yes.
i have been arguing for years that statewide elections should always be modeled as independent from each other and that the current status quo - to model them as dependent events and use conditional probabilities - is a massive error. i have also identified this error as a major cause of polling misses.
somebody might come by here and think i made a mistake. i would aggressively argue otherwise - i think the consensus position is a mistake.
i have a pretty good track record. i make some mistakes. we'll see if i'm right (again) or not soon enough.