If you chaps can prove that you get different scores to those you see in reviews with identical hardware components, software versions, and test configuration then you have a point - and should challenge the reviewer directly, as they do sometimes make mistakes. Otherwise, the simpler and likelier explanation is that without replicating their set-up and testing exactly, you're bound to get different scores.
The above is one of the reasons I do have so many components, so I can test under identical conditions as I have little faith left in some reviewers.