I have several grievances/ points of advice:

1. Be succinct.
Internet reviews (not just game reviews, either!) are generally snoozefests, because the writers are not confined by space budgeting and other limitations placed on print media. A good review will not contain a complete itinerary of every event in the game, nor will it contain irrelevant and self indulgent analysis. Roger Ebert doesn't spend three pages reviewing a movie, but you still get a good idea of what a movie is about, if it was successfully executed, a few high points, and a general feeling of whether or not the movie will be compatible with your tastes.

2. Don't bother trying to be funny, unless you actually are.
I know the relative success of "Penny Arcade" tends to defy the above dictum, but they are filling an important void left by the deaths of Charles Shultz and Bil Keane and are exempt from the rest of this paragraph. That said, don't try to spice up your bland review with humorous asides or clever in-jokes unless you truly have a gift in this department.

Some corollaries to #2:

2a. Fuck off with sarcasm.
Bleh. I don't know when the sarcasm trend spun wildly out of control, but it must come to a stop. First and foremost, the amount of sarcasm present in recent reviews is just...childish and stupid. It's unimaginative. It's formulaic. Not only is it formulaic ("I'd rather [insert harsh, unpleasant activity] than play this game another  [insert small increment of time]!!111"), but these Mad Libs-style insertions are just embarrassing in their simplicity and limited vision. "I'd rather stick my hand in a blender than play this game another microsecond!!(111)" If you (as the reader) can put aside the cringing embarrassment that unsophisticated sarcasm forces upon you, you are still left with a diluent to the review. How much sarcasm must you wade through to get useful information? How do you know that information isn't tainted by the author's attempts at humor? And why should you trust a person who is so thoroughly unsuccessful in their aims (write good review/be funny/appear knowledgeable), yet so proud of their failed product that they display it for all to see in spite of everything?

2b. Don't rehash other peoples' humor.
Have some pride in yourself. Good grief. It's bad enough that people insist on rehashing "Penny Arcade" text. It's worse to other people cramming in things that were stolen from equally unfunny forum posters. No, I am not referring to anyone at the Retrogaming Roundtable, but I've seen it quite a bit elsewhere.

3. Drop the pseudonyms and cartoon images.
I don't know if Gamefan or Gamepro first started this (probably Gamepro), but it's gotten completely out of hand, especially on the internet. Perhaps if a person had their own name and face next to a review, they would be more careful about their defiance of rules 1 & 2, as well as the overall quality of the review. Come to think of it, no wonder Gamepro instituted this idea of virtual anonymity. They were the originators of shitty, meaningless reviews. I mean that quite seriously.

4. Don't be obvious when trying to show off.
This could be taken as a good rule of thumb, in a variety of manners. What I refer to specifically is the increasing tendency of amateur (and some professional) writers to lace their text with obscure references to "old school" games. We get it. Again, it's more than just annoyance with the videogame analogue of name dropping. It also dilutes the review, and throws doubt upon the integrity of the review as a whole. "This guy obviously wants me to think he's cool, so why wouldn't he just use a game review as a vehicle to spew obscure trivia within?"

5. Use absolute descriptives.
Telling me that Game X is a mixture of [obscure game 1], [obscure game 2], [generic game 1], and [generic game 2] tells me nothing. Describe a game within the confines of itself, as much as possible. If it's a legitimate clone of something else, that's acceptable, but if not, describe it in "unrelated terminology". Just remember Die Hard. Speed was described as "Die Hard in a Bus!", Air Force One was "Die Hard on a plane!", there was a train movie, and plenty of others. What if a person never actually saw Die Hard? How about descriptions that don't rely on a person's prior knowledge of prior games?

6. Be objective.
I know, I know. It's an obvious idea, but one that rarely makes it to execution. I'm not talking about system loyalty, or anything along those lines. More to the point, a review should never start out by telling you what a "fucked up piece of shit" it is. Not if it's supposed to be informative and to be used for purchase decisions (ostensibly the reason for any review).A review should be able to cover the positives and negatives, and their reasons, followed by a summary impression at the end. A "fucked up piece of shit" style review will only give further reasons (typically sarcastic..) to reinforce the opening statement. Ranting and raving about what a terrible game it is does nothing to inform the reader, and only detracts from the review's credibility and usefulness. Give the reader the high and low points, and let the reader decide based on that information. Maybe a thumbs up or an overall score, but be fair. It's too easy to go "xtreme" and exaggerate everything. Maybe a balanced review isn't as shocking or tart, but it's a more useful tool.

7. Don't bother with fucking decimal points in the score.
Incredibly pointless tendency. It's not going to influence anyone to score in tenths of points. Why not hundredths or thousandths? The only reason why this practice is even around (aside from overzealous, obsessive stat maniacs) is so that reviewers won't have to answer for why game A and game B both got the same score that month.Go with a common sense approach: Poor, fair, average, good, excellent.