In the spirit of the 9 Circles of Scientific Hell, and inspired by the evidence showing that scientific peer reviewers agree only slightly more often than they would by chance, here's a handy tool for randomly generating your review.
Feel free to print it out and throw darts at it, or maybe make a roulette wheel kind of thing, or perhaps a ouija board. It seems to be in widespread use already, so there must be an easy way to use it.
1. The Power of Love: You love this paper! Well, you love the author. Maybe it's a romantic thing, maybe they once saved your ass by lending you their expertise/equipment/data, or maybe they bought you a drink once at a conference. Either way, they're awesome, so their paper must be fine.
2. Bee-in-your-Bonnet: You don't really care about this paper, but you do care, very strongly, about something else which is vaguely related. Many say that you're obsessed by it, though not to your face, because that would start you off talking about it. The problem with this paper is that it doesn't cover your pet idea. If the authors want it published, they'll need to change that, pronto. Major revisions are called for.
3. The Pedant: The paper is atrocious and doesn't deserve to be written on a scrap of toilet paper let alone submitted to this great Journal... in terms of spelling and formatting. Scientifically, you think it's probably pretty good, but it was hard to tell because of the amount of red ink you put all over it. English isn't the author's first language? That's their problem. Isn't that what "minor corrections" are for? No! That's what the bin is for.
4. Cite Me, Me, Me!: The problem with this paper is that it doesn't reference the right previous work... yours. Unless the authors change it to cite everything you've written in the past 10 years, they can get lost. If they do, the paper will be immediately accepted - to reject it would harm your citation count.
5. The Tortoise: You'll review this paper when you get back from holiday. And finished writing your own paper. After that conference. When you've finished your teaching for the year. Maybe. Until you submit your review, the authors are stuck in a horrible limbo, but luckily you're anonymous so they won't know who to send hate mail to.
6. The Cheerleader: This paper is awesome because it supports something that you yourself are about to publish. It's full of methodological holes? Never mind, that will only make your paper better by comparison. It's barely readable? Suggest edits to make it just about comprehensible so people can tell how well it supports you. Then accept a.s.a.p.
7. Wrong End of the Stick: You think you understand this paper, but actually you don't. So your review completely misses the point. When the authors point this out, you have two options: a) blame the paper for being confusing, and chuck it out or b) decide the whole thing is much too complicated to spend time over, and accept it.
8. The Perfect Reviewer: You are an intelligent, informed expert, new enough to the field that you have no axe to grind, and you take the time to read the paper fully, and return a constructive, perceptive review within a couple of weeks. Well done. Unfortunately, there are 1 or 2 other reviewers, and there's only a 1 in 8 chance they'll be like you...