There's no real need to open comparisons to "esports" broadly. Which ones? Some work with consistent teams and would have ratings in accordance with that, but that's different from the individual ratings, as I said above. For rating individuals in team games where your partner randomly changes from game to game, there's a well-designed system ready to go, which would work much more accurately if it just assessed the team's chance of winning a game--and the rating adjustment that comes with it--according to the strength of the team.
(For finer tuning, we could explore how disparate teammates really affect the play. I believe the weakest player at the table has much more potential to bring the team down than the strongest player has to bring it up. (In the example I gave of 150 & 450 vs 300 & 300, this would mean the 300s are actually favored.) But even just a linear first approximation, using the teammates' average rating, would be a tremendous improvement over the status quo.)
(For finer tuning, we could explore how disparate teammates really affect the play. I believe the weakest player at the table has much more potential to bring the team down than the strongest player has to bring it up. (In the example I gave of 150 & 450 vs 300 & 300, this would mean the 300s are actually favored.) But even just a linear first approximation, using the teammates' average rating, would be a tremendous improvement over the status quo.)