Protected: Two ways to estimate primary outcomes without polls (transcript)
This content is password protected. To view it please enter your password below: Password:...
East Coast primaries – open thread
Based on polls and border counties (see summary), I expect Donald Trump to get over 85% of all the delegates to be voted upon today. I estimate that he will gain about 150 delegates (43 of these are district-level Pennsylvania delegates, whose rate of faithfulness I estimate to be 0.8). Trump’s total number of delegates might exceed 1000 tonight. [live results at HuffPost]
More calculations from N. after the jump.
N. has been playing around with dropping various states from the Google Correlate algorithm, and finds that dropping the three candidates’ home states tends to reduce the error of the imputation. N. found that distinguishing between primaries and caucuses didn’t help. She then estimated the overall error by dropping one additional state and calculate its imputed value, then repeating for four states that have voted so far (ID, MO, WI, and NC). The average error was 3.4%.
Here is the scorecard that N. is keeping.
For the record, I do not expect to do particularly well on N.’s measure of vote share. My goal this year was to estimate delegates.
11:00pm: N. may not be able to spell my name, but she can predict election results without using any polls at all.
Topics:
Democratic primary in Rhode Island looks like it has a spread in predictions. All predictions have Clinton negative, from -1.1 to 8.8, while the RCP poll average is +2.5. This should be interesting test of prediction methodology.
Looking at Trump’s margins, it sure looks like Google Correlates will be the hands down winner, even over polls (I mean, the Brown U. poll had Trump at 38% just two days ago). This is incredible. My compliments to N. It sounds like Voodoo, but it works. Please ask N to update her correlates and give us new projections for the upcoming states. Also, does this mean that Trump should redirect whatever money he would have spent on commercials to sponsoring reruns of DeGrassi?
It may be time to remove the Cruz factor from the model.
Thought Kasich would do better in Pennsylvania, given that metro Pittsburgh borders Ohio.
Joel, you forget that Pittsburgh is now a huge university town, and that the new major industries are banking & health care. The academics are all Sanders people, while the industry folks are all Clinton people. The rural voters loved it when Trump promised to bring steel back. No constituency for Kasich here, sorry. Your friend at Pitts, m.
Is there a reason that PA has 71 delegates but the AP is only awarding Trump 17? Typo?
They’re unbound. Sam discusses this elsewhere. Here’s chapter and verse:
http://www.thegreenpapers.com/P16/PA-R
Thanks. I figured that out a few minutes ago and felt foolish! Probably not as foolish as the people who adopted the Republican PA delegate rules. They’re really strange.
Jim: I dunno if they’re feeling foolish – some might well be positively giddy about the idea of PA delegates getting special attention from campaigns (although chances of it mattering may have mostly faded now).
Are you sure that N. isn’t really Hari Seldon?
A little man in a wheelchair and with a cigar…
Hillary increased her lead to more than 300 pledged delegates last night – and almost 750 if superdelegates are included in the count. I think it’s safe to say that her chances at clinching the Democratic nomination have climbed well above 99 %.
You guys are outstanding. I am a hard core political junkie, and you guys are it.
Knowledge is power. Armed with your output – done for FREE (thanks) – I’ve reached a point with 538 and the news channels that I’m like, “damn, don’t you guys read Wang?”
What the hell.
Dr. Wang you mention 43 unbundled delegates. Isn’t it 54? (71-17), or is there a third class of delgates that comprise the other 11?
54*0.8=43