Sales figures deduced from pre ordering and first week after release. Sales figures are not known directly, they are deduced. That is what I wrote.
You wrote "deduced from first week and sales", this confused me. But apropos: if you're going down this route you could take VGChartz (which no one in the industry takes seriously) or NPD, since multiple gaming networks do have an account with NPD (no idea why you wouldn't mention that since it seem pretty important to your theory). The sources Telegraph claims are either from GTA IV's publisher (this happens a lot, as sneak-PR) or one of those experts that bandies about numbers that are rarely right. Even NPD, who do it for a living, are at best guesstimates, and focus only on North America, which is making them increasingly irrelevant (at least they decided to encompass digital download a while back).
Regardless, games that are expected to do bad sometimes do well, and vice versa. Let's take Killzone. Hyped as a "Halo killer" in many previews, it was eaten up once released, for 70% average on Metacritic, but still sold enough to spawn 3 sequels so far. The reviews for Killzone have since matched up more closely with the franchise's popularity, in the 80s to 90s range. The way this went doesn't match up with your theory.
But here's the thing, you're offering a conspiracy theory and I'm asking you "based on what"? Have you heard this from a professional game reviewer? I know a lot of them since, uh, I am one, but no one's ever mentioned this gig to me. Meanwhile, I'm offering you a much simpler explanation: high sales and high reviews tend to coincide because mainstream reviewers have the same tastes and demands as mainstream gamers, the audience they're writing for.