'After going through multiple scenarios, we believe a game rating, in most cases, is not a reliable tool for predicting game sales,' they said. 'There are isolated examples of strong correlation, but they are just that - isolated. We believe a naked game rating without context is largely useless. The more specific the comparison (controlling for as many variables as possible), the more likely a statistically significant correlation may exist. However, the more specific the sample, the less useful the hypothesis becomes.'
I always wondered about this. I suspect that with a little manipulation it is possible to show that game rankings do correlate to sales. Jason Hall went out on a limb and said he plans to tie Warner Bros. licensing fees to game review results.
But I've personally seen too many games with low ratings sell well and more often just the opposite - games that reviewers think are really cool and don't sell squat.