For much of this graphic (before the “70 days until Election Day” mark) Sinema was looking solid. She consistently had high single-digit leads across a range of polls and was facing no major opposition from within her party. McSally, on the other hand, was facing off against Kelli Ward (a Tea Party candidate who likely would have had huge problems in the general election) and Joe Arpaio (who was convicted of criminal contempt but got a pardon from President Trump).

After winning the primary, McSally’s win probability jumped, declined a little bit (partially due to the timing of various polls) and then started to increase again. In our full model, McSally’s chances are about twice what they were on Labor Day.

There are a few possible explanations for why the model (which is primarily driven by the polls) moved in this way.

The easiest explanation for the early movement is that the GOP consolidated behind McSally after her primary win. Some Republicans might have been holding out for Ward or Arpaio, but now are willing to pick McSally over Sinema.