Awhile back I said that pollster house effects can mislead about a partisan pollster's performance in a cycle that has a polling bias in the same direction. In a cycle with a polling bias in the other direction, that pollster will be one of the worst performers. Historically average polling bias is close to zero, and inconsistent election to election.
Trafalgar had a great 2016 and 2020. Both years where polls overestimated Dems. In 2018 Rasmussen had an awful year where polls underestimated Dems.
In 2022 polls again underestimated Dems.
This Trafalgar poll is especially bad.
Fwiw while 538 includes both pollsters, the model adjusts for their house effects.
While I think Silver does good polling analysis, I think he shouldn't rate these pollsters as high based on their past performance. Trafalgar especially is very opaque.