We need to identify and clearly (in code, to be reproducible) define a handful of scenarios where multiplanner is able to pick the best plan because it filled a batch during trial.

We need to have the scenarios coded as javascript to be able to reproduce them manually but also be embeddable into synthetic benchmarks test suite.

Ensure both multiplanner was used and EOF was NOT hit from the explain plan.

Code reference: https://github.com/10gen/mongo/blob/eb44e53e46ca4beed28f8f29806771f3e68847df/src/mongo/db/exec/classic/multi_plan.cpp#L435-L438