Transparency in Tech Reviews, and How to Read Between the Lines.

Semi-Ranting Part

It irks me these days. Any time someone asks about a specific computer part, if I can't find it on one of my tried and trued websites, I groan a little inside. There's multiple ways a benchmark can be bad, and the worst is a vague review to make a mediocre or bad product look good.

Why do I hate these worse than bias? Well, frankly, because bias is usually pretty easy to spot. A site or reviewer that looks desperately for a way to make something look better than it's competitors is fairly obvious, since direct comparison is required. Any time there's direct comparison, you can apply a little bit of critical thinking and elementary math to see the discrepancy for yourself.

The problem with a vague review is that some people do accurate vague reviews, and some people use a vague review to obfuscate a desire to continue getting free stuff by finding a way to make the facts look better. Now an accurate vague review, while frustrating, can at least be useful, for finding potential issues that will let you make an informed decision on something. When I was looking into purchasing the Antec Lanboy Air case, I ran into several of these. While it drove me completely insane to not see much information on the thermal and acoustics of the case, those weren't my main concern. I was worried about build quality, since the case is a veritable treasure trove of gimmicks.

Unfortunately, you aren't always trying to find stuff that a vague review will provide well, and when this happens, it can be downright annoying. By nature, a vague review or benchmark is never transparent. There won't be enough comparison, or the testing methodology will either be vague, variable, or just downright not discussed. When this happens, you'll be stuck looking at a chart, scratching your head, and trying to make an educated guess.

The problem, of course, is that a lot of "reviewers" think that the way to keep getting free or discounted stuff to review is to find a way to make the stuff look good. But if nobody reads your reviews, and nobody links to it, are you really doing yourself or the manufacturers a favor? No, of course not. You're just making yourself look like an idiot. If you aren't going to compare something to a proper baseline, you might as well not benchmark or review it at all.

Now obviously, the odds of any of the people I'm complaining about reading this are slim to none, so what value does this have? None, really, aside from possibly educating people about how to properly use several vague reviews to get more data than people were trying to give out.

Educational Part

So, here's to the tips. If you can't find a half-decent review of a product, but do find 3-4... shall we say less than useful ones, there's ways to drag the hidden information out of them.

1: Look for common compliments. Certain words, like innovative, really just mean "non-standard", which also frequently means "we couldn't stop swearing at it". If there's no good pictures of whatever is "innovative" you know it's a damn nuisance. "Lots of fans" or "Amazing airflow" mean "Poor mans surround sound for the movie Twister." Get the idea?

2: Look At Picture Angles. If the pictures avoid certain angles like the plague, look at what you aren't able to see. No picture behind the motherboard tray? Cable management blows. No picture of the Hard Drive bays? Good chance they might be either noisy or inconvenient. Anything they don't want to give you pictures of from a rational angle is either inconvenient or badly engineered.

3: Look for what isn't mentioned. This one only works if you read a decent number of good reviews as well. Anything that normally gets discussed in a good review that doesn't in a vague review is probably an issue. Good examples would be acoustics or thermals on a GPU, cable management or HDD mounting in a case, noise quality or OC'ed performance on a CPU cooler.

Basically, it's all about reading between the lines and having a base knowledge of what sort of things a review or benchmark should have for a given component. If you find a site that does good reviews, bookmark them. Use their search feature. And don't just throw away a vague review. Find several of them, keep them open in tabs next to each other, and apply these tests to all of them at once. If the hidden stuff is consistent, you know that it's not an oversight, and you can extrapolate what's wrong with the product.