One of my goals this election cycle is to keep my readers from being surprised—that is, to make sure that you hear about all the potentially competitive races, have a gauge on exactly how competitive they are and don’t feel blindsided by a foreseeable event.

To that end, I’ve written up a couple paragraphs about each of the competitive (or potentially competitive) Senate races that you probably haven’t read too much about yet. Note that I’m not talking about Texas or Florida (which have gotten a lot of coverage) or races that you might have recently read a lot about here or elsewhere (e.g. you can read my pieces on Arizona and Tennessee, Andrew Egger’s coverage of Missouri or John McCormack’s latest dispatch from Nevada if you want to learn about those races). The goal here is to help fill some gaps and cover races that have arguably gotten a bit less attention than they deserve.

Indiana: The Most Underrated Race of the Cycle

Indiana is probably the most competitive yet still under-the-radar race of the cycle. Polling is sparse in Indiana (due partially to the state’s laws), but the most recent polls have shown Mike Braun, the Republican candidate, slightly ahead of Joe Donnelly, the incumbent Democrat. SwingSeat (THE WEEKLY STANDARD’s stats-driven Senate Forecast) looked at all of that data (along with older polls that were better for Donnelly and other data sources) and projected that Donnelly had a roughly 60 percent win probability. That signals a very competitive race where Donnelly has an advantage but either candidate could win.

Donnelly has been trying to paint himself as a more conservative Democrat (he literally quoted Ronald Reagan and defended ICE in a recent ad) and often votes for Trump’s priorities, but he’s still a Democrat. In a red state like Indiana, that counts. And if pollsters choose not to keep measuring public opinion in this race, we could get blindsided by a late-breaking Braun surge (remember how late the GOP surge happened in the state’s 2016 race).

Montana: A Weird State Where We Need More Polls

Democratic senator Jon Tester is in a weird position: In some ways he’s a bad fit for his state and in other ways he’s a great fit. The top-line numbers suggest Tester should be in rough shape. Montana voted for Trump by a 20-point margin in 2016, and Tester is a Democrat. Moreover, Tester isn’t as moderate as someone like Joe Manchin. His DW-NOMINATE score puts him only a bit to the right of the average Democrat, and although he holds some conservative stances (e.g. he’s generally pro-gun) he votes against President Trump most of the time.

But in other ways he’s a good fit for his state. Montana is deeply Republican, but there’s a sort of latent union-Democratic sentiment there, and Republicans have been willing to elect Democrats to other statewide offices. (Steve Bullock, the state’s Democratic governor, won re-election in 2016 while Trump won the state in a landslide). Moreover, Tester is a cultural fit for the state. He wears cowboy boots, has slaughtered his own food, and lost three fingers in a meat grinder.

These competing factors make it hard to know who is going to win without the help of polling. And while the polling suggests that Tester is ahead (he leads by 4.5 points in the RCP average), my model suggests that there’s a real possibility for an upset (Matt Rosendale, the Republican, has a 16-percent win probability in my model, which discounts some of the earlier polling and uses non-polling factors). This is, like Indiana, a race where we’ll need polling data to gauge where last-minute undecided voters are heading.

West Virginia: Almost Heaven, Fundamentals

Right now, the polling in West Virginia looks solid for Democratic senator Joe Manchin, but the model still leaves some chance for an upset.

Manchin is one of the only senators who might still credibly claim to be a moderate. He’s voted with President Trump about 60 percent of the time, and he’s the only Democratic senator who voted for Brett Kavanaugh’s confirmation to the Supreme Court. In fact, Manchin voted with Trump on two of the arguably most important social/cultural votes of the last two years (the Supreme Court confirmations of Kavanaugh and Neil Gorsuch) but voted with his party on two of the most visible economic/policy votes (the Affordable Care Act/Obamacare repeal/replace effort and the tax reform bill). That gives you a sense of the line that Manchin is trying to walk: pleasing an electorate that feels culturally alienated from the Democratic party but that still has union-style/blue collar instincts on economics.

Polling suggests that Manchin is in very good shape, but the fundamentals (most notably, the underlying partisanship of the state) make SwingSeat think that Republican Patrick Morrisey isn’t out of this race. In about 10 percent of my model’s simulations, some combination of polling error and late movement allow Morrisey to close the gap and take the Senate seat. A scenario like that isn’t implausible—if undecided voters break toward Morrisey or voters decide that their loyalty to Trump outweighs their loyalty to Manchin, then he could still win. But Manchin is still the right bet for now.

New Jersey: Corruption Hurts! Which is a Good Thing!

In a year like this (where public opinion favors the Democrats), a New Jersey Senate race should be a non-event. But Democratic senator Bob Menendez seems pretty corrupt, and voters are noticing.

I re-ran an older version of the model that only looks at head-to-head polling, and the polls suggest that Menendez has a roughly 90 percent chance of holding onto his seat. The full model is much more bullish on Menendez, giving him a roughly 98 percent win probability. Menendez is the clear favorite in both models, but the fact that it’s not a 100 percent win probability in the polling-based model is notable. It suggests that Republican Bob Hugin has an outside shot of winning (though I suspect that the blue-ness of New Jersey will reassert itself and that Menendez will win).

Minnesota: GOP Fingers Crossed for Ticket Splitting

I won’t spend much time on Minnesota because I personally believe an upset here is highly unlikely. Democratic senator Tina Smith (appointed to replace Al Franken) has been posting good-but-not-election-ending numbers against her Republican opponent, Karin Housley. It’s possible that Housley pulls off a win, but that seems unlikely given that Amy Klobuchar, the state’s other senator, is also up for re-election and has been posting eye-popping numbers this whole cycle. There may be some Housley-Klobuchar voters, but it would take a boatload of them to keep Smith from winning this one.

Mississippi: The Democratic Hail Mary

On Election Day, Mississippi is going to hold a non-partisan primary where every Republican and Democratic candidate runs in the same race (this is a special race triggered by Thad Cochran’s retirement—Roger Wicker looks solid in his regularly scheduled re-election bid). If no candidate gets above 50 percent, then the top two head to a runoff later in the month.

Right now, it looks like the GOP-appointed incumbent, Cindy Hyde-Smith, and Democrat Mike Espy will make it to the runoff. Recent polling from Marist suggests that Hyde-Smith will win that race by a solid margin and keep the seat in GOP hands.

But there are two ways this could go wrong for the GOP. First, the GOP could have a bad Election Day and only end up getting to 50 seats—meaning that the Senate majority would come down to Mississippi, where Democrats could knock that number down to 49. In that case, we could imagine supercharged Democratic turnout (one of the factors that allowed Doug Jones, a Democrat, to win a Senate race in Alabama, a demographically and politically similar state, last year) to help carry Espy over the finish line. Second, it’s possible, though unlikely, that Republicans push Chris McDaniel, a highly problematic Tea Party Republican, through to the runoff election. That would give Espy a much higher win probability.