Anyone think more Objectivity is needed?

[tuteja1986]

Originally posted by: n7

Originally posted by: keysplayr2003
There isn't anything wrong with HardOCP's benchmarking methods. It's only different. Details are given on each test as to what was turned down or not. To this day, nobody has proven their benches to be false or misleading.

Of course not.
Kinda hard to when you review things in a completely messed up unreproduceable way.
So here we have apples, oranges, pears, some plums, & some bananas...

I used to actually read their reviews for information.

But then they decided to do everything "real world".
Problem with their idea of real world is that it's not reliable.

Like how when Intel's Core 2 Duo came out, what'd we see there?

A review of the two systems, an A64 X2 & a C2D @ 1600x1200, maxed graphics settings.
Their conclusion as i recall it was that it doesn't matter which high end CPU you have, as GPUs are the main limiting factor.
Well no sh!t sherlock :roll:

Anyway, HardOCP really isn't worth arguing about, so i'll shut up.
But one thing i don't want to have to keep seeing is more Wreckage + pro-nV reviews every few days.
It's getting very old, & everyone here knows his agenda.

Firingsquad editor , Bit-tech editor , beyond 3d and 100 Thousand people think Hardocp is nvidia biased. Hardocp once said the everyone was wrong and that started a war as firingsquad wrote up an article showing how hardocp was wrong. then few hrs later some hacked firingsquad website and changed the firingsquad name in front page to firngsquid. Also hardocp then wrote an article showing that firngsquad article on them was stupid.

Word of Advice :! I mean if you go to any press-confrence of ATI / Nvidia. You say "There isn't anything wrong with HardOCP's benchmarking methods." to a senior editors at major tech website you will all respect from them.

 

[superbooga]

How is HardOCP biased? They said the x1950xtx was superior to the 7950gx2 when most sites said otherwise. Perhaps their benchmarks are too limited, and it is silly of them to say "product X delivers a much better experience than product Y in game Z" when the difference is only like one minor quality settings (they do this in every review).

Anyways, perhaps apples-to-apples comparisons with a frame rate logs are the best. The minimum/average framerates don't 't tell you enough. Sure, if your minimum is 60, then you know it will be playable. But in that case, couldn't you raise the resolution or quality settings and still get a playable framerate? Now if the minimum drops to 30 fps, is it still playable? Two cards can both get a minimum of 30 fps, but one can be playable and the other unplayable.

 

[z3R0C00L]

Every forum I've posted this in... the overwhelming response has been.. that HardOCP/Fudzilla/Bjorn3D are major problems.

Overwhelming.. I hope Hard/Fud and Bjorn3D members have seen exactly the type of responses I've gotten. Cuz it don't look pretty towards them.

I mean for pete's sake.. XFX's official support forums are at Bjorn3D...lol And XFX is plastered all over the site.

 

[apoppin]

it reminds me of the Cold War ... when the USSR entered their Lada against GM's Cadillac in a State-sponsored "comparison"

well, of course, The USA car won every single test ... but the Soviet Press proudly printed:
"Big Car Competition, USSR comes in 2nd, overall ... USA 2nd to Last"

a lot depends on the presentation

 

[betasub]

Originally posted by: apoppin
it reminds me of the Cold War ... when the USSR entered their Lada against GM's Cadillac in a State-sponsored "comparison"

well, of course, The USA car won every single test ... but the Soviet Press proudly printed:
"Big Car Competition, USSR comes in 2nd, overall ... USA 2nd to Last"

Point taken, but I think the facts are a bit out: the Lada was/is a budget brand - the Cadillac "big car" imitator was the Zil.

 

[Keysplayr]

CP, which games do you own that are in H's article. And which games do you think play as slide shows at H's given maximum playable settings? I'd like to see exactly what you mean for myself what you are seeing. I am asking you because of your blanket statement and didn't specify. I have a feeling that they would not say a game was playable at certain settings, when it was indeed not playable.

 

[Cookie Monster]

Originally posted by: Extelleron
I agree with the OP that HardOCP is not a reputable site and definately favors nVidia, but I don't agree about only testing those "popular" games.

I don't know about you, but I haven't touched Battlefield 2 in a while, it's been a month or so since I've played Doom 3, I haven't touched F.E.A.R. in quite some time either. I have, however, played STALKER in the last few days.

Sites should test with a wide variety of games based on different engines, but they should definately test newer games. I'm probably not going to go back and play a 2 year old game on my new video card, but I will play a new game like STALKER, so I care about its performance.

The reason HardOCP is not reputable is because their way of benchmarking allows for benchmarks to be "fixed" based on which card they like better. Each card has strengths and weaknesses. HardOCP knows the HD 2900XT sucks at AA, so they use AA in their tests (which I actually agree with.) Let's say the HD 2900XT sees a big drop from "High Shadows" and the 8800GTS sees a minor drop. However, the 8800GTS sees a huge drop from "Ultra Lighting" but the HD 2900XT sees a small drop. HardOCP will use "High Shadows", but not "Ultra Lighting". Their test method is flawed. Reputable sites should use MAXIMUM IN GAME SETTINGS and not fool around with individual settings in-game. It's just too hard for bias to take over when you can fool around with individual settings.

Proof? I agree with some of your points but to make outlandishly statements like above without proof is quite pointless.

BF2 is alot popular than you think. (since i do a part time job at a pc cafe)

However i dont see anything wrong with their method of benchmarking but infact i question their conclusions based on their benchmarks. They make it sound 10 times worse than what the numbers actually portray in some occasions.

Well, hardocp does have its apple to apple comparisons if you guys missed that.

 

[apoppin]

Originally posted by: Cookie Monster

Originally posted by: Extelleron
I agree with the OP that HardOCP is not a reputable site and definately favors nVidia, but I don't agree about only testing those "popular" games.

I don't know about you, but I haven't touched Battlefield 2 in a while, it's been a month or so since I've played Doom 3, I haven't touched F.E.A.R. in quite some time either. I have, however, played STALKER in the last few days.

Sites should test with a wide variety of games based on different engines, but they should definately test newer games. I'm probably not going to go back and play a 2 year old game on my new video card, but I will play a new game like STALKER, so I care about its performance.

The reason HardOCP is not reputable is because their way of benchmarking allows for benchmarks to be "fixed" based on which card they like better. Each card has strengths and weaknesses. HardOCP knows the HD 2900XT sucks at AA, so they use AA in their tests (which I actually agree with.) Let's say the HD 2900XT sees a big drop from "High Shadows" and the 8800GTS sees a minor drop. However, the 8800GTS sees a huge drop from "Ultra Lighting" but the HD 2900XT sees a small drop. HardOCP will use "High Shadows", but not "Ultra Lighting". Their test method is flawed. Reputable sites should use MAXIMUM IN GAME SETTINGS and not fool around with individual settings in-game. It's just too hard for bias to take over when you can fool around with individual settings.

Proof? I agree with some of your points but to make outlandishly statements like above without proof is quite pointless.

BF2 is alot popular than you think. (since i do a part time job at a pc cafe)

However i dont see anything wrong with their method of benchmarking but infact i question their conclusions based on their benchmarks. They make it sound 10 times worse than what the numbers actually portray in some occasions.

Well, hardocp does have its apple to apple comparisons if you guys missed that.

there IS nothing wrong with their comparisons ... it is a great idea ... BUT, being allowed to "fiddle" with the individual setting without posting more one single setting result is not only "suspicious" but rather "in line" with their out-of-sync conclusions.

Derek Wilson has posted in our long 'in house' thread that AT disagrees with HardOCP's testing methods/conclusions and mentioned has a plan to do something similar, --the "AT way' [ie. "the right way", imo] - but he won't attack them as he has way too much class

no one trusts HardOCP anymore except hardcore nvidia fans
--doesn't that tell you something about their methods and their slant?

 

[CP5670]

I think the best benchmarking method would consist of comparisons at equal settings, but with the type of graphs that HardOCP uses. As much I don't like their comparisons of different settings and arbitrary notions of "playability," the graphs are an excellent idea and provide far more information than a single framerate number.

 

[Sable]

I've always liked techreports reviews. Fair and balanced in my opinion.

 

[CP5670]

I like Techreport too. The only problem is that they sometimes don't bench enough games.