Our AK-1 model returns Democrat Eric Hafner to a 0% chance after four days

Our model now gives Democrat Eric Hafner a 0% chance of winning Alaska's at-large House seat. In the run of October 6 at 12:32 AM ET it gave him a 9 in 100 chance. That is a move of 8.9 points toward Republican Nick Begich III over four days. Our model now gives Begich a 100% chance, up from 91%.

The shape of the forecast moved with it. Our model's mean margin is now R+23.9, compared with R+11.1 in the earlier run. Its 5th-to-95th percentile range now runs from R+37.4 to R+10.3. Four days ago it ran from R+25.0 to D+2.6, so the Democratic end of the range has dropped out.

The polling in the payload comes from two Alaska Survey Research ballot tests of 1,512 likely voters, fielded October 1-3. In the two-way version, Begich takes 52.7% and independent Bill Hill takes 47.3%. In the four-way version, Begich takes 49.8%, Hill 39.9%, Hafner 6.9%, and other candidates 3.4%. The payload does not say when these polls entered our average, so it does not establish whether they account for the change between the two runs.

The district-level result also feeds our national House estimate, which gives Democrats a 99% chance of winning a majority of 218 seats. That figure is marginally firmer than it should be. Our simulation correlates error nationally, by census division, by state and by district, but it lacks a term linking demographically similar districts in different places, because we have no cluster data to build one.