Originally Posted By Mr X "According to our final polls-plus forecast, Hillary Clinton has a greater than 99% chance of winning the Michigan primary." <a target="blank" rel="nofollow" href="http://projects.fivethirtyeight.com/election-2016/primary-forecast/michigan-democratic/">http://projects.fivethirtyeigh...ocratic/</a> Bernie Sanders just won Michigan.
Originally Posted By ecdc No. There were seven solid polls in the last month in Michigan showing Hillary up by anywhere from 12 to 30(!) points. Unclear so far how the polling from numerous outlets went so wrong.
Originally Posted By Mr X Sure. But then there was also his claim a few months back that Trump stood around a 2% change of getting the nom (including six supposed hurdles, most of which Trump seems to have blown through at this point). Just sayin...methinks that aura of science in his mathematical gymnastics is starting to fade.
Originally Posted By Mr X <a target="blank" rel="nofollow" href="http://fivethirtyeight.com/features/donald-trumps-six-stages-of-doom/">http://fivethirtyeight.com/fea...of-doom/</a> Trump has already surpassed stages 1 through 4.
Originally Posted By Kar2oonMan that aura of science in his mathematical gymnastics is starting to fade.<< I dunno. He said 99% chance, not 100%. It's sort of like that scene in Dumb and Dumber, Jim Carrey asks if there's any chance the leading lady could fall for him. "One in a million?" "More like one in a billion." He grins and says "So, you're saying there's a chance…" ; )
Once more, Silver's strong prediction (90% chance as of last update) of a Clinton win, this time in Indiana, goes south. http://projects.fivethirtyeight.com/election-2016/primary-forecast/indiana-democratic/ I still say he's lost that aura of invulnerability this cycle bigtime.
Perhaps. But primaries are notoriously more difficult to call accurately than general elections. His record there is still excellent (I think 49 states out of 50 in 2008 and all 50 in 2012, IIRC). Primaries are harder for many reasons, the primary (see what I did there!?) reason being that determining who a "likely voter" is for a primary is much more difficult. I don't think there's anything wrong with his algorithms (unless he really falls down in the General); it's just that primaries are more likely to defy algorithms.
I guess for me the main thing is his style. He claims there are these gigantic hurdles for Trump, and that he only has a snowball's chance. Now the egg on his face has dried completely, since Trump "beat the odds". Except "the odds" were based on a whole crapload of unprovable assumptions to begin with, not hard data like he'd become famous for using. Then he comes up with these lofty percentages like 90%, or even 99%, only to be proven wrong. Heck, even if he wrote it as a range I think it'd be more reasonable. He seems cocky to me, is the point I guess — thus when proven wrong it seems like a bigger deal.
To his credit, he admits how wrong he was, and tries to fathom why. Why Republican Voters Decided On Trump
