SoylentNews is people
SoylentNews
Submitted via IRC for chromas
The Clinton presidential campaign used a complex computer algorithm called Ada to assist in many of the most important decisions during the race.
According to aides, a raft of polling numbers, public and private, were fed into the algorithm, as well as ground-level voter data meticulously collected by the campaign. Once early voting began, those numbers were factored in, too.
What Ada did, based on all that data, aides said, was run 400,000 simulations a day of what the race against Trump might look like. A report that was spit out would give campaign manager Robby Mook and others a detailed picture of which battleground states were most likely to tip the race in one direction or another — and guide decisions about where to spend time and deploy resources.
Of course, the results are only as good as the data. Since the outcome of the election was different than most poll predictions, it seems like Ada may have had a Garbage In, Garbage Out problem.
(Score: 2) by wisnoskij on Saturday November 19 2016, @02:20PM
Poll: 98% chance of winning for Hillary.
So you are implying that the margin of error of election polls are 98%+?
(Score: 3, Touché) by jmoschner on Saturday November 19 2016, @03:30PM
People tend to forget that a 98% chance of winning is still a 2% chance to lose.
(Score: 3, Insightful) by wisnoskij on Saturday November 19 2016, @05:00PM
But real events are not statistical models. If we went back in time 99 times to Nov 7th we would not find that Clinton won 98 times, Trump once. We would find that Trump won with a landslide victory 100% of the time. Trumps chance of a victory was always 100%. The polls just try to approximate this number. So, the polls were off by 98%. That shows a really really bad polling method.
(Score: 2) by JNCF on Saturday November 19 2016, @06:10PM
I think the sources projecting a 98% chance of Clinton winning were drinking their own Kool-Aid. FiveThrityEight seems to do alright at guessing outcomes of elections, and they were giving Trump a 28.5% chance of winning going into the election. They weren't projecting a Trump victory, but they gave him better odds than a critical failure on a d4; a totally plausible outcome. Some folks were even criticizing them for giving him too much of a chance. One reason for them giving him better odds than some other sources is because their simulations are run with poll data that has been adjusted for the historic bias of the institution running the poll. Knowing that at least the Democrats have been conducting new research about which parts of the population to oversample in polls [wikileaks.org] in order to skew the results even more in their favor as recently as 2008, I could see how even FiveThirtyEight's attempt at adjusting for historic bias might not have accounted for the current bias. The current net bias might go significantly further than the historic bias has gone.
I was following FiveThirtyEight's numbers throughout this election, I'm not just using them as a source in retrospect. Here's their full post-election analysis. [fivethirtyeight.com] I'm not claiming to know jack-shit about statistics.
(Score: 2) by JoeMerchant on Sunday November 20 2016, @10:07PM
Each of those polls (in the states) was within a margin of error going either way, or the pollsters are just making crap up and getting paid for it, which I would also believe. The 98% chance of Hillary winning was probably a Scott Adams style media stunt to skew things in Trump's favor - looks like it worked.
🌻🌻 [google.com]