Those one or two that score a game much lower than everyone else (or score a bad game highly) are called "
outliers." When aggregating game reviews, they can't be ignored, but when deciding whether or not something is "universally acclaimed," they can. Many games, therefore, could be considered "universally acclaimed" despite one or two reviewers who didn't like them (or even didn't like them as much as everyone else).
Another reason why to not base things on review scores: there are
entire genres where the average score is going to be skewed down or up simply because of the genre. You can basically ignore half of the reviews when looking at games like
Pokémon Mystery Dungeon and
Shiren the Wanderer, because with Roguelikes, you're going to get a large number of reviews by people who can't handle the difficulty of a Roguelike and therefore give it a low score without actually rating the game. (Don't believe me? Every single low-scoring review for
Pokémon Mystery Dungeon: Blue Rescue Team on Metacritic complains only about the Roguelike aspects, and
maybe the graphics and dialogue - and Roguelikes aren't known for excelling in either of those.) Similarly, look at 2D Metroidvania games - regardless of the reviewer's opinion, these games get scored highly almost every time. Many reviewers disliked
Castlevania: Order of Ecclesia's comparatively-ridiculous difficulty and stated that they wouldn't recommend it to anyone except die-hard fans because of it - but that didn't stop them from rating it as highly as they did.
While shmups usually fall into the favorable or average ranges, there are some common complaints from reviewers about those as well, all of which seem to miss the point. Look at the (generally favorable) reviews for the GameCube version of
Ikaruga. Every one of them in Metacritic's yellow range complains about the difficulty and/or the length ("only" five levels that are so hard they'll throw even shmup veterans for a loop). The PS2 compilation
Gradius III & IV? Complaints abound that the difficulty is far too high (they're basically arcade-perfect ports, and the arcade version of
Gradius III is considered one of the hardest games ever made in a genre known for its difficulty) and that "it doesn't push the limits of the PS2," as if that has anything to do with whether or not a game is good. (You'll notice this complaint is mysteriously absent with games like
Odin Sphere and
Disgaea, which isn't pushing the limits of the PS3 either.)
Also,
God Hand. My point is, professional critics are definitely not something you and I should be listening to when trying to decide whether or not a game that could go either way is good. (Do listen to them when looking at, say,
The Orange Box, or something else that's obviously good.) Listen to someone you know who's reputable, who had time to play through the whole game (including sidequests, if need be) and who can give you an honest review of it, not someone who had review code for it and five other games they had to play within a week, or an aggregate of many of those reviewers.
(By the way, looking at Metacritic's top DS games,
Mario & Luigi: Bowser's Inside Story is now lower than
Chrono Trigger and tied with
Mario Kart DS and
Flipnote Studio at 91. If more reviews come in, unless they're 100s by Metacritic's system, it'll likely be knocked down even further - I wouldn't be surprised to see it end up tied with the four absolutely excellent games that currently hold scores of 89.)