You have to think of it through the eyes of the committee. Their tools give them limited data to go on in order to make decisions between the auto-bid seeds. What they look for first is Quad1 and Quad2 wins. At the 14/15/16, merely having any can be a differentiator. And in this case, Harvard would have 1 (at Penn, which would be ~125 if it wins out and loses to Harvard) and Penn would have zero (again Dayton could hang on, though). Penn beating Harvard in the IL Tourney final wouldn’t help Penn on that front, because it would only be a Quad3 game for them.

Then, they look at RPI and SOS/NCSOS. Harvard would be slightly better in the RPI as champ than Penn (~100 to ~110). It would be WAY better in SOS and NCSOS (right now around 150 in former and Top 100 in the latter). I think they’d both be 15s right now, regardless.

From an actual resume perspective (Wins Against Bubble or other strength of record measures), Penn has a much stronger resume than Harvard (and actually could be a borderline 13-seed case). There will be an SOR metric on the team sheets this year, but it’s unlikely that the committee will pay it much heed.