(1) The whole country is voting on ANGER. That sentiment has been commonplace in American Presidential elections, at least since 2008, but the depth and breadth of that anger are very unusual in midterms. They are even more usual for times marked by unprecedented prosperity, low unemployment, and no major wars. Both parties have stoked the fear and rage, figuring it is better to mobilize their base voters than to rely on centrists. Both are now locked in to this divisive appeal, not just for this election but for 2020. That means the election is less about who you are voting for than who you are voting against.

(2) The midterms have been nationalized. That, too, is very unusual. Midterms are usually localized, focusing on issues in a particular district or state. The reason for the change is obvious. No modern President has been more polarizing than Donald Trump. That’s a big plus for Republican Senate candidates in red states, a big negative for Republicans in suburban House races, where educated, higher-income voters are repulsed by Trump’s rhetoric and personality, not by his policies.

(3) Several big Senate races are polling within the margin of error: Florida, Indiana, Missouri, and Nevada. Several others, including Montana and West Virginia, are also close. The narrow polling margins mean the outcome depends on turnout and campaign ground games. Both sides have known that for a long time, so few will be caught by surprise.

(4) The Democrats, who normally beat the drum about “big, rich donors” and the Supreme Court’s “awful decision on Citizens United,” are now mute on the subject. That’s because the big, rich donors this year are left-wing billionaires from New York and California. They seem very interested in like-minded candidates across the South, especially this year’s JFK, Beto O’Rourke in Texas. The interesting question: will the strong, anti-Trump stances that draw big money from these donors actually hurt their candidates by pulling them away from positions that are more popular with local voters?

(5) Polling is getting more and more difficult to do, despite steady technical improvements by pollsters. The problems lie in so many non-responses, the variability of turnout, dishonest answers to polling questions (especially voters’ reluctance to say they support pro-Trump candidates or oppose some minority candidates), and, in this election, the need to poll each House race separately. The turnout issue is especially vexing because the raw numbers have to be tweaked to account for expected turnout, which varies significantly by sub-groups and has been increasingly hard to forecast.

(6) Despite these uncertainties, pollsters and betting markets give the Democrats high odds of taking the House. The respected Cook Political Report is also showing growing numbers of contested races in once-reliable Republican districts. Again, turnout is the key since marginal votes come disproportionately from Democratic sub-groups.

(7) A narrow victory in the House for either party means legislative deadlock. The House Republicans are too divided to pass much. The House Democrats would have less trouble passing bills, knowing they are all rhetorical exercises that will die in the Senate. Of course, a Democratic House can—and will—investigate, investigate, investigate, but the Senate is far more important since it alone can approve Trump’s nominees for courts and executive branch. In the first two years, the courts have been a major win for Senate Majority Leader Mitch McConnell and Judiciary Chairman Chuck Grassley. The Republican base couldn’t be happier with those results. That’s one reason the Kavanaugh hearings were so important in restoring Republican enthusiasm. It reminded them of the stakes.

(8) A Democratic House with a narrow majority would be a nightmare of investigations against Trump but, paradoxically, an electoral godsend for him and the Republican Party in 2020. That’s because a Democratic House would cater to their enraged base, as they did in the Kavanaugh hearings, and overplay their hand. In the process, they would alienate many centrist voters and become an easy target to run against in 2020.

(9) The victories or defeats of three very progressive candidates in big states–Georgia governor, Florida governor, and Texas senator–should have a major impact on how Democrats position themselves going forward. Victories would move the party even further left, losses would compel rethinking—and perhaps more candidates like Pennsylvania’s Connor Lamb, who fit the local voters, not the national donors.

(10) If the Democrats do not win the House, the old leadership will be amputated, without anesthetic. If Democrats do win but by only 5-6 seats, they could still face a real mess in the Caucus about electing Nancy Pelosi speaker. That’s because the national Democratic campaign committee adroitly backed centrist candidates for purple districts, and those candidates explicitly promised not to vote for Pelosi as Speaker. If the Democrats take the House, those newly-elected representatives will have to ask themselves which is their greater fear: (a) betraying their voters and risking the next election, or (b) betraying Nancy Pelosi and risking exile to Siberia if she becomes speaker? The Democratic Caucus is deeply divided by age, radicalism, and identity politics. But they do have something to unite them: their hatred of Donald Trump.

Bottom line: Tuesday is all about turnout, and turnout is all about what turns you off.

Charles Lipson is the Peter B. Ritzma Professor of Political Science Emeritus at the University of Chicago, where he is founding director of PIPES, the Program on International Politics, Economics, and Security. He can be reached at charles.lipson@gmail.com.