Alaska Survey Research shows Peltola ahead; our forecast holds at 26 in 100

Alaska Survey Research polled 1,352 likely voters from September 10 to 12 and released two sets of numbers. In the first, Mary Peltola takes 51.9 percent to Dan Sullivan's 48.1. In the second, drawn from the same sample, Peltola has 48.4 percent and Sullivan 3.7. The payload does not say what distinguishes the two questions, so we will not guess; the second set plainly is not a straight head-to-head, and the first is the one that behaves like one.

Our forecast did not move. The model gives Peltola a 26 in 100 chance of winning the seat and Sullivan 74 percent, the same split as the previous run earlier today. The mean margin shifted from R+6.7 to R+6.8. The 5th-to-95th percentile range runs from R+23.7 to D+10.2, which is another way of saying our model does not treat a Peltola win as unusual so much as unlikely.

That gap between the poll and the forecast is the thing to hold onto. Our average leans several points more Republican than the Alaska Survey Research topline, because it weighs this survey against the rest of the polling and against state fundamentals rather than reading it alone. A single survey showing a near-four-point Peltola advantage is consistent with a race our model still expects Sullivan to win more often than not.

Nationally, our model gives Democrats a 56 percent chance of Senate control, with a mean of 50.7 Democratic seats. Democrats need 51 seats for control and Republicans 50, because the vice president is a Republican and a 50-50 tie goes to the Republicans. Alaska is one of the seats that decides whether that mean rounds the right way.