(CNN) All the signs suggest that 2018 is going to be a very good year for Democrats.
Midterm elections are historically terrible for the president's party. In 18 of the last 20 midterm elections, the president's party has lost seats. In those 18 elections, the average seat loss is 33. Those numbers are even more daunting for presidents under 50% job approval -- as Donald Trump is right now. Since 1946, the average seat loss in the House in that situation is 36 seats.
But before Democrats get too delirious about the election to come, they should read this paragraph from David Wasserman's terrific analysis of the 2018 election on FiveThirtyEight:
"Even if Democrats were to win every single 2018 House and Senate race for seats representing places that Hillary Clinton won or that Trump won by less than 3 percentage points — a pretty good midterm by historical standards — they could still fall short of the House majority and lose five Senate seats."
That's absolutely stunning. And reflective of the advantages Republicans have going into 2018 -- one, in the House, built on having largely controlled the 2010 redistricting process, and the other, in the Senate, based on how great the 2006 and 2012 elections were for Democrats.
In the House, there are 23 districts currently held by a Republican that Hillary Clinton won in 2016. (There are 12 seats held by Democrats that Trump won.) Of those 23, just eight went for Barack Obama over Mitt Romney in 2012 as well.
Even so, if Democrats won all 23 of the seats Clinton carried that are represented by Republicans -- and lost NONE of the dozen seats Trump won that are held by Democrats -- the party still comes up a seat short of the majority.
Republicans, quite simply, did a very good job in drawing the congressional lines in states where they controlled the entire process after the 2010 census. Large populations of Democrats are, as Wasserman notes, packed into urban districts while Republican voters are more spread out among suburban and rural seats.
That reality, coupled with the fact that political tribalism is on the rise, means that there are just far fewer chances for Democrats to make gains than there were a decade or two ago. (In 1996, there were 108 "crossover" districts where the member of Congress was from a different party than the presidential candidate who carried the seat.)
A good midterm election wouldn't be enough to switch control of the House. Democrats would need a great one. Which is possible -- especially given Trump's dismal approval ratings and the lack of legislative accomplishments for Congress -- but never a certainty.
On the Senate side, where redistricting isn't a factor, Democrats are a victim of their own successes. Democrats won six Republican seats in the 2006 election. In 2012, Democrats picked up two more seats -- if you include Maine Sen. Angus King, an independent who caucuses with Democrats. And I do.
That embarrassment of political riches means that in 2018 there are a whopping 25 Democratic seats up as compared to a meager eight for Republicans. In other words, 52% of all the seats Democrats control are up in 2018 while just 15% of Republicans' seats are up.
And it's not just the raw numbers. It's where these seats are. Ten of the 25 (40%) are in states Trump won in 2016. TEN.
Five of those 10 -- North Dakota, Missouri, Montana, West Virginia and Indiana -- are states Trump carried by double digits. By contrast, only one Republican up for reelection -- Dean Heller of Nevada -- represents a state that Clinton won last November. Only one other Republican-held state -- Arizona -- was even marginally competitive in the presidential contest. (Trump won Arizona by 3.5 points.)
In short: The political environment is looking very, very good for Democrats. But the math is unchanging -- and bad -- for them.
History suggests that political environments can sometimes overwhelm raw numbers. But, for that to happen you need a gale-force wind blowing in one direction. And that is a very rare thing.