538 Democratic primary model is up

[jaichind]

I am surprised given the high upward, even though very unlikely, potential for a Bloomberg surge Super Tuesday followed by sweeping other states that Bloomberg is nowhere in this model.

[jaichind]

Given where Bloomberg is in the polls this model should really include him if they are going to include Buttigieg.

[jaichind]

So the candidate named "No one" is leading. How can I donate this wonderful candidate?

What is Governor No One's position on health care ?

[jaichind]

Something else I noticed:

In almost every Super Tuesday state, Biden is ahead of Bloomberg in terms of probability of *winning* the state (though still behind Sanders in most cases), yet in every non-Southern Super Tuesday state, and some Southern ones too, Bloomberg is ahead of Biden in the *average* number of delegates won across all runs of the model.  How can this be?  Presumably, it's because Biden has much higher variance than Bloomberg in terms of how strong he'll be on Super Tuesday.  We already know that Bloomberg is getting 0% in Nevada and SC, since he's not competing in either state.  But Biden's strength going into Super Tuesday is highly variable, because whether he's a winner coming out of SC will make a huge difference.

So in the median case, Bloomberg will do better than Biden on Super Tuesday....not necessarily winning anything, but coming in 2nd place to Sanders all over the map.  But Biden has a wider range of outcomes, so has a better chance than Bloomberg of actually winning some states (and a better chance that Bloomberg of totally bombing).

A lot of this is because the model gives a significant chance of Biden dropping out by ST.   The model gives a significant chance of Biden dropping out by SC primary which I think is way too high.  There is no way Biden drops out before SC primary.  I think the model overestimates the chances of a Biden drop out by ST but I guess he could always totally bomb NV and then SC.

[jaichind]

The model still has a non trivial chance that Biden has 0 delegates from SC which is him dropping out. That makes no sense given Biden performance in NV. Overall the model way overestimates the chances Biden will drop out before ST.

[jaichind]

That makes no sense given Biden performance in NV.

The model was frozen just before NV voted though.  It hasn't incorporated the results from the state yet.  Presumably, once it does, its odds on Biden dropping out pre-SC will change.

That's fair.  Hopefully they also adjust downward the chances of a Biden drop out before ST which is pretty much zero at this stage.

[jaichind]

https://fivethirtyeight.com/features/what-the-race-looks-like-if-biden-wins-or-doesnt-win-south-carolina/

If it is the Biden wins big in SC this model seems to put the Sanders march to victory into jeopardy as it would produce a situation where Sanders would be on pace to win less than 40% of the delegates where it would give some legitimacy toward some convention maneuver to come up with a compromise candidate.   

[jaichind]

"No one" continues to gain at the expense of Sanders, while Biden continues to inch higher.

No majority 53
Sanders 28
Biden 17
Bloomberg 2

The model seems to very heavily invest in a pro-Biden momentum from a significant victory in SC.  Main risk there is not the significant Biden victory in SC which now seems likely but if Biden have enough infrastructure in ST states to take advantage of a post-SC surge.