I am trying to understand the betting. I have no doubt that if the election was tomorrow Trump would be at least 20/1 based on current polls.
Yes all the social media stuff is affecting the way folk make up their minds but a poll is a poll. The betting markets clearly do not think that today's polls are a reliable indication of people's views in 2 months' time, but I think they would accept that they are a fair representation of their current views.
They think it is all up for grabs. I have mentioned a few possible game changers in Trump's favour between now and polling day. There may also be the view that he is not beyond dirty tricks. Maybe take out another Iranian general or a faux trade dispute with China.
I disagree with some of the predictions on here and think an evens bet on Biden is pretty good odds at the moment. Recall that the overall polls in 2016 were reasonably accurate at the overall vote share. The failure was on the state by state calculations and the electoral college. But the outcome at the time was rated as a 30% chance - not out of the range of possibilities. And the estimate is about the same now - again not out of the realm of possibilities but would be unlikely.
But if we forget our biases and leanings, looking purely at the betting markets rather than polls there is an internal inconsistency. I can't square the following;
Overall winner : pretty much dead even. But popular vote winner is about 1:4 for Biden. So the betting market is predicting quite a significant likelihood of Biden winning the popular vote.
So that requires Trump to repeat 2016 and win the electoral college with a minority popular vote. Two ways he can do this;
- Repeat 2016 and win Florida, Penn, Wisconsin, Michigan, Georgia, Arizona, North Carolina
- Not win all of the above but swap out a number of smaller states for some of the bigger ones which would require winning Navada and Maine but still requires Florida, Georgia, Arizona, North Carolina.
So looking at the betting markets for the states. The states he needs to win in either route
Florida (even)
Arizona (even)
Georgia has Trump at 1.3 so pretty safe but North Carolina at 1.66 is less conclusive
For route 1 above
Pennsylvania has Democrats favourites at 1.6
Wisconsin has Dems @ 1.7
Michigan has Dems @ 1.5
For route 2 above
Maine has Dems at 1.25
Nevada looks pretty safe
So I can't reconcile markets which (a) show high probability of Biden winning popular vote, (b) shows a lot of critical states at evens which Trump would need to win ALL of irrespective of which route to 270 electoral votes and(c) shows Democrats favourites to win swing states in both routes to 270 which he would also need to win ALL of (in either route) and yet.... has Trump as evens to get to 270.
On a pure stats / probability basis - there has to be an arbitrage there. From a stats point of view, (Fl @ 0.5) x (Ar @ 0.5) etc etc can't equal 0.5 probability of overall win
EDIT : had originally stated 290 electoral college votes as amount required as spotted by Duke. Have changed