The steelmanned version of the argument for playing Stastny on the wing would be to load up 3 lines and allocate the maximum amount of icetime for your best players and giving the 4th line 6-7 mins a night like Maurice has done in the past.
I looked at EvolvingWild's GARs for Jets forwards and did a calculation of how many GARs the team would produce under
Scenario a) Put 4 Cs on 4 lines with normal'ish icetime
Scenario b) Stastny on the wing with 3 loaded up lines and minimal icetime for the 4th line
This is a bit simplistic as it doesn't take into account player chemistry etc but gives an idea. Lines were taken from practice today and slightly modified for each scenario to make it work
Under this the "loaded up" scenario b) results in the Jets getting about 3.37 more GARs over 48 games.
And this is using EvolvingWild's metric of player value. Maurice most likely believes that the difference in value on ice of Scheifele over Perreault is more than the 0.122 GAR/60 than what EvolvingWild's model says. So in the eyes of the coaching staff the difference is even more magnified.
Even if we put Harkins on the 4th line instead of Thompson, the loaded up lines with Stastny on the wing comes out ahead by 1.22 goals. And EvolvingWild's model thinks Harkins is .339 GAR/60 player and is possibly overestimating him due to small sample size and sheltered deployment:
Given the horses we have I don't mind rolling 3 lines and shoving the 4th line under 8 mins a night.