Based purely on maximizing votes across the widest # of states (max prob of winning, min risk) here is my list
Buttigieg/Abrams or Martinez
Klobuchar/Abrams or Martinez or Castro
Harris/Buttigieg or Castro
Warren/Castro or Abrams
Joe Biden is a mess disguised as a safe pick
Buttigieg/Abrams or Martinez
Klobuchar/Abrams or Martinez or Castro
Harris/Buttigieg or Castro
Warren/Castro or Abrams
Joe Biden is a mess disguised as a safe pick
You'll note: I'm presenting tickets. I think the ticket matters a lot and that diversity on the ticket (race/gender/ethnicity matters a lot lot. I argue in my book that had HRC picked a more progressive or diverse running mate in '16, she prob could have won.
Also to note: this is a list. It isn't meant to be an end all be all & it could change if conditions change. It also is unlikely, bc right now Biden is hanging tough. But boy, that debate performance was tough. It was like watching a sports star come out of retirement and flop
But if Biden's the pick at least Trump also has major mental acumen issues. Really critical that Biden pick @staceyabrams as the VP though. CRITICAL.
Oh god yes, I was thinking of the current governor of New Mexico and had my own Biden moment! Michelle Lujan Grisham NOT Martinez.
@staceyabrams Had some questions as to why Warren is low on the list & no Sanders. Again, I am solving only 1 equation: vote maximization + risk minimization. If you're curious about the rationale here is the ideology of the overall electorate. Although there has been quick growth the last 10 
years in the percent of people who describe their ideology as "very liberal" has gone up (doubling in fact) it is still lower than the total number who say "strongly conservative." More importantly, the modal ideology category for Ds, is, in fact, moderate, lean liberal.
This contrasts sharply w the modal category for Rs-which is NOT moderate, lean conservative, but is conservative. So there is a much bigger appetite among Ds for moderates than for "very" liberal candidates (nationally). Also of importance is the ideology of Indies, which is
moderate, lean conservative followed by moderate, lean liberal. Combined, it is overwhelmingly moderate. Of course, the 538 analysis does show that there are many different types of Indies and they are almost certainly more open to populist messaging than the establishment of the
D Party thinks. And in many ways, the progressive wing of the party will already "win" 2020 bc whoever emerges as the D nominee will do so as the most unabashedly liberal nominee the candidate has seen in decades. And both the enfranchisement of millennials and Gen Z voters, as
the fact that the incumbent is terrible make conditions ideal for a liberal candidate to win, bc the Rep nominee could not be more unattractive to many Indie voters. But in terms of risk mitigation, you probably want a center-left nominee and not one that represents the furthest
ideological faction of the party bc right now, much of the American middle is ripe for the picking and probably most important, you want your nominee to be seen by the media as unquestionably more electable than Trump. Yielding that advantage is not a risk mitigation approach.
But again, this is my personal calculations, based again on solving only one equation-maximizing potential to win while minimizing risk. As you know, in my forecast I argue that Ds will have the edge regardless of the nominee. But there are easy paths, and there are hard paths.
To beat Trump Ds need massive turnout of Dem coalition voters, including Indies that sat out of 2016 and the movement of pure Indies to the D side. The ticket combos above are my best guesses at achieving those dual goals.
