The mistake here is in assuming reviewers (professional or otherwise) have more game-playing experience than anyone else. They don't. And, while experience can be an issue, there's something else to consider:
Some games get worse over time. Some games get better over time. And when you play a game for the very first time, it's impossible to tell which category it will fall into.
Contemporary reviews are often written as a sort of "time capsule", capturing the writer's enjoyment at a particular singular point. Because of that, it's hard to gauge which games will become classics, and which ones will fade in power -- because the reviewer hasn't let enough time pass to know any better. That's why it's so easy to point and laugh at old magazine reviews.
Just look at the Nights/SMB64 phenomenon... when they came out, both were highly rated, but SMB64 came out on top. And the reasons given made a lot of sense (one being that it was potentially the start of a brand new genre, and thus more important). But now a lot of people recall Nights more fondly, because it has never been replicated, and also because it's just much easier to pick up and play than SMB64.
After playing a game for the first time, a reviewer can make educated guesses, but they have no way to *know* a game's (1) genre impact, or (2) enduring replayability. So if someone "re-reviews" a game, taking those two things into account, it's understandable that the score would change. It doesn't mean they were stupid or hypocritical; it just means that new information is being considered, which wasn't available at the time of original review.
And that can happen no matter what game is being reviewed. Unless you've let the game sit and fester in your mind for a year (or more), you may very well find yourself looking back and wondering what the hell you were thinking by giving it a 9, 10, or whatever.
