Wow, think you're rather digging yourself in deeper there Raven, firstly the problem you showed with Anandtech who I've question in regards to accuracy for about a year now as they've gone WAY downhill.
Is they DID retest cards in that test, and you just have to be careful as too see which ones. Oh yes, the 480gtx 49.9 in the first test , 68.7 in the newer one, which also moves it from WAY behind the 5870 to literally a nose hair infront of it.
So they retested EVERY Nvidia card with newer better drivers on an optimized game(likely with the drivers that cause actual stuttering in gameplay but do great in benchmarks).
Yet they used old results for AMD, making them look as bad as possible.
SO what you've done there, is found another site that is using completely and utterly unfair benchmarking and in no way is, and this is the KEY POINT, using the same method for AMD as they are for Nvidia benchmarking.
You did help me put the final nail in the coffin for Anandtech, I was never truly over enamoured with their benchmarking style, 3-4 line graphs followed by the table of results they used for years missing out lots of important info, then more recently in the last couple years sticking mostly to their long list of barcharts, which are irregularly updated, on top of spouting a lot of crap in a lot of reviews.
[H] are much the same, VERY odd apples to oranges benchmarking, complete inability to draw accurate conclusions, highlighted by their 480gtx review, "its a huge power hog, its crazy hot, its loud as hell" "in furmark 480gtx sli barely uses 100W more power, these cards clearly don't use anywhere near the rumoured 300W", next review 470gtx in sli in metro 2033 uses 250W more than a single 470gtx, STILL haven't noticed and still haven't updated their mixup, which is Furmark + sli/xfire = not even close to useful results.
Benchmarking websites in general, aren't particularly good, but at least get it fair, get it right and update ALL the results, its their JOB for gods sake. There ARE sites that rerun ALL tests with updated drivers in new reviews, techpowerup do, and they have issues, Wizzard used the wrong drivers for the, 480 review and got a lot of stick, but he redid it and updated it and said sorry, plus it made for funny April fools "i'm quitting the site" forum antics.
Your argument of "Anandtech did something even worse so Hardware Canucks are fantastic" is flawed in every single way possible.
Infact your other link that shows they did(or at least pretended) to redo results at a date inbetween, and were too lazy to use their latest set of results in a more recent review actually, to me, shows they are even less professional.
At first it was , they were too lazy to update any Nvidia or AMD results, even though they made a flat out lie that they did so. Your second results suggest these professional and accurate people have a newer set of results, and couldn't be bothered to include them, thats WORSE.
If you want to find a 3rd even worse site that skews the results even worse, go ahead.
If you posted your own results and said "x sites results are crap" I wouldn't have given a monkeys, but don't use another set of inaccurate and not remotely trustworthy results to disprove other inaccurate and not trust worthy results.
I'm not saying the ones you're arguing against are right, and that a 5870 is easily faster than a 480gtx or any other nonsense, all I'm saying is using a dodgey site to disprove another dodgey site, doesn't prove anything at all.
If a site happily says, we benched in 10.6's for everything, but found nothing was faster than 10.4's, except BC2 which is horrible on 10.4 and great on 10.6, but for all around performance we thought it was fairest to use 10.3, fine, thats logical, makes sense and if they do that, using the uber fast 10.6 in just BC2 to give a comparison would also be fair.
But when a site simply claims to use driver X, and its lied, or when it updates all results that go up almost 50% on one set of cards, and not update the results for another set of cards, its just rubbish, utterly rubbish.
Heres a hint, I used to do hardware reviews, its INCREDILY difficult to do fairly, it means the person doing it has to know EVERY single tweak, trick and be able to identify a problem even if the results don't obviously point to something. Its also very difficult to know what game is faster on what driver, when, with which tweak, and with which setting. Its a minefield, but the way to do it is, be fair, be HONEST and tell the person reading what you've done, when you fail to do those three things the review literally is worthless. if HC SAID they used their old AMD results, because they no longer have the card, or didn't have time as they only got the Nvidia card in a day before the review was due, that would let you know why the possible discrepancy, in the HC review its why the 460gtx looks truly fantastic compared to the 480gtx, when in reality its no where near as close as the review suggests.
However again if they TOLD the reader they had old results and even updated the results later on with new results for older/main amd/nvidia cards, again it gives you an idea of what the results mean. Lying, immediately invalidates everything they say. Anandtech updating results that give a 50% performance boost to Nvidia cards only and do not update AMD results is pretty bad, but they EXPLAIN this saying they see no performance improvement over 10.3a, in the link Krugga gives you it shows this, in the review.
Although personally I don't think thats an accurate statement at all. They also missed the fairly obvious problem of, the driver may not have improved, but the game may have improved with a patch since then. Its very hard to keep up to date, but Anandtech should really be using either the latest driver at that point, or at least showing it in ONE benchmark with the same results to prove they aren't any better. Though if they do perform the same theres really very little reason not to update them anyway, its very lazy on their side.
EDIT:- Just for the funny I'll point out that, Nvidia, 49fps on older drivers, 68 fps on newer ones, still only just beats the 5870 it costs significantly more than, Nvidia drivers, just work, especially as getting the performance the card should have, apparently makes the game stutter........ Its funny, as more people buy Nvidia cards, the forum has been full of Nvidia problems the last couple weeks........ really strange how that happens, even more so for a company whose drivers have no problems.
Nvidia's own forums have FAR more people with FAR more problems than AMD's forums for the record.