Currently, there are a whole host of things wrong with game reviewing in general. Whether the complaint decries the length, the lack of depth, or vice-versa, there is something clearly wrong with the approach many game publications take to reviewing games no matter which way you look at it. 

First of all, video games, from a consumer's standpoint are expected to be multiple hour-long pieces of entertainment that have either substantial replayability by their multiplayer, or provide a memorable, or exhilarating single-player experience. Though these are not unreasonable expectations for a $60 product, is it ultimately necessary to read a seven or eight-page review in order to get a sense whether a particular game is for you? Especially how some reviewers take the time to detail the sound, graphics, and the peripheral multiplayer suite clearly thrown in at the last minute, is when I begin to call a certain review or reviewer into question. 

More often than not, game reviewers will find themselves swamped by the sheer amount of games during Fall or so, and will most likely play until the reviewer has gotten an understanding to the rhythm of the game and as to what the game does well and poorly. And because most games don't have the eyebrow-raising twists like those found in something like a Metal Gear Solid game, I typically don't have a problem with reviewers not completing the game they're reviewing. But there is indeed a fine line between concise and outright lazy writing. 

The whole purpose of video game reviews (which might vary from publication to publication), is to let potential players know whether a game is mechanically sound, meaning, that the game won't crash, and that it is generally devoid of bugs and glitches, and if it is something that might interest any potential players, without giving out too many hints about the story and plot. 

What I see not in fact being the purpose of game reviews is the obsession and granularity of review scores, the specific rating of such things as graphics, and sound -- which developers certainly deserve commendation for, but to have those individual ratings influence the total score, is missing the point entirely. I feel the 'total score' indexes have been misconstrued over the years, and should not necessarily indicate a game's quality, but rather, the amount of fun or entertainment a given game has provided. 

Now on to that whole thing about whether game reviewers need to complete the game they're reviewing -- the answer is a resounding no.