Kyle Kondike at Sabato's Crystal Ball published the first professional analysis of the 2020 Senate map I've seen this election cycle, but I need twenty posts to include links, so I will just quote the bottom line predictions and hope that the next person links the article.
I largely agree with this, however I think that Jones in Alabama is basically in the same position in 2020 that Scott Brown was in 2012: unless he gets another amazingly awful opponent, he's a goner. Unlike MA in 2012, it's likely that the president won't be popular nationally, but also unlike MA, Alabama is a very
inelastic state and I get the impression Jones has worked with Trump far less than Brown worked with Obama.
I also think that if Trump's popularity stays where it is now, Montana, South Carolina, and Texas could be competitive depending on the Democratic candidates. Montana is a very elastic state, and Kondike mentions that if Steve Bullock runs it could be competitive.
Meanwhile Texas and South Carolina aren't quite as red as people tend to think they are, but they are relatively inelestic states, which helps the GOP for as long as they remain even slightly red. Graham in particular could be a weak incumbent because he is a former NeverTrumper who did an about-face. Cornyn and Graham's speeches made a strongly negative impression on me at least, and I no longer think of Cornyn as "the sane one." I'd say that Cornyn is still safer than Graham, but If Cruz goes down in this election, Cornyn may look even more out-of-touch,
and it might also convince the national Democrats that Texas is worth investing in. What I don't know is whether there are Democrats in Texas and South Carolina who are willing to run and who could repeat O'Rourke's success in fundraising and building statewide name recognition.
If Trump's popularity were to go down further, I feel like it is most likely to affect the agricultural states, which are already feeling the effects of his trade war. This could potentially make Kansas, South Dakota, and Nebraska competitive, especially if Sasse faces a strong primary challenge. However the remaining Southern states plus Idaho and Wyoming I can't imagine being competitive even if Trump gets Bush-in-2008 numbers, except Alabama, where Jones would no longer be a definite goner.
Conversely, I don't think that if Trump's popularity were to go up (even if his favorable and unfavorable numbers reversed), it would make any of the blue states Kondike rated as safe competitive, though his popularity going up combined with a retirement or a Menendez-level scandal
could do it. But I do think that a rise in his popularity to that level removes all potential Democratic pickup opportunities except Colorado, Arizona, Maine, and North Carolina, while helping Republican challengers in Michigan and Minnesota and possibly New Hampshire. Even if his numbers end up evenly split, I would take Kentucky and Georgia off the list, and potentially Iowa, though his trade war may throw a wrench into that race.