"useful to know" isn't the same as being a quality factor. same with price. other media reviews usually mention length as well. the problem is more that "replay value" is somehow seen as a "good thing" in itself.
a bad game you can play in a dozen different ways is still ultimately a bad game and probably doesnt warrant multiple playthroughs (or even one) anyway. on the other side of things, ghost trick (for example) is a great game but has zero "replay value" because being a linear puzzle/adventure game, that's just not part of its design. should developers really be punished for making linear gams?
You could say all of this about "Graphics" or "Music". Replay Value is sometimes given higher weight than other metrics, but if you score each metric separately the person reading the review can just juggle the weights to their own preferences instead.
If Ghost Trick was exactly the same game, but offered more incentive to give it a second time around, it would be a better game than if it was just the way it is now.
It's not about punishing people who make linear games, but rewarding people who make games with more depth and breadth of content.
It pretty much goes without saying that a maximum score in every field for a game that simply isn't any fun to play doesn't change the fact the game isn't any fun to play.