Based on this information in this comment, comments should be added to both the rank_fusion_upgrade_downgrade.js and score_fusion_upgrade_downgrade.js files to indicate why queries are accepted/rejected respectively in the 'whenBinariesAreLatestAndFCVIsLastLTS' scenario.

Testing coverage for basic $scoreFusion queries (without $score/$scoreDetails metadata and $minMaxScaler) should also be added to show that these queries would be accepted in the 'whenBinariesAreLatestAndFCVIsLastLTS' scenario (ex: changing  normalization to "none" instead of "minMaxScaler").

If possible, testing coverage for $rankFusion queries (ex: with $scoreDetails metadata set) should be added to show that these queries would be rejected in the 'whenBinariesAreLatestAndFCVIsLastLTS' scenario.

Ultimate goal: achieve parity between testing coverage and documentation (ex: which combinations of queries are rejected/accepted) for both rank_fusion_upgrade_downgrade.js and score_fusion_upgrade_downgrade.js. See if we can create combinations of both $rankFusion and $scoreFusion that are both rejected and accepted based on what FCV-gated concepts they reference that will execute on the shards. Also write clear comments as to why the behavior is why it is in each case.