IDK about the first set, but the second set isn't a good comparison. WI has a dem incumbent running.
A polarizing incumbent with very mixed fav/approval numbers, yes. Is a 44Fav/42Unfav spread (in the latest Marquette which showed him leading his R opponents) the kind of split that will make a WI Democrat defy partisanship to the point that it justifies a Lean D rating on a night when (as this model also suggests) Republicans are winning the HPV by
six points? Also note that this "model" gives Steve Sisolak a similar 64% chance of winning reelection, and I don’t think there’s anyone on here who would argue that Sisolak is stronger than a generic D or benefits from actual crossover appeal. So what’s the reason for these estimates? As you can probably guess, it’s the fact that Silver's model (and their House forecast confirms this) predictably places way too much weight on "incumbency" (or lack thereof) — this is also why I adduced the Meijer/James races as another ludicrous example (and there are more) illustrating that same pattern.
OR is an open seat with a deeply unpopular governor (and the house majority leader is the dem nominee - at least somewhat tied to that governor), there is also an independent who has strong financial backing running named Betsy Johnson. She is a democrat turned independent who will take a lot of votes - though it's unclear from who.
I’m aware of that independent, but as you noted, it’s very unclear to what extent she will actually hurt Kotek. There’s also very little reason to believe that her current numbers are sustainable given the history of third-party/independent candidates collapsing during GE campaigns and/or underperforming their polling numbers. In fact, there’s a case to be made that the prospect of a Republican governor will eventually scare enough D Johnson supporters into voting for Kotek. Besides, the Democratic base in the Pacific NW has been extremely reliable, and the state has not elected a Republican governor since 1982 (longest-running winning streak for Democrats), whereas Democrats only won WI-GOV 2018/WI-PRES 2020 by a point under extremely favorable conditions.
She is backed by Phil Knight (NIKE) and as I am visiting my family in Oregon since early June, Betsy Johnson has absolutely bombarded the airwaves with ads, with nothing from Tina Kotek or Drazan so far (I'm in Salem). So basically, it is very uncertain.
Yes, some "nuance" is certainly warranted (and I agree that there is a lot of uncertainty regarding the OR race), but not when it misses the big picture — OR is a solid Democratic state with an extremely high D floor and no recent history of having elected opposite-party governors (unlike KS in 2018) and WI is a state with an extremely high R floor that leans slightly Republican and would elect any halfway competent Republican over a generic Democrat in a R+6 year.
I could understand "Toss-up" for WI and "Lean D" for OR, but not this. On their "Classic" forecast ("What Election Day looks like based on polls, fundraising, past voting patterns and more"), they even rate WI-GOV "Likely D" (giving Evers a 75% chance of winning), NV-GOV "Lean D" (giving Sisolak a 64% chance of winning) and OR-GOV a "Toss-up" (with each candidate having a 50% chance of winning) — all of this on a R+6 night. There’s nothing rational about this, it’s just absolute nonsense.