Originally posted by French Rage
View Post
+ The Presidential vector, same for ever election this year
+ The Seat's vector, same for every election of that seat for every year (within some interval)
+ The Candidates' local vector, particular to that one-time matchup.
So in ME for example you have the Presidential vector of D+6, that Seat vector which is maybe R+3, and then the Collins-Gideon vector which is unknown. Assuming the Candidates' vector is less R than R+3, Gideon should win. But we won't know until we know, Mr. President.
But as the national vector moves, all composite election vectors move in tandem.
So to extend the analogy there are "cliffs" in the Senate prediction that map to the Presidential vector. If Biden wins by 8 we could get 50 Senate seats, by 9 we get 51, but by 10 maybe we get 53 and by 11 maybe 56. The latter varies directly but not linearly with the former.
Comment