I want to talk a bit about Partisan Voting Index, because there's a bit of a subtlety to how they're computed that seems to be causing a lot of confusion, and I don't think the corrections being given in this thread are actually all that helpful and may even be harmful to understanding.
When we think about election results, most of us think in terms of margins, i.e., the difference between the candidates. If an election results in:
Alice: 51%
Bob: 49%
We think of Alice as winning by 2 points.
This, however, is not what PVI refers to. PVI does three things. First, it factors out all third party votes. Second, it compares local results to national results. Most importantly for this discussion, it's based on vote shares, not vote margins. In the example above there's only two candidates, so we don't need to worry about factoring out third parties, but instead of comparing Alice's votes to Bob's, we compare her votes to 50%. Adapting the language of PVI we would say this result is "Alice+1," not "Alice+2."
When people say "Montana is R+20 but this election was only R+6" what they mean is Trump won by 20 points and Gianforte only won by 6. This is an improper use of "R+X" but it is at least an apples to apples comparison.
When Cook says Montana is R+11 what they mean is that, in a hypothetical election where the vote is evenly split between Democrats and Republicans, we would expect the Republican to get 61% of the two party vote to the Democrat's 39%, that is, pretty close to a 20 point margin assuming a "normal" third party vote. When you say "Montana is R+11 not R+20" you're correcting the first statement but not the second, which encourages the reader to compare vote shares to vote margins, an apples to oranges comparison.
If we want to translate this election into the language of PVI, we shouldn't call the result R+6. Rather, we compute that Gianforte got about 53% of the two-party vote (53.2% to be more precise) so the result was R+3.
That's quite the underperformance in an R+11 district. Of course there are a lot of cautions in order when dealing with PVI due to its inherent limitations (I could easily write another long post about cautions one should take when interpreting PVI), but this is consistent with evidence of a strong political environment for Democrats.
