I’m confused by this. Are you referring that times where Jack-Sam are linemates and Jack goes for a line change before Sam (or vise versa at the start of the shift)? If so, do you believe this skews the data enough that we can’t come to a conclusion about Sam’s effect on other lines?
Other than those occasions, Sam and Jack would be on different lines. We know that they’ve been separated for 429 minutes (5 on 5) over the last two seasons (not very much). And we also know that he played on the Sobotka line a good amount last year. We know he played on the Mitts line. And we know the amount of time he’s played with each player (e.g. Sheary which is 318 minutes). Considering those players get slaughtered in GF% on a regular basis (especially Sobotka) that seems like enough info to deduce that Sam has had a positive effect on those lines or else his number GF% would be worse than it is away from Jack.
I’l try to clarify things better. Jack/Sam is not a line. Its 2/3rds of one. The other player matters quite a bit to help parse out player impacts.
Take last season and the GF% for Jack’s two most common line combos.
Skinner/Jack/Sam (541mins) ——> GF% - 58% / xGF% - 55%
Skinner/Jack/Pommer (249mins) -> GF% - 59%/ xGF% -58%
Now look at Skinner/Jack away from Sam and Jack/Sam away from Skinner.
Skinner/Jack no Sam (304mins) -> GF% - 52% / xGF% - 54%
Jack/Sam no Skinner (213mins) -> GF% - 41%. / xGF% -> 53%
Its pretty clear from these fleshed out numbers it was Skinner last year, not Sam, that had the big positive impact on Jack’s line combos winner their matchups.
Now compare that to just using a narrowly focused GF% of last season for Jack with/without Sam
Jack with Sam (754mins) ——> GF% 53% / xGF% 54%
Jack without Sam (407mins) -> GF% 45%/ xGF% 51%
That gives the impression Sam was a driver of Jack’s success. Something the fuller picture shows isn’t the case. Jack’s good numbers with Sam are due mostly to the 77% of those minutes Skinner was out there with them.