The Inquirer says it has seen some documents which outline how AMD wants to manage and control the Athlon FX/Athlon 64 review process. The exact phrase used is “conduct all reviews directly with reviewers.”
This ought to bother you, a lot. You’re going to inevitably end up with a biased picture with Big Brother one way or another sitting on your shoulder.
In a world where integrity mattered, AMD would be universally told to take their systems and shove it. But I fantasize.
The sad reality is many, if not most people, both reviewers and reviewees, don’t want the truth. Truth can hurt. Truth doesn’t respect your biases and prejudices.
What many, if not most people want is to be told what they want to hear. They want their biases and prejudices reinforced. They want to be told, “See, I was right,” and when they are, they’re not going to look too closely at the details.
How To Be Manipulated, A Little At A Time
What we are beginning to see and will continue to see are subtle manipulations, slants and biases in the reviews. Since the average person really doesn’t read with 100% accuracy or memory or comprehensions, those manipulations will take advantage of that.
The idea is to create this generalized warm, fuzzy feeling of AMD being back on top again, and where do I sign up?
The reality is an Athlon FX running at 2.2GHz will probably do a little better overall (somewhat to rather better in some things, somewhat to rather worse in others) than a 3.2GHz PIV, and a little bit better than that for games, and an Athlon 64 running at 2.2GHz will do a little bit worse, and not that little bit better for games (single-channel memory, you understand).
That’s not bad at all, and better than we thought even a month or so ago, but it’s hardly enough to cause a feeding frenzy.
So good isn’t good enough. Time to get biased.
It can be as simple and crude as using mostly or entirely benchmarks that unduly favor the Hammer setup, or disfavor the PIV.
You’ll likely see a lot of Athlon FX comparisons, and not too many socket 754 comparisons. This will leave the average gamer with that warm fuzzy feeling when looking at socket 754 systems when it really shouldn’t.
Rather than compare a Hammer to a 3.2GHz PIV, you’ll see comparisons with a 3.0C processor.
It’s going to be a matter of a few percentage points here, a handful there, but when you have a situation where:
a) in reality, it’s close and
b) the two processors being compared don’t really consistently compare well against each other (each processor has pronounced strengths and weaknesses compared to the other)
. . . it’s easy to slant a case heavily in favor of either.
The Radeon 9700 Experience
A while back, ATI pioneered this sort of approach. They took control of the review process, and blatantly manipulated the initial statistics to create a very, very good first impression.
At the time, I rather strongly objected to this (well, to be honest, I waz probably closer to stark raving mad at the time. :)), and wrote a lot about it, making all sort of points that in the end proved to be correct.
I got some hate mail for it, but what really blew my mind was the almost universal incomprehension of what I was talking about.
In a nutshell, ATI was using unrepresentative testing which gave the impression that the Radeon 9700 was twice as fast as a GeForce Ti4600. Later testing showed the difference was more like 25%.
To me, that was outrageous, but to practically anyone outside of a nVidia diehard, it was like, “You see, it is faster, what’s your problem?”
If the difference between 25% and 100% is but a trifling detail in the minds of the audience, how likely is it that people are going to catch is a shift of a mere 10-15%, a few points here, a few there? Especially when a lot of people want to believe it.
The Upcoming Spectacle
What we’ll see over the next few weeks is a rollout as managed as a Super Bowl halftime show. You’ll have puppets on strings saying pretty much whatever AMD wants them to say. Some puppets will resist a little and try to say things between the lines in the hope that the Big Brother censor won’t catch it.
That kind of approach might have been quite understandable for a dissident in the Soviet Union, but for the most part, this is an American company dealing with (mostly) American or at least Western reviewers.
And what’s the goal of all this control? To make a product which probably is 3-8% better then its competition look 10% to 15% better than that, and to make you think that another product which actually is a little worse is better, too.
Do you know what? They’ll probably not only get away with it, people will be happy about it. Their side is winning again. End of story.
The victims are just too willing.
Often enough, I get emails that essentially say, “You’re not cheerleading for AMD, why are you so biased?” and the writer is so clueless as to how unintentionally hilarious he is or what the word “bias,” much less “objectivity” actually means. I guarantee you I’ll get more after this article.
Be smarter than that.