Biased Reviews ?

[Johny Boy]

For a while i was trying to figure out how few reviews sites publish contradicting info's with miss match data among them.Is it possible for them to have biased one sided reviewing mindset when comparing products from sponsors ?

I have seen how processor reviews fluctuates from reviewing product being branded " Decent to Rubbish " while reading reviews from one website to another one.

Is it possible that synthetic vs actual performance have large variations ?

For example - FX piledriver processor have very bad review results from Bit-Tech but decent from Legit-review's.

Just wanted to know how full proof this reviewing sites publish their data that becomes buying benchmark for buyers ?

[Cacao]

Reviews are very personal, everyone does their own testing and everyone has their own opinions. There wouldn't be much point in reviews if they all said the same thing. Find someone you trust and see what they have to say about the product.

[Johny Boy]

I have been told reviews with Cinebench are biased towards one particular platform due to compiler.

[Emperor_Piehead]

I have been told reviews with Cinebench are biased towards one particular platform due to compiler.

Well you can have synthetic benchmarks which have been programmed to be more favorable with intel over amd sometimes.

[Johny Boy]

Well you can have synthetic benchmarks which have been programmed to be more favorable with intel over amd sometimes.

Yes but max temperature / power consumption too have mismatch.I know every chip is different from other but still one review quoting 100+W over its counter product but second site's info has mere 40W difference.

Same goes with GPU comparisons where brand quality takes first stage over performance.

[Emperor_Piehead]

Yes but max temperature / power consumption too have mismatch.I know every chip is different from other but still one review quoting 100+W over its counter product but second site's info has mere 40W difference.

Same goes with GPU comparisons where brand quality takes first stage over performance.

Well for each card you can't really change much as a board partner so things like the warranty and bundled software may cause a boost in score then say a  card of a different vendor that is O/Ced higher, but only about 2-5% faster.

[Johny Boy]

Well for each card you can't really change much as a board partner so things like the warranty and bundled software may cause a boost in score then say a  card of a different vendor that is O/Ced higher, but only about 2-5% faster.

That's why i said brand quality over performance.Having better quality part's with finely tuned software can provide extra premium on reviewing but just warranty or software should not hold validation over primary things like performance.

But why does every software that reviewing site uses are less supportive on AMD's cpu's ? If reviewer knows that certain part of software wont optimize the full power of a product then why to use it .......is it just to show how badly one is over other one ?

[Emperor_Piehead]

That's why i said brand quality over performance.Having better quality part's with finely tuned software can provide extra premium on reviewing but just warranty or software should not hold validation over primary things like performance.

But why does every software that reviewing site uses are less supportive on AMD's cpu's ? If reviewer knows that certain part of software wont optimize the full power of a product then why to use it .......is it just to show how badly one is over other one ?

because there isn't many programs that are synthetic that haven't been coded for better intel performance. Synthetics don't mean much anyways when comparing cpus. Most of the reviews I read though are pretty unbiased on most products.

[TechFan@ic]

It's usually ignorance rather than bias, I really detest reviews that bash you in the head with synthetics even though most of them acknowledge how little synthetic benchmarks reflect on real world performance & then expect you to make an educated buying decision based on relative figures of performance that don't exist in reality.

I prefer reviewers that make their methodology for testing public & test their products with applications people actually use in the real world, like Linus/Slick , Tek Syndicate & GameGPU.