Because if you look closer at the data in the Geekbench browser, it’s kind of shit. The iPad entries are probably not too far off, but there are a ton of entries that are obvious garbage, like a Pixel 7 Pro with a Ryzen 9 5900X. Also, a lot of system names are VM hypervisors. In a VM, you can control the realtime clock that the Geekbench profiling software sees, so you can just kind of dial whatever performance number you want.
Geekbench obviously just takes the average, but the average of garbage is still garbage.
What I’m hearing from you is that because some of the numbers are faked then the entire data set is unreliable and you don’t believe this information that is coming in.
That’s fine.
I guess I thought given your assertion you had … more.
The data is unreliable. If we knew how much of the data was faked we could compensate for it, but we don’t. We could discard the outliers, but we don’t know if we’re discarding valid data, and someone who is deliberately tainting the dataset would submit a bunch of samples that are only a little bit off as well.
And while some of the numbers must be from trolls, manufacturers (and shady investors) are heavily incentvized to sway the listings.
Why?
Because if you look closer at the data in the Geekbench browser, it’s kind of shit. The iPad entries are probably not too far off, but there are a ton of entries that are obvious garbage, like a Pixel 7 Pro with a Ryzen 9 5900X. Also, a lot of system names are VM hypervisors. In a VM, you can control the realtime clock that the Geekbench profiling software sees, so you can just kind of dial whatever performance number you want.
Geekbench obviously just takes the average, but the average of garbage is still garbage.
So your assertion is these numbers are faked rather than ran by an actual device then?
Some of the number are faked. The only person who knows the accuracy of these one are the people who posted them.
What I’m hearing from you is that because some of the numbers are faked then the entire data set is unreliable and you don’t believe this information that is coming in.
That’s fine.
I guess I thought given your assertion you had … more.
Basically it’s your feeling.
The data is unreliable. If we knew how much of the data was faked we could compensate for it, but we don’t. We could discard the outliers, but we don’t know if we’re discarding valid data, and someone who is deliberately tainting the dataset would submit a bunch of samples that are only a little bit off as well.
And while some of the numbers must be from trolls, manufacturers (and shady investors) are heavily incentvized to sway the listings.