At this point (or if not pretty soon) the reason to put non-zero weight on fundamentals is more because polling is just wrong sometimes because of non-response-bias/uncertainty about turnout/etc than it is that polling will converge toward fundamentals
Wouldn't that be a reason to increase uncertainty in the model? It seems like you're saying 'fundamentals' get used just because it's available data, not because of positive predictive power
-
-
Simple example: Predicting house national vote from 1946 to 2014 in midterms with a polling average of all generic ballot polling in the last two weeks and a "fundamentals model" of "Party of the president tends to get 48% of the vote". Fundamentals get 56% weight.pic.twitter.com/E63XU3dZsG
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.