Is this happening? It’d be a really good thing if this were happening, since the two GOP-controlled toss-ups tonight — Nevada and Arizona — are shaky. Matt Rosendale coming from behind against Jon Tester to flip Montana back to red would almost singlehandedly kill Democrats’ dreams of a Senate takeover regardless of what happens in AZ and NV.

Yesterday the Democratic firm Change Research put Rosendale up 49/46, a major surprise given that he hadn’t led in any poll since late August. Today the Trafalgar Group has him down a single point to Tester. They’re the firm that shocked the world two years ago by spotting Trump’s late movement in the midwest. Now they’re seeing late movement for Rosendale:

Add all of that up and it’s Tester 50.1, Rosendale 49, well within the margin of error. “Why is Montana, of all places, a toss-up to begin with?” you may ask. That’s a … good question. Beating an incumbent is always a challenge but incumbency doesn’t seem to be doing much for Tester’s Democratic colleague, Heidi Heitkamp, across the border in North Dakota. A clue:

A split of 49.6/49.1 for a Republican president in a deep red state like Montana would be underwhelming even in a lukewarm economy. In a great economy it’s bananas. Democrats there feel much more strongly in their opposition to Trump than Republicans feel in their support, too. An obvious possible explanation for Trump’s tepid popularity is a backlash to his trade war among Montana’s farmers, but that becomes less obvious when you remember the state of play in North Dakota. The trade war is affecting them too and yet Kevin Cramer led Heitkamp by nine points in Trafalgar’s poll of the state in late October. POTUS’s approval rating there is a solid 56.8/43.8.

The explanation may be as simple as Tester’s and Heitkamp’s relative approval ratings. Per Morning Consult, as of last month, she stood at 47/43 in North Dakota whereas he stood at 52/37 in Montana, about a five-point difference. That difference may be enough to send one back to the Senate and the other into retirement.

One more note about Montana. The data sites like RCP and FiveThirtyEight give Tester an average lead of between three and five points, but that lead is primarily the work of a single pollster. Harris Interactive and Scott Rasmussen have been polling the state together daily and finding comfortable leads for Tester in the range of 6-7 points. They’re the *only* outfit that sees the race as not-very-competitive, though. A Gravis poll taken in late October found Tester up three; then came the Change Research poll finding Rosendale up three; and now here’s Trafalgar’s poll finding Tester’s lead down to a single point, essentially a toss-up. It may be that Harris/Rasmussen is operating off of flawed turnout assumptions and in fact the race has been tight all along, with Rosendale well positioned to win. We’ll know in a few hours.