Every year and every game is different. If we simply did what you suggest we should do, most games should be 10.0s, including Madden, at this point. Standards change each year as they are raised by other games, and year to year progression has to be accounted for. A review score for a reviewer is a reflection of the value a game brings...if a game moves forward but still has major flaws which existed in the prior version (or are troublesome to the final product) then it will have to get dinged. On the same token, a game which was an 9.0 one year releases the next with a roster update...is it still a 9.0 or is its value far less since you can get the same game with slightly older rosters for a much lower price?
It's review theory and there are as many approaches as there are outlets to how to approach scores, some don't even mess with them because people concentrate too highly on a singular number and not enough on the content within the review -- which is a sad reality at OS and most other outlets that choose to use scores for reviews.