Pertaining to the problem presented in section 3.3 "team competitions" of the Elo-MMR paper:

[2101.00400] An Elo-like System for Massive Multiplayer Competitions (arxiv.org)

EbTech/Elo-MMR: Skill estimation systems for multiplayer competitions (github.com)

In other words:
Is there a smarter credit-assignment scheme that doesn't motivate players to seek credit at the expense of a team's overall performance?

Or:

Is there a smarter credit-assignment scheme that ensures every individual's incentive is perfectly aligned with winning as a team?

Where:
A smarter credit-assignment scheme precisely estimates individual performance, P_i, in a domain-agnostic way.

"Precisely" is lenient for the purposes of this question. I'll use my discretion, but generally anything better than the current default of equal credit to all teammates qualifies. This means anything that most of the time (>50%, but my discretion pushes this up to maybe at least 67%) assigns better crediting than the default.

Put simply:
Is there a way to (somewhat) determine how much each player contributed to a team's performance, without using game-specific metrics, in a way that can't be exploited and still requires full dedication to winning as a team?

Resolves YES if someone mathematically shows such a scheme exists.

Resolves NO if someone mathematically shows such a scheme can't exist.