Former Georgia gubernatorial candidate Stacey Abrams speaks during the National Action Network Convention in New York on April 3. (Seth Wenig/AP)

For Democrats, the obvious best-case result from the 2020 election is that President Trump is defeated and Democrats retake the Senate while holding the House. The worst case involves Trump being reelected. What’s unclear is how Democrats would rank another option: retake the White House but Republicans hold the Senate. Is that closer to best case or worst case?

A number of Democratic voters appear to be frustrated by the fact that candidates who would enter contested Senate races as strong contenders are deciding not to run (at least for Senate). Former state lawmaker Stacey Abrams in Georgia. Former congressman Beto O’Rourke in Texas. Gov. Steve Bullock in Montana, who obviously already holds statewide office. Abrams, for her part, nearly won the Georgia governor’s seat last year, and O’Rourke nearly triumphed against Sen. Ted Cruz (R-Tex.).

There’s one theory about why these candidates might balk at running for the Senate in red states. In 2016, there were 34 Senate races on the ballot. Not a single one of those races saw a Democrat win in a state won by Trump and not a single one of those races saw a Republican win in a state where Hillary Clinton got more votes. It was the continuation of a trend: Fewer and fewer split-ticket results between the Senate and the most recent presidential results in states since 2010.

(Philip Bump/The Washington Post)

If you’re thinking about running for Senate in, say, Montana — where Trump will almost certainly win in 2020 — that’s daunting. Are Trump voters going to cast a ballot for Bullock?

But then 2018 happened. Last year, there were more split-ticket results than there had been since 2010.

(Philip Bump/The Washington Post)

You’ll notice that there’s a pattern in which there are more split-ticket states in off-year elections, precisely because the presidential candidates aren’t on the ballot. But the jump last year was remarkable, with Democrats winning Senate seats in Arizona (Trump +4 in 2016), Michigan (Trump +0.2), Montana (Trump +20), Ohio (Trump +8), Pennsylvania (Trump +1), West Virginia (Trump +42) and Wisconsin (Trump +1). Several of those were incumbents being reelected such as in the most-red states, West Virginia and Montana. Others won precisely because Trump’s victory was so narrow. But it’s a reminder that “red” and “blue” are malleable terms.

Another thing that happened in 2018 is that the margin of support between the presidential and Senate candidates in each race widened after narrowing very consistently since 2004. In other words, not only have red states been electing Republicans to the Senate more consistently (and blue/Democrats), the margins by which those Senate candidates win have mirrored one another more closely.

In 2018, that difference jumped back up.

We can look at this another way. In 2016, there were 13 states in which the margin by which Senate candidates won was within five points of what the candidate of the same party got in the state’s presidential contest. In 2018, only five states were similarly close.

To retake the Senate, Democrats need to hold their existing seats (including in Alabama, which saw a split-ticket result in 2017′s special election) and win five additional seats from the Republicans. Cook Political Report has two states with Republican incumbents where the race is expected to be a toss-up, Arizona and Colorado, which backed Clinton in 2016.

In Maine, the seat held by incumbent Sen. Susan Collins (R) is listed as “lean Republican” — slightly safer for her party, though her state also voted for Clinton. There are three seats that are “likely Republican,” including the one in Georgia, North Carolina and Senate Majority Leader Mitch McConnell’s seat in Kentucky.

The bad news for Democrats? Since 1990, there have been seven elections in which one party won blue-state Maine’s Senate seat while the other party triumphed in the most recent presidential contest. In the other five states above, it’s only happened 11 times — and only three times since 2006, including once in Colorado. (In Montana, there have been eight split-ticket votes since 1990 and three in the past decade.) If Democrats pick up Maine and Colorado — beating incumbents — they still need three more red-state victories to retake the chamber.

The good news? The 2018 results suggest that 2014 and 2016 may have been exceptions, not markers of a new uniformity in Senate voting. What’s more, since 1990, presidential years in which an incumbent was on the ballot (1992, 1996, 2004, 2012), saw a quarter of Senate races involve split-ticket results. That’s closer to the off-year rate (29 percent) than the rate in presidential election years without an incumbent (2000, 2008, 2016): 18 percent.

What do Democrats need to win? In part, good candidates.