Control of the Senate is being decided Tuesday, with Georgia holding a pair of Senate runoffs. Democrats need to win both to gain a 50-50 Senate and effective control of the chamber. And if they were to win, they would control not just the chamber, but all three levers of lawmaking power in Washington.

But to do so, they’ll need to beat history.

Despite President-elect Joe Biden’s narrow win in the state in November, Democrats took fewer votes than Republicans in both Senate races — by about one percentage point in the race between appointed Sen. Kelly Loeffler and Raphael Warnock (when you add all the Republicans and Democrats in that race together), and by nearly two points in the race between now former senator David Perdue and Jon Ossoff. So to win the two seats, they’ll essentially need to come from behind.

But over the past quarter-century, that has been a tall order for the Democrats. Only once since 1992 have Democrats done better in runoffs after a November election than they did in the general election. Over and over again, runoff turnout has favored Republicans.

There are reasons to believe this could be different, which we’ll get to. But first let’s run through the numbers.

Since 1992, we’ve seen 10 Republican vs. Democrat runoffs after a November election in which no candidate received a majority of the vote. (Often, the runoffs have been forced because of Libertarian candidates taking enough votes to pull the major-party candidates below 50 percent. Sometimes, special elections have featured multiple candidates from one or both of the major parties, as the Loeffler-Warnock special election did.) Of those 10 races, eight have been for statewide office and two have been for U.S. Senate.

Of those 10 races, Democrats have exceeded their Election Day performance in just two, and they’ve won only one: a public service commissioner race back in 1998. They’ve won just that one race despite leading on Election Day in five of the 10 races.

As for those two Senate races: Democrats took more votes on Election Day 1992, only to have Sen. Wyche Fowler lose narrowly and surprisingly in the runoff to Republican Paul Coverdell. There was another runoff in 2008, in which Republican Sen. Saxby Chambliss led Democrat Jim Martin by three points on Election Day and then trounced him by nearly 15 points in the runoff.

That same year, Republicans also flipped the vote in a public service commissioner race in which they had trailed by less than a point on Election Day, winning the seat by a nearly 14-point margin.

Those were two of the biggest general-election-to-runoff shifts in the GOP’s favor since 1992, but they’re hardly alone. Across the 10 races, Republicans have gained an average of 5.2 points on their Election Day margins. The only big exception to the trend is the 1998 Public Service Commission race, which went from a 16-point Democratic margin to a 31-point margin.

And the only other race in which Democrats did better between Election Day and the runoff was a local race — a 2010 special election for the district attorney for the Macon Circuit, in which the party did nominally better (by about 0.4 percentage points) but still lost. And besides being a low-turnout local race, that one was a bit of a special case in that the initial race featured one Democrat and three Republicans who split the conservative vote — and whose supporters would logically have been more accessible to the Democrat.

We have already seen a few post-election runoffs since November, but only one of them could be seen as even a little bit instructive, given almost all but one featured matchups between candidates of the same party. A district attorney’s race in the Athens area initially featured two Democrats and a nonpartisan candidate favored by the local GOP. The Democrats took about 30 percent more votes on Election Day but won by just four points in the December runoff.

That wouldn’t seem to mean much given how local it was and the lack of an official Republican nominee, but it doesn’t exactly downplay the Democrats’ challenges on Tuesday.

But we are in some uncharted territory, in which the stakes in this race are unmatched by any of the races detailed above, with control of the U.S. Senate on the line.

Fowler’s 1992 loss, by contrast, was the difference between a Democratic majority and a slightly smaller Democratic majority. Chambliss’s blowout 2008 win was pretty significant, too, in that it deprived Democrats of an effective 60-vote, filibuster-proof majority in the chamber — which was a key GOP argument against electing Martin.

It was perhaps understandable that what was then a more conservative state would bristle at that prospect, even if President Barack Obama over-performed the state’s fundamentals on Election Day. But Republicans have made a similar argument in this race, including Trump saying that Loeffler and Perdue are the “last line of defense” (despite quickly contradicting himself on that point).

Democrats failed to turn out in nearly as strong numbers in virtually all of these races, but perhaps that could change with control of the Senate at stake. And perhaps a state that is now more evenly divided could provide a different proposition, in which lower turnout might not so obviously favor one party. There’s also the matter of Trump’s constant and baseless complaints about the legitimacy of Georgia’s election, which Republicans fret could depress turnout by convincing enough Republicans that it’s just not worth it to vote.

But something I keep coming back to is Democrats having taken fewer votes on Election Day — and the fact that Republicans arguably have more to vote for, in that this is their last hope for controlling at least one lever of power in Washington. When we saw such a pivotal vote in 2008, the GOP had some of its best runoffs of the past 30 years.

If Democrats can win these races, it would be a significant statement about how the state has shifted politically.