by Coons.
If she beat off Castle going head to head, fewer Demcrats are going to take the risk of swinging both ways for Castle if he and Coons get into a three way with her. They would be betting that he’d wind up on top, which is a more risky proposition.
I believe that if Castle enters as a write-in, we will see the death of the GOP in Delaware. A similar thing is occurring in Colorado with the teagging GOP nominee for governor running a distant 3rd (after anti-illegal immigrant Tom Tancredo entered the race) and handing the race to Democrat John Hickenlooper. And the CO GOP is also risking being classified as a minor party if their candidate, Dan Maes doesn’t pull 10% of the vote.
“If she beat off Castle going head to head, fewer Demcrats are going to take the risk of swinging both ways for Castle if he and Coons get into a three way with her. They would be betting that he’d wind up on top, which is a more risky proposition.”
*tips hat…. that sir, is poetry!
Running as a write in is stupid. Castle is not stupid. I am right about that? Really,he isn’t is he. I mean he can’t be?????
Anyone hear Dave Burris on with Allan Loudell today? He never let on that he left the party. Touted as a conservative blogger, he said that Mike Castle should not re-enter the race as a write-in and suggested that he would bet that Castle will not do so.
Allan suggested that wasn’t it logical for the GOP to want Castle in the race if the polling showed that votes for him would be pulled off of Coons?
In the 2006 Senate race 258,053 voted statewide with 254,099 voting in the US Senate race for someone on the ballot or casting an O’Donnell write-in vote.
http://elections.delaware.gov/reports/agprpt_2006.html
If turnout is in that range a write-in candidate would need 86,394 write-in votes to get 34%. If turnout is higher the number needed moves up.
___________________________________________________________
From the Tommy Wonk archives:
http://www.tommywonk.com/2006/11/christine-odonnells-write-in-votes.html
Christine O’Donnell garnered 11,127 votes write-in candidate for the U.S. Senate, 4.3 percent of the total votes
________________________________________________________________
Her write-in campaign was pretty successful. Could a Castle write-in campaign really be 8 times as successful? That appears to be what would be needed to be competitive. I doubt close to 90,000 people would write-in .
Thanks John! The math does look difficult. I’m sure Castle has sane people on his team, so he won’t run. The only reason I think he’d run is to hurt Coons’s chances.
John – mmmm. math. In a three way race Mike Castle needs a minimum of 85,000 people to write his name in the little box at the bottom, and NOT push a button. 33.5% of 254,000. Not just 11,128. 30,600 already voted for COD, and I suspect they will do so again. COD already has 12% of the electorate vs. the 2006 general where 252,000 voted. She only needs 55,000 more or 22% of the voters + previous to push the R button and win a three way race.
In 2006 Democrat Spivak received 97,600 votes 38.8% of 252,000 votes. In a normal race I would expect Coons to poll at least that much. I posted that Castles BEST scenario would be 33 33 33. I think it’s a lot worse. Geezer is probably right at 10%ish for a Castle write in. But we still haven’t seen what $2+ million (really?) in negative advertising can do to Coons (and Castle, should he re-enter). I seem to recall that the Dems dropped about 2 million for Carper v. Roth, and that was fairly pleasant.
I fear that the Tea Party advertising against Coons will be so deceitful, so heinous and so prevalant that Chris’s own mother may think twice before voting for him. (Ask Max Cleland if he’s soft on homeland security.)
The Democrats have a GOTV, and they will have it ramped up. If you don’t vote by 6pm, expect at least 3 robo calls (my guess Obama, Clinton and Markell, if you’re a voting Dem). I’m curious what robo calls the I’s and R’s get (Bloomberg?). If you’ve voted consistently in prior years, expect a human call too.
The Greenville Republicans have outsourced their GOTV. Media (not volunteers), the free market, bombarding you with advertisements, telling you that Coons is a (fill in the blank) and supports Obama! An excellent strategy, if they have Hannity, Limbaugh and Palin pitching their candidate (for FREE). Because they reach the angry, motivated crowd. (My in-laws) In a best case scenario for Castle, they cannot get 85,000 people to write a name on a ballot. And ask Castle how well that advertising buy worked for him on WGMD and WDEL.
In a 3-way, Castle beomes a potential spoiler with a bargaining chip. I don’t think he’ll ever endorse Coons (pull an Arlen light), but he does know DC, and could help himself as a lobbyist, if he doesn’t run a write in, and accidentally cede the seat to COD. And yes, the good staff can get jobs there, too.
Castle takes Democrat votes away from Coons; O’Donnell the loser, Coons Castle battle it out.
“I fear that the Tea Party advertising against Coons will be so deceitful, so heinous and so prevalant that Chris’s own mother may think twice before voting for him. (Ask Max Cleland if he’s soft on homeland security.)”
Blue is so right about this. I hope Coons has a plan. He is right where Castle was this far out against O’Donnell. Sitting on a huge lead in the polls with everyone thinking that his opponent is an unelectable doofus. We all know what happened when Fox News poured it on.
“Anyone hear Dave Burris on with Allan Loudell today? He never let on that he left the party.” Who is surprised that Burris is walking back his contempt for Tbagz? Republicans always come home to the ticket by election day.
Thanks for this posting — I’d been looking for numbers like this. It’s hard to see how Castle running could benefit Coons. Do these numbers indicate percentages of all possible voters in each category or percentages of likely voters?
Nate is good, but this response to his post makes a better case for who wins and who loses if Castle jumps in. The short answer: Coons has the most to lose.
I don’t see Castle getting 30% of the vote by write-in. I see a lot of wasted votes, instead. I think that you have to be very precise in the name that you write-in, for example, a vote for “Mike Castle” would not be the same as vote for “Michael Castle” which would not be the same as a vote for “Castle.” And that is assuming that each ballot is completely legible and properly spelled.
I think that getting voters to write the proper name in legibly and constantly one reason it is hard to mount or win a write-in campaign.
Anonone – One of the “reasons” you are supposed to register as a write in candidate is so they will count all write in votes that resemble the name of the candidate: Cassle, Kassel, Mike, et al. Still, it takes a lot of dedicated voters to go from pushing a button, to sliding the little door and writing in anyones name. I don’t know how many names each paper tape can hold. If it gets filled up, do they have extra paper tape on hand? Are there pencils/pens in the booth?
If Mike runs a write in campaign, he better have volunteers at all the polls handing out pens with his name on it. (Which would violate election rules about bringing candidate swag into the polling area – so maybe just pens.)
Thanks, BC. I did not know how that worked.
I Delaware has an intent law, so misspellings of the name could be counted. I think Delaware also displays the names of write-in candidates somewhere. However, writing in a name takes way more work. Plus, the voters have to be really motivated to do it. Does Castle really think he’s going to get those people who don’t pay attention to politics to write in his name? Why does he think he has an enthusiastic base of support if he couldn’t even get them to turn out to vote in the primary?
As a rule, independents are not generally enthralled with politics so a strategy that relies on them is pretty tough.