Curious about how they’d approach this.

In practice, even with the same system specs and game settings, if you run through a game in a slightly different manner than someone else, it might paint a different performance picture for you than it does for them. The more a game allows free roaming, the more variation there will be in results. I doubt they’ll ask everyone to run a benchmark for each game (and to further that, not every game has a benchmark capabilities built in to begin with).

At least they have the benefit of potentially having huge data sets on their hands, so things would probably even out.