The problem with the ratings system employed by most reviewers these days is that they stand over too large-a-period of time. If the scores were broken down into years, for example 2008 compared to 2007, then a 10/10 would be possible for _that_ year. Technically speaking Pong would be a 10/10 and thus perfect, which it isn't, because nothing came before it and automatically it was the best that could be, and thus every game released after that is compared to it as it's the first 10/10.
I don't see Pong in any 100 games of all times list crap that gets spewed out every so often - unless they're trying for the old cop-out of "it was important and thus gets the No. 1 spot" bull crap that they use to make themselves feel intelligent.
Arguably a great game is always a great game; Legend of Zelda: Ocarina of Time and Portal I'm looking at you. However comparing them to the games released 20 years from now, it's a fair call to assume technology will allow for new and interesting things in games and should, if handled correctly, produce better games that surpass what we have now. Half-Life 2 is superior to Half-Life in almost all regards, yet if Half-Life received a 10/10 what should Half-Life 2 get?
Then again I've always felt that using a number system to relate the quality of a game to be lacking - used now-a-days by "official" magazines and websites that are funded by advertisements paid for by the gaming companies producing the games they sport advertisements, competitions, give-a-ways and promotions for, which just happen to be the very games they've dedicated 15 pages of jargon to and it just happens to be given a 10/10.
Oblivion was released. And until the next Elder Scrolls game, I've got my 10/10.