The problem of election forecasting models is that you can never really trust a model just because it was in sync with one election, but there's always a sneaking suspicion that a model with multiple observations is counting a world that no longer exists.https://twitter.com/BenjySarlin/status/1225435542701137921 …
-
-
Replying to @baseballcrank
And this, my friends, is why you need to put enough variance in your model -- not just enough to account for sampling error, but enough to account for the unknown electoral shifts
1 reply 0 retweets 0 likes -
Replying to @DongJohnsonIV @baseballcrank
And this, btw, explains Nate Silver's success is 2016. All the models predicted a Hillary victory. But those models with the most variance showed the least certainty and therefore "won" 2016.
1 reply 0 retweets 0 likes
Replying to @DongJohnsonIV
Right. I expected Hillary to the end - everyone did except the loons - but Silver & Trende convinced me that Trump was still in the game in the final weeks. But people wanted to buy what Sam Wang was selling.
7:24 AM - 6 Feb 2020
0 replies
0 retweets
2 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.