Gamers Nexus tests Noctua fan rumors.

[mr moose]

2 hours ago, Bcat00 said:

8 fans is actually a decent sample size for a single reviewer.  No one is going to buy a hundred fans and test them all, its unrealistic and for the love of god, come back it reality and think twice.

It's not though, and not being able to afford more is not an excuse to do half a job.    Whilst I think we all agree that the reality is there probably isn't much of a difference between the two factories, and lets be honest we are not betting on sheep stations here, the actual fact of the matter is this test is at best only highlighting that one production run in china has produced almost identical fans to one production run in Taiwan.   

 

As I already said:

4 hours ago, mr moose said:

Actually, I believe the sample size statistically speaking is 2.  Because 4 fans in a row from one production line should be almost identical and the 4 fans from another should be almost identical also.  To properly test production line quality the samples would have to be random from different production runs from each country. 

 

It's like testing a lake for algae by getting all your samples from one corner at the same time and saying this lake has less algae than that lake.

 

Given the question is about quality from the factory, they need to test samples from different production runs, not 4 samples from the one run.   Especially given one of the Chinese fans was different raises enough concern to warrant testing more from a different batch.

[Doobeedoo]

Well ok that is good to know. 

[WereCat]

@mr moose

While I agree with what you said I must also mention that testing like this may be enough under some circumstances.

Gamers Nexus did their own testing. I have no doubt there will be another journalists who will do their small sample testing.

As a reader you can seek those out and made a conclusion on a much larger sample scale.

(and yes, the testing methodologies will most likely differ)

[Brainless906]

10 hours ago, jcw150 said:

I don't have a problem the sample size here but I agree with your point regarding GamersNexus, he seems to live up his own arse.

Ah something i can (almost) completely agree with. His testing methodology seems to be only as strict as how bad he'd like to shit on a product.

His sample size on the other hand i do take issue, especially when almost 13 percent of his sample size was a full 1db louder and he simple brushed it off.

8 fans from the same runs with one having a significant increase in db isnt going to make me run around screaming like the world is going to end but saying that 13 percent of your damn sample size can be ignored and nothing is different seems like a pretty shit conclusion to me, especially from somebody who loves nothing more than to shit on other channels testing methods.

 

[jcw150]

5 hours ago, tom_w141 said:

No. Longer more in depth videos = good. If everyone becomes LTT with 6 minute product reviews we will never learn anything meaningful. You can't be thorough and quick. I'd rather a 20 minute review to a 5 minute review. I watch LTT for entertainment but not really for serious reviews/knowledge.

I completely agree with everything you say there but that's not the point I was making. Gamers Nexus is not interesting enough to make a 20 minute video enjoyable, the stuff he is often saying is useful and interesting but he so boring to me I won't watch it. I was watching his 22 minute video on threadripper and the information was good but presented in such a boring and not at all concise manner (concise does not mean short, just means to the point).

[Vode]

Meh, as expected. People freak out over nothing. Same as the bugs on Ryzen and Skylake...

[Heesleemer]

12 hours ago, Misanthrope said:

The scope is beyond what I would consider reasonable for an independent tester yes, but that doesn't means that you should call it significant. That's a blatant misrepresentation. He is trying to posse as an authority on testing methodoloy when it suits him but other times he completely ignores his laughable sample size (hint: When it's time to shit on AMD he's all about methodology, when it's time to defend Noctua however, it's "good") 

 

Example: There's a smaller youtuber OztalksHW that I watch and just today he posted a video where he attempted to test i3 vs pentium vs Ryzen 3 in multitasking environments. He specifically mentions how the GamersNexus guy went out of his way to basically tell him that a specific test of having Chrome is inconsistent and would produce meaningless results because there's inherent inconsistencies in the way the load balances on a browser due to caching and other things.

 

Here's the video:

 

  Reveal hidden contents

 

 

Is this fair advice? Perhaps. The variables are though to narrow down but it seems like he was too technical since using a browser while gaming is a fairly common task users have interest in and the variance wouldn't be that dramatic but I said ok he made his point about inconsistency in testing and the importance of proper methodology to control as many factors as possible, that's fair.

 

Then hours later he releases this piece of shit video "testing" just 8 fucking fans and even calling it "good sample" when nothing can be further from the truth. On it's own I could say "Ok he's just a reviewer I wouldn't expect him to get 1000 or 2000 fans I'd be prohibitive and mostly untestable for a small team within a reasonable timeframe" But putting both together, do you see the fucking inconsistencies? 

 

He's ok with calling out a smaller channel publicly about their testing methodology while at the same time testing with 8 fucking fans only. How full of yourself do you have to be to intentionally interfere with your peers while you release pointless fucking testing that's worst yourself.

Id have to agree, it would be nice to have a body of evidence first. I wonder if a meta analysis is possible if multiple small channels conduct same tests. 

[mr moose]

7 hours ago, WereCat said:

@mr moose

While I agree with what you said I must also mention that testing like this may be enough under some circumstances.

Gamers Nexus did their own testing. I have no doubt there will be another journalists who will do their small sample testing.

As a reader you can seek those out and made a conclusion on a much larger sample scale.

(and yes, the testing methodologies will most likely differ)

 

So long as the other tests take random samples, oither wise not only will they have the same problem but it would be very hard comparing fans measured under one set of conditions (equipment and test setup) to another set measured under conditions with an unknown difference.  

 

 

However before we end up in a huge debate, my comments are more because I take a hard line on scientific method and I personally can't handle seeing flawed testing methods being used.  As I said before, there is likely nothing wrong with the fans and people are hardly being duped if they buy theirs from the Chinese factory (unless they die early due to some other weakness).