What Sucks About HardOCP Video Card Reviews

I wouldn't mind seeing more apples-to-apples comparisons. I know [H] prides themselves on deviating from the standard and showing best playable in game settings, but the problem with this is that what is acceptable is different for everyone. Some people prefer FPS over looks. The apples-to-apples comparisons help tell which games I can run in the 120-140fps range (and at what settings), which is more useful to me than how much can I crank to maintain 60fps. I'm not saying there is anything wrong with the real world benchmark data, but it would be great to see both. I know [H] does that, but it frequently seems like it's only a small sampling of games, and an afterthought.
 
Honestly Kyle, I think you do an excellent job as it is. You and Anandtech are probably the best for giving the best and most-informative reviews and, what's more, you don't pull any punches when companies pull some bullshit.
 
Derangel said:
I like that idea, in theory. There are a couple problems with it. You're equating FPS to the only important thing for determining real world performance. [H] has been trying to say that that is not the case for the last 13 years. On top of that you cannot directly compare FPS when using "best playable" settings as a basis of a review. I suppose you could use apples-to-apples for it but that goes back to the issue of thinking that FPS is the only thing important.
Click to expand...
FPS is not the only thing important, but is nevertheless a VERY important metric when looking at video cards.

Derangel said:
The other big issue is the prices themselves. GPU prices can change weekly, hell sometimes daily. They could put all the work into a chart and then within an hour of the review being published a $20 shift in prices completely invalidates it. This is especially true when we're talking 4-6+ months from the launch of a card.
Click to expand...
I highly doubt a $20 shift in pricing will radically alter their recommendations when you're looking at video cards that cost anywhere from $300-$1,000. People would simply need to be aware that the Dollar/FPS chart contained within the review was valid at the time of its posting.
 
id like see a REAL flight sim like Xplane added this could double as an OpenGL test

VR performance is going to be important now too so that will need to have tests
 
DJS4000 said:
first things first, [H] is the only place on the web that i trust when it comes to video card reviews.

i also know that [H] is a site for enthusiasts. but i'd still like to see comparisons to somewhat older cards, like a 770, to establish a baseline. i mean, 1080p is still the standard resolution for most people, so seeing the cards perform at this resolution would add immense value, i'd wager.



cease the peasantry. eyes & brain are capable of perceiving frame differences in the range of 200+ fps. anyone who games regularly will be able to tell 80fps and 144fps apart.
Click to expand...



I can definitely see 10 fps increments up to 90 fps, and from there I can see about 20 fps increments, up to 180hz afaik. That's far as I've overdriven a panel.
 
To answer the original question? Nothing really sucks about [H] video card reviews. I'm a long-time reader since the Radeon X800 days and I really miss those days when you all reviewed even budget cards. But I understand that time is money, and not everyone is into that. I also like that most of your video card reviews, by and large, have not changed. I really loved reading your articles back then, and I still read them today.

Actually... Dare I say? I still go back and read your old reviews here and there for nostalgia. :D
 
Also can you add a GTX 780 to one of the newer reviews, I'd like to see how it performs now.
 
  • Like
Reactions: Elios
like this
Silverghost said:
Opening pictures in a new tab sucks. They also kind of low in quality.

Frame times as well would be great.
Click to expand...
i think a lot of us use chrome...including kyle....pictures always are fine with chrome for some reason
 
I cant say the existing format of reviews sucks. SOmetimes less than optimal would be better worded though less attention grabbing.

My biggest gripe is that the reviewers all like different games then I. Not just say shooter X vs Shooter Y its a whole genre off. Truth be told that is not unique to the [H].

Like many I will often go right to the conclusions, then a day or two later as I have time I will go back and read the whole thing.

Shooters and RTS type games push the hardware one way, RPG/MMO games another. I wish I could see more numbers from games I play or games like those I play. That said I get it is a whole lot easier to test games with an offline, single player mode and get consistent meaningful data.

I think the image quality reviews would be the best place to mix in some of the other type games. (I have been very happy to see such IQ articles in the past)

I bet many people would love to see an IQ and FPS review of Black Desert Online right now. It is fairly new to the American/EU game space and has a big rep for being truly awesome to look at.

The other big gap I see with the reviews is the spread of hardware. Where to draw the line is probably a big behind the scene topic. Just review the latest two models from AMD/Nvidia or go back further?

I suspect I am not the only one who needs to skip every generation or two (three) when upgrading my GPU. I know not every single card can be included it would be nice to see maybe one card from each of the past 2-3 generations from each vendor. For example with the GTX 900 V AMD (is it 390 currently?) also do the 780 and 680 along with the AMD equivalent that would help us with older cards fine tune when to pull the trigger on an upgrade. I dont know it would even need to be every new card on the market to include the old ones, maybe 2-3 times a year?

Anyhow... I have been happy with the work of Kyle and others and since you are asking for feedback I am happy to give some.
 
I am just going to say this

First off I like this site and if that was not the case I would not have been here for nearly 12 years Over all I think the data prevented is very good.

But you all got issues with AMD that I personally think should not be. Radeon Fury for example. you all knew what it's limitations were but you all thrashed it anyways. You guys would have never tore NVidia a new bung hole like that if they had released the same thing... They were literally between a rock and a hard place... you put that RAM on a slower card and voilà you can beat your top end card in some situations.

AND has issues, everyone gets that but when I got my r280 for $160 (including $20 MIR)a year ago, that was an amazing deal. I picked one up and I have had ZERO issues with drivers. Everything works all the games I run work just fine and I have NEVER had a card catch on FIRE due to a driver disabling a fan.

NVidia gets away with some SERIOUS crap and they are rarely called out to the extent AMD is
 
Last edited:
It's funny you should say that because in a number of threads I complained about [H]'s repeated failure to include VRAM usage figures, in which Nvidia would shine because the Titan X has 12 GB whereas at the time AMD GPUs had a maximum of 4 GB, though that was not the point of my complaint. A Nvidia fanboy would point to that as favouring AMD. [H] have listened to me and now include VRAM usage figures, for which I am duly grateful.
 
YeuEmMaiMai said:
I am just going to say this

First off I like this site and if that was not the case I would not have been here for nearly 12 years Over all I think the data prevented is very good.

But you all got issues with AMD that I personally think should not be. Radeon Fury for example. you all knew what it's limitations were but you all thrashed it anyways. You guys would have never tore Nvidia a new bung hole like that if they had released the same thing... They were literally between a rock and a hard place... you put that RAM on a slower card and voilà you can beat your top end card in some situations.

AND has issues, everyone gets that but when I got my r280 for $160 (including $20 MIR)a year ago, that was an amazing deal. I picked one up and I have had ZERO issues with drivers. Everything works all the games I run work just fine and I have NEVER had a card catch on FIRE due to a driver disabling a fan.

NVidia gets away with some SERIOUS crap and they are rarely called out to the extent AMD is
Click to expand...

Can't say I agree.

Nvidia didn't release a un-competitive product at an absurd price. However you feel about the limited options AMD had, the one move they could/should have pulled was price. The Nano is laughably over-priced, and I think actually [H] went out of their way to review it in the fairest light possible, frankly showing bias towards AMD. But we all knew the truth about Nano. I mean, No-one here bought it, right? Anyone? Seriously? Who didn't just buy it for the curiosity or sake of?

I've never felt any bias in [H]'s reviews. That they buy into the nvidia produced FCAT testing is evidence of this.
 
If I could.......

It's not that HardOCP sucks at all in reviews. In fact I think this site has the best reviews there are.

I would just offer, why not have a poll to gather information on what the readership actually runs at home. Sort of like the Steam data.
Find out what resolutions are used the most, what setups people use........why go to all the trouble to test 6 GPUs at 4K when 12 people on your website actually use 4K.

Is Surround more popular? Is SLI more widely used? For example.....

Who does what might be really helpful, just my 2 cents worth.
 
DJS4000 said:
cease the peasantry. eyes & brain are capable of perceiving frame differences in the range of 200+ fps. anyone who games regularly will be able to tell 80fps and 144fps apart.
Click to expand...
The brain is capable of perceiving these differences, the eyes are not. When you're gaming the difference between 80fps and 144fps are felt through the central nervous system, not actually seen. It is important to have those extra frames in high level gaming. But I still remain firm in my claim that 80fps should be the standard based on the limits of the eyes.
 
Vendor neutral game suite. Or to make a point to have an even amount of GW and GE titles in the suite would put any of that bias doubt to rest. The game suite [H] had the past year was mostly Gameworks and initially put AMD in the worst light possible.
 
magoo said:
I would just offer, why not have a poll to gather information on what the readership actually runs at home. Sort of like the Steam data.
Find out what resolutions are used the most, what setups people use........why go to all the trouble to test 6 GPUs at 4K when 12 people on your website actually use 4K.
Click to expand...

I run some games at 4K. Not many, as many don't play well. I really would run more if I could, though.

My next upgrade will be able to play current games at 4K very easily. That's going to be one of the selling points. I'm glad to see those benchmarks. 1080P? My kids use that resolution, and I think of it as minimum for this site. If it can play well at 1080P, it'll do great at anything below.
 
spine said:
Can't say I agree.

Nvidia didn't release a un-competitive product at an absurd price. However you feel about the limited options AMD had, the one move they could/should have pulled was price. The Nano is laughably over-priced, and I think actually [H] went out of their way to review it in the fairest light possible, frankly showing bias towards AMD. But we all knew the truth about Nano. I mean, No-one here bought it, right? Anyone? Seriously? Who didn't just buy it for the curiosity or sake of?

I've never felt any bias in [H]'s reviews. That they buy into the nvidia produced FCAT testing is evidence of this.
Click to expand...

new tech comes at a price.....and I never said it wasn't over priced.... and how they handled the whole AMD issue was unprofessional.
 
I have had a few days to think since my first post.

It is clear FPS is still very important to many people as an effective metric.

How can the reviews maintain what clearly works (at least partially) for just about everyone who posted here, and we all know persons speak up much louder and energetically making sure everyone (and the horse they rode in on) knows what they dont like, and then add some, much less all, of what we all would like to see?

When I get a new game the first thing I do, often before I am in the world, is max every setting my current system will allow me to turn on, set to high/max/uber and then jump in the world and see if I can even move.

I imaging I am not alone in this.

I then start scaling back settings till I can move, spin in place, and otherwise generally interact with the game world.

I dont even look at the FPS till I am trying to fine tune AA or other very small adjustments... where I cant immediately visually and by moving around tell the difference.

I know from experience in many open world single player/mmo worlds that 60 fps is generally not needed to have a smooth in game existance. I have found depending on the game that 15-20 is generally the minimum workable area and 30-40+ is optimal.

Once I have played a few days I will look at the optimization type guide, ini settings or what have you.

Skyrim is really two games for me... the play the quests, build my characters, exist in the world game and then get the .ini files just right game.

Now that I have given some insight to my specifics, I am wondering how can a review of hardware help the people who are FPS oriented and give them the information they need while at the same time give persons who are all about the image quality and smoothness of play what they want to know. Both groups of course are thinking (I suspect) is it time (or will it benefit me enough) to upgrade now with this new toy?

I'd like to see some reviews for AMD systems as well, that is after all what us budget conscious persons usually have.

I suspect Kyle and others dont want to spend 2 months 12 hours a day building and changing hardware, recording numbers, ignoring friends and family to make a comprehensive, makes everyone happy, 100 page review that everyone just jumps to the conclusion at the end reading nothing else.

-edit : added a few commas
 
Last edited:
I like the reviews, while I find the 'best playable settings' idea to be a little counterintuitive when trying to gauge performance of one product vs another, I appreciate that you also append an apples to apples comparison on there.

My biggest gripe with your reviews is the comparisons.

In one 390 review you have pitted it against a 970, both Max OC and stock. In a 980Ti review I checked, it was a 4-way battle between 980Tis

Recently I've been looking at your reviews specifically to try and get some OC vs OC data on 390/390x vs 970/980, and while I did get it, it was a little... counterintuitive to find

otherwise, great stuff
 
One thing I will point out is that [H] likes AA. On my high dpi 28" 4K monitor, I find that I don't seem to need it as the higher dpi works to the same end. So it would be very useful to find out what I could enable instead. [H] have done some very useful in depth graphics tests on a few games and these can substitute but only for those games and only should I play them. But I would appreciate an expert eye.
 
Rvenger said:
Vendor neutral game suite. Or to make a point to have an even amount of GW and GE titles in the suite would put any of that bias doubt to rest. The game suite [H] had the past year was mostly Gameworks and initially put AMD in the worst light possible.
Click to expand...

Come on man. Gameworks can be turned off. Anyway, [H] uses games a lot of people play and it wouldn't make sense to use obscure games just because they weren't sponsored by NV or AMD. Fuck the ADF and the NVidiots. I want to see how my card plays games in the reviews and I could care less about who sponsored those games. I play games not stupid PC hardware company politics.
 
My criticism of [H] GPU reviews is the same as other sites' GPU reviews: you generally compare cards within the same generation. While this is understandable from an apples-to-apples standpoint, it doesn't help me figure out the relative performance improvement I should expect to gain when I upgrade from my three gen old GPU. There is no reference point; I have to go back to your older reviews of the GPU I own, look at the benchmarks there, then attempt to extrapolate the old GPU's performance when applied to the new benchmarks of the new GPU.

It's fine and dandy to tell me that New GPU A performs 10% better than New GPU B at a lower cost, but that doesn't tell me if the total final cost of New GPU A is worth the performance improvement over Currently Owned GPU A.
 
Ieldra said:
I like the reviews, while I find the 'best playable settings' idea to be a little counterintuitive when trying to gauge performance of one product vs another, I appreciate that you also append an apples to apples comparison on there.

My biggest gripe with your reviews is the comparisons.

In one 390 review you have pitted it against a 970, both Max OC and stock. In a 980Ti review I checked, it was a 4-way battle between 980Tis

Recently I've been looking at your reviews specifically to try and get some OC vs OC data on 390/390x vs 970/980, and while I did get it, it was a little... counterintuitive to find

otherwise, great stuff
Click to expand...

I agree. The best playable settings comparisons don't add value for me. Apples and oranges when you do that. I would prefer a section where you use that information and SHOW images where the different settings result in visual differences. HBAO+ is off to get a 60 fps min? SHOW me HBAO+ on and off. Otherwise the fps graphs aren't much use unless all in- game settings are identical. IMHO.
 
c3k said:
I agree. The best playable settings comparisons don't add value for me. Apples and oranges when you do that. I would prefer a section where you use that information and SHOW images where the different settings result in visual differences. HBAO+ is off to get a 60 fps min? SHOW me HBAO+ on and off. Otherwise the fps graphs aren't much use unless all in- game settings are identical. IMHO.
Click to expand...
HardOCP normally does an outstanding overview of a new title in multiple parts with performance/settings/image quality which then the title may or may not be part of the HardOCP future hardware reviews. These are some if not the best reviews I've seen on the internet dealing with a new game and game settings. I don't think this needs to be duplicated for each review unless something dramatically changes which I've seen happen and was put in later. Keeping it more focus on the hardware vice going into each game setting in the review, keeping it as streamlined as possible is probably much better in the long run. Anyways what you ask is done.
 
DooKey said:
Come on man. Gameworks can be turned off. Anyway, [H] uses games a lot of people play and it wouldn't make sense to use obscure games just because they weren't sponsored by NV or AMD. Fuck the ADF and the NVidiots. I want to see how my card plays games in the reviews and I could care less about who sponsored those games. I play games not stupid PC hardware company politics.
Click to expand...


Only reason why I brought it up was because it would just eliminate any questioning from others. I am tired of the fanboy wars myself as you should know that from me of all people. Level the playing field then the fanboys can't cry about it. Gameworks can be turned off... Async can be disabled by running DX11 etc, who cares because the reviewers will run the features at max settings/enabled which you know can't be done for both camps on all games.
 
Well isn't that the idea of a review to know what each card is capable of or not? If games have features that only Nvidia or AMD can use why not test it and show the results? Makes it easier for one to decide if one wants that feature or not. My only real complaint, which I hate repeating over and over again is the number of sampled games or data points which would cover a broader selection (which I do believe would be very hard for HardOCP to do every time due to the current method - hence cycle through games on the reviews giving a slightly larger selection).

As for GameWorks that can be covered on a whole different topic but for games that use it - go for it. I would like to know as well if a game is hindered or not on my hardware before I buy it.
 
Rvenger said:
Only reason why I brought it up was because it would just eliminate any questioning from others. I am tired of the fanboy wars myself as you should know that from me of all people. Level the playing field then the fanboys can't cry about it. Gameworks can be turned off... Async can be disabled by running DX11 etc, who cares because the reviewers will run the features at max settings/enabled which you know can't be done for both camps on all games.
Click to expand...

Yeah, it's too bad people can't just except reality and forget what they "think" things should be like. It's amazing how so many think hardware companies owe something to them.
 
Rvenger said:
Only reason why I brought it up was because it would just eliminate any questioning from others. I am tired of the fanboy wars myself as you should know that from me of all people. Level the playing field then the fanboys can't cry about it. Gameworks can be turned off... Async can be disabled by running DX11 etc, who cares because the reviewers will run the features at max settings/enabled which you know can't be done for both camps on all games.
Click to expand...

Difference is Async Compute is a part of the DX12 spec. Gameworks is Nvidia built code.

Async Compute will be used here on out. Gameworks will hopefully die and go away.

MSFT @ GDC:
 
I never expected this thread..
I'm glad to see a willingness to hear out input as opposed to ignoring it.
 
Terpfen said:
My criticism of [H] GPU reviews is the same as other sites' GPU reviews: you generally compare cards within the same generation. While this is understandable from an apples-to-apples standpoint, it doesn't help me figure out the relative performance improvement I should expect to gain when I upgrade from my three gen old GPU. There is no reference point; I have to go back to your older reviews of the GPU I own, look at the benchmarks there, then attempt to extrapolate the old GPU's performance when applied to the new benchmarks of the new GPU.

It's fine and dandy to tell me that New GPU A performs 10% better than New GPU B at a lower cost, but that doesn't tell me if the total final cost of New GPU A is worth the performance improvement over Currently Owned GPU A.
Click to expand...

I totally agree, and its not just a [H] thing... this is why I find AT's "bench" tool to be very appealing and valuable, even if the comparisons need a big grain of salt to go with. A lot of times the comparison is made right when a new generation comes out at then we never see it again despite new games/drivers etc coming along. Throwing a previous gen flagship in, or equivalent card from the previous generation or two would be nice... even if it was a separate article (not in every comparison). Even just one card that can be used as a kind of control to see how far things have come would be helpful to add context to previous reviews at times (i.e. A is faster than B by 10%, B was 10% faster than C 2 years ago, so now I know C is going to be ~20% slower than A)
 
I like the reviews as is. I'm an enthusiast, the sheer amount of data in the reviews is what I love to read, the graphs, the settings, the picture quality, memory issues, all of it. I don't skip any of it.
 
Brackle said:
Gameworks will hopefully die and go away.
Click to expand...

Unfortunately that will never happen! Nvidia is too large and has too much of a market share to let GW die off.

For the same reason PhysX is still around, VisualFX and OptiX will remain in use.

The only way to kill off GW is for GPUOpen to be adopted by all AAA devs. The problem is AMD is way behind the power curve on GPUOpen when compared to GW and NV keeps getting AAA games to use one part(s) of GW or all of GW.
 
Cr4ckm0nk3y said:
Unfortunately that will never happen! Nvidia is too large and has too much of a market share to let GW die off.

For the same reason PhysX is still around, VisualFX and OptiX will remain in use.

The only way to kill off GW is for GPUOpen to be adopted by all AAA devs. The problem is AMD is way behind the power curve on GPUOpen when compared to GW and NV keeps getting AAA games to use one part(s) of GW or all of GW.
Click to expand...
Not entirely accurate. Tomb raider had GW but opted Tressfx(Purehair) because it was open and able and free to be modified.
 
Except GW coded titles have had some very problematic releases with performance issues as well as being extremely buggy. What ever the developer uses it needs to work fairly well, black box libraries does not look like it is cutting it. Signing license agreement handicapping what you can do and what kind of ownership you have is also another issue. In other words GW may not be paying off in the long run and needs to be revised. Wait that is what happening, it is now more open from Nvidia.

What ever system makes the most money or gives the best results will be used, maybe a combination. Also GW does not look like it helps the developer to develop console games as much. So we will see where GW goes, I expect Nvidia will be more open with their code and maybe even optimize or make sure it works with other hardware (AMD, Intel) while using their own hardware effectively.
 
Love the reviews the way they are, no bs!
Only thing I'd like to see would be run threw videos for each game in the test suite. That way people can see what sections you're using and the the course you take. You could just record one run from each game, doesn't matter from which card or even what settings it's at. It'd be just for demonstration purposes. Then have a link in the game info section. You'd only have to do this once for each game and reuse the video with each gpu review. At least until you replace a game in the test suite. Then make a new one during the next gpu review you do. Hope that makes sense.
 
You must log in or register to reply here.
Back
Top