NVIDIA Starts Disinformation GPP Campaign

Derangel said:
Beyond that there are not numbers to back up the stats. Valve does not reveal its selection methods, it does not reveal how many people they ask each time, it does not reveal user demographics (how many people by region), it does not reveal the percentage of people that responded, there is no information if one set of data is weighted more than another, and so on. There is nothing remotely scentific about the Steam survey which is revealed to the public.
Click to expand...

Corrected.
 
Be nice if AMD made some good GPUs already. Or anyone else, for that matter.

I will very, very begrudgingly buy the next round of NVIDIA cards. I'm calling it quits on Gsync forever now though. Not getting locked into that crud, especially at the premiums they demand.. for some reason.
 
Pieter3dnow said:
I have been thinking about this for a while but to what extent. many virtual machines with just 1 videocard or just a boatload of ghost profiles?
Click to expand...


Nvidia has a much higher amount of cards on the street if you will.

Miners didn't just abandon nvidia cards. Card use is being misreported for both vendors.

The steam statistics should be fine because the sheer amount of correct reporting likely boisters fringe cases, and variation wouldn't account for a full percent in either direction.
 
Shop_vac said:
Nvidia has a much higher amount of cards on the street if you will.

Miners didn't just abandon nvidia cards. Card use is being misreported for both vendors.

The steam statistics should be fine because the sheer amount of correct reporting likely boisters fringe cases, and variation wouldn't account for a full percent in either direction.
Click to expand...

Check what Kyle and Derangel said. That means that unless they show their polling numbers to prove that they have a good cross section of the demographic they sample then you can't make any conclusion on their data.

If you can cheat it would make it worse then it does not matter how many numbers you see unless you have a good idea on how much of the data is from cheaters and then you would have to check where that goes as well.

You know what I wouldn't put it past Nvidia to pay for the data to be skewed this way.
 
Pieter3dnow said:
Check what Kyle and Derangel said. That means that unless they show their polling numbers to prove that they have a good cross section of the demographic they sample then you can't make any conclusion on their data.

If you can cheat it would make it worse then it does not matter how many numbers you see unless you have a good idea on how much of the data is from cheaters and then you would have to check where that goes as well.

You know what I wouldn't put it past Nvidia to pay for the data to be skewed this way.
Click to expand...


I love how hilariously subtle the line between unreliable data and NVIDIA PAYS THEM TO SKEW THE RESULTS is.

https://en.m.wikipedia.org/wiki/Persecutory_delusion
 
IdiotInCharge said:
It's always possible- I think AMD could put out a GPU that is competitive with Nvidia in gaming all the way up and even more today if they weren't also trying to target professional applications with the very same silicon dies.

Granted the fact that they build 'fuller' GPUs has meant that they've been selling every GPU that they can make to miners and that they've been successful with plenty of general compute markets.
Click to expand...
?

Nvidia uses the same die across all applications as well? The only split has been for fp64 with the Gx100 dies each generation.

This is a moot point.
There's nothing in Vega or Polaris that make them better for general compute, as for mining I'm fairly certain AMD's major advantage was the faster execution of certain algorithms making heavy use of circular bit shifts.

The only feature I can think of is packed math on Vega but that's slowly being used in games
 
rgMekanic said:
My argument on it has always been this... how long have you been on Steam? How many times have you been asked to do the hardware survey? I've went 5+ years in between being asked. It wouldn't be hard for them to do a popup to everyone asking them, just like they do with sales twice a week..... why don't they?
Click to expand...

i'm at 13 years and never been asked once.. figured they just pulled the hardware info straight from the steam client every time you login.
 
  • Like
Reactions: Aenra
like this
sirmonkey1985 said:
i'm at 13 years and never been asked once.. figured they just pulled the hardware info straight from the steam client every time you login.
Click to expand...

I've been prompted for the hardware survery extremely regularly, it definitely isn't the most reliable data in the world but I'm yet to see ONE coherent argument that can justify why such an inherently random and uncontrolled trial can result in such a strong skew towards nvidia.
 
The Steam stats for GPUs aren't really that much different than what we've seen in the sales market share data by places like Jon Peddie or whatever, especially if you consider that a decent chunk of AMD sales over the past 2-3 years have gone to mining rather than gaming on Steam.

I guess what people are implying is that a certain percentage of NVIDIA's Steam marketshare is actually GeForce cloud gaming.
 
Pieter3dnow said:
Check what Kyle and Derangel said. That means that unless they show their polling numbers to prove that they have a good cross section of the demographic they sample then you can't make any conclusion on their data.

If you can cheat it would make it worse then it does not matter how many numbers you see unless you have a good idea on how much of the data is from cheaters and then you would have to check where that goes as well.

You know what I wouldn't put it past Nvidia to pay for the data to be skewed this way.
Click to expand...


Steam has enough saturation to provide a working sample. The variance of outliers would be applied to both vendors, since it's logical the same thing could happen to both vendors. This is called a standard curve, and accounts for all the unknowns, so long as enough of a sample group remains.
 
Shop_vac said:
Steam has enough saturation to provide a working sample. The variance of outliers would be applied to both vendors, since it's logical the same thing could happen to both vendors. This is called a standard curve, and accounts for all the unknowns, so long as enough of a sample group remains.
Click to expand...

The problem is there is zero data on the size of their sample group or any demographics of them. Valve does not surface ANY data about their metrics. That makes the Steam numbers unreliable. There is nothing to back up the data. There is a reason that any reputable statistical paper will take pains to list their methods, the response numbers, and so on. Without that the statistics might as well have been pulled out of Gabe Newell's ass because that's all the value they offer.
 
amenx said:
Funny when shady business practices happen with other products/companies no one gives a shit. Only when it involves GPUs, controversies seem to explode with very upset, indignant people. We never saw massive anti-Intel threads when they were doing their crap, which was far worse than Nvidia. No one boycotted Intel, but virtually all enthusiasts bought Intel. Many were aware of Intels practices, but seems they set their principles aside and went for what they viewed was the better product. Wonder how this will shake out if/when Nvidia release Turin soon and it turns out to be a monster performer. And/or if Navi disappoints. Will people discard their lofty principles and do as they did when insidious Intel had the top product?
Click to expand...

Thats a big ass assumption on your part, I discontinued any business with Intel due to that. A lot of people complained about it once they knew about it, but Intel kept it pretty hush for quite some time. This allowed Intel to pocket billions and give people a piss poor experience compared to AMD at the time. You want both companies to compete on a level play field not a stacked one.
 
Derangel said:
The problem is there is zero data on the size of their sample group or any demographics of them. Valve does not surface ANY data about their metrics. That makes the Steam numbers unreliable. There is nothing to back up the data. There is a reason that any reputable statistical paper will take pains to list their methods, the response numbers, and so on. Without that the statistics might as well have been pulled out of Gabe Newell's ass because that's all the value they offer.
Click to expand...
So.. "How do we know Gabe didn't just make up the stats". That it? I'm sure he totally cares about partisan GPU bickering enough to stick his thumb on the scale for either camp. /s

Remember kids: when you don't like the stats, say the stats don't matter.
 
Last edited:
Derangel said:
The problem is there is zero data on the size of their sample group or any demographics of them. Valve does not surface ANY data about their metrics. That makes the Steam numbers unreliable. There is nothing to back up the data. There is a reason that any reputable statistical paper will take pains to list their methods, the response numbers, and so on. Without that the statistics might as well have been pulled out of Gabe Newell's ass because that's all the value they offer.
Click to expand...

Interestingly enough the steam survey seems to match buying trends reported by tech outlets.

I get the feeling people think 3 million under the table amd gpu sales could somehow skew a statistical model that accounted for 125 million users back in 2015, and has grown an estimated 27 million a month in new accounts since then. At the same time believing an nvidia dominant market (66% reported tam) wouldn't have at the very least that many statistical outliers as well.
 
It's funny.

If AMD really is 10% of the market, it makes GPP even worse, not better. They simply do NOT need play dirty.
 
Shop_vac said:
Interestingly enough the steam survey seems to match buying trends reported by tech outlets.

I get the feeling people think 3 million under the table amd gpu sales could somehow skew a statistical model that accounted for 125 million users back in 2015, and has grown an estimated 27 million a month in new accounts since then. At the same time believing an nvidia dominant market (66% reported tam) wouldn't have at the very least that many statistical outliers as well.
Click to expand...

It probably is fairly close, but without the information on how they got to those numbers the charts are pretty worthless because there is nothing to back up their findings. They need to surface the background information on these surveys in order for them to be valid statistical data. For the most part, the data is consistent month to month with very few huge swings. That does speak to their respondent numbers being fairly consistent across the board. The major outlier being the first month where they started surveying Chinese Steam users and even after that things leveled out a bit, which seems like they sent out more surveys to people in China or weighted those numbers higher than other regions for that month to account for the population. But, again, without the methodology information to back it up its all just guessing. The exact same problem exists with the ESA's yearly video game industry reports.
 
Alkaladur said:
I love how hilariously subtle the line between unreliable data and NVIDIA PAYS THEM TO SKEW THE RESULTS is.

https://en.m.wikipedia.org/wiki/Persecutory_delusion
Click to expand...

Is that a link with the algorithm used by STEAM to calculate the gpu usage? Can I suggest you try and read what I wrote. I clearly stated something else then what you even quote me on.
Pieter3dnow said:
You know what I wouldn't put it past Nvidia to pay for the data to be skewed this way.
Click to expand...
That is what I wrote not what you make of it.
 
Pieter3dnow said:
Is that a link with the algorithm used by STEAM to calculate the gpu usage? Can I suggest you try and read what I wrote. I clearly stated something else then what you even quote me on.

That is what I wrote not what you make of it.
Click to expand...

Steam doesn't care about GPU usage, you can check that with GPU-Z or Afterburner if you're so inclined. PM me and I can teach you how.

You wrote that you wouldn't put it past NV to pay Valve to skew the results, and I think that's one of the most profoundly hilarious and unintelligent comments I have read posted on this forum.
 
Pieter3dnow said:
You know what I wouldn't put it past Nvidia to pay for the data to be skewed this way.
Click to expand...

tinfoil-hat-and-tinfoil-cat.jpg
 
viscountalpha said:
The thing I don't understand is why push gpp? It seems greedy and pointless to try to over push your advantage unless amd has something coming that they might be worried about?
Click to expand...
It's easier to fight someone at their weakest than at their strongest, and it makes it harder for the weak guy to gain strength in the future.

These things are cyclical anyway. Nvidia isn't dumb enough to think AMD will never have a competitive product at the top end ever again.
 
deton8 said:
It's easier to fight someone at their weakest than at their strongest, and it makes it harder for the weak guy to gain strength in the future.

These things are cyclical anyway. Nvidia isn't dumb enough to think AMD will never have a competitive product at the top end ever again.
Click to expand...


When you're at the top, nowhere to go but down.
 
I like how Tech of tomorrow comes out with a video titled
Why I know longer accept review samples from AMD. A few days after saying on air he heard Kyle accepted cash to break the GPP story. Sounds like more nvidia damage control to me.
 
cjcox said:
The tinfoil on the man is approved under GPP, but not on the cat. Please remove the tinfoil from the cat if you want full benefits from the program.
Click to expand...

That's not true.

The cat can wear tin foil as long as it's the store brand.

Reynolds Wrap is reserved for the man's head.

Get it straight.
 
Qthulu said:
I like how Tech of tomorrow comes out with a video titled
Why I know longer accept review samples from AMD. A few days after saying on air he heard Kyle accepted cash to break the GPP story. Sounds like more nvidia damage control to me.
Click to expand...

And how many new cards from AMD are announced exactly for this year?
The idea that people that say they know Kyle and then come to the conclusion that he took money from anyone for that matter is rather weird isn't it?
 
Qthulu said:
I like how Tech of tomorrow comes out with a video titled
Why I know longer accept review samples from AMD. A few days after saying on air he heard Kyle accepted cash to break the GPP story. Sounds like more nvidia damage control to me.
Click to expand...
Who knows. Worth mentioning, I did reach out to him to discuss his conversations with NVIDIA and he has not returned by mails.
 
Qthulu said:
I like how Tech of tomorrow comes out with a video titled
Why I know longer accept review samples from AMD. A few days after saying on air he heard Kyle accepted cash to break the GPP story. Sounds like more nvidia damage control to me.
Click to expand...

It's certainly strange to make a video now about something that happened in the past regarding someone that no longer works for the company. If this was an issue, he should have brought it up at the time.
 
Last edited:
what a coincidence. he releases a video suggesting kyle being paid by amd regarding GPP, then a few days later makes another video suggesting shady tactics by amd

it looks like someone did get paid, and that person is not kyle
 
DarkLegacy said:
I'm going to guess that Nvidia will be sending you pretty much nothing to review going forward lol. Nvidia is simply too big to give a shit at the moment. They'll get a slap on the wrist at best but will continue their shady ways since AMD simply can't do anything about it.
Click to expand...


AMD can sue Nvidia before the the federal courts for violating the the Sherman and Clayton antitrust acts for non-competitive predatory practices. But with all their corporate money Nvidia will most likely get off with a slap on the wrist. We live in the most corrupt country on earth and these clowns have the nerve to talk about the Third World. Who do you think those Third World dictators are emulating??
 
You must log in or register to reply here.
Back
Top