2016 Postmortem
Related: About this forumSam Wang (Princeton) just poked Nate Silver on his prediction model...again
If you don't know their history, you may not know who this is directed at, but, trust me, it is a stick at Nate:
It seems to me a good prediction should do more than follow current conditions. Better to change slowly over time https://t.co/sgZuZoNV8b
https://twitter.com/SamWangPhD/status/778720189458231298
He has really not been happy with the new Nate predictive models at all this cycle.
Loki Liesmith
(4,602 posts)Basically says Silver's estimate "probability" behaves nothing like an accumulated estimate of binary outcome probability:
https://twitter.com/nntaleb/status/771372006914846720
and
https://twitter.com/nntaleb/status/778559824305000448
RAFisher
(466 posts)Wang has Ohio and Florida toss-up. Silver doesn't do tossups and has them both red. The only real difference they have is ME-2. Wang has it blue and Silver has it red. What's mildly funny in all of this is that between the two neither actually has a degree in statistics. But neuroscience does use alot of statistics.
Loki Liesmith
(4,602 posts)The difference in the forecasts is just how likely the map is.
For what it's worth David Rothschild of Predictwise is also trolling Silver:
https://twitter.com/DavMicRot/status/778543535545053184