bavor

GloR1ouS_

kevinc313
Sorry man, I'm with Bavor on this one.  Everything you're doing is too inconsistent to conclusively sus out that the driver is causing lower single digit percentage variations.  Sure, if there's large variation due to a major driver defect what you're doing will pick it up, but beyond that no.

A couple google searches would show you that a lot of people have had nightmarish experiences with 3000 series drivers. EVGA's forums used to be top tier with knowledge but it seems things have changed. Check reddit, nvidias forums, linus ect and tell me I am wrong. I have watched many videos on this and I know it makes a difference with a well optimized windows. Keep being sheeple as far as I am concerned.

You are wrong because your testing methods are inconsistent, not repeatable, and don't account for the other variables.  That's the issue we have with your claims.  It doesn't matter that there have been occasional driver issue in drivers.  The simple fact is you made claims without detailed info about your testing methods until after your testing methods were called into question.  Then you later claimed to have certain testing methods that you didn't originally state.  Even then your testing methods can't be considered reliable and accurate.  When testing different drivers you have to account for every external variable.  You aren't doing that because of your testing methods.  Not we can't take your word that you originally used those testing methods because anyone can make something up after the fact.  In addition to that, the video also shows inconsistent and inaccurate testing methods that don't account for all variables.  
 
Your methods are being called into question so that makes your results unreliable at best.  You are taking it personally and resorting to name calling instead of repeating the test with better testing methods.
 
If the variations in game performance youa re claiming existed, thousands of people would notice it and would have reported it.  Searching various forums and Google hasn't yielded any results showing that the driver versions that were listed caused significant issues for a large number of people or anyone really.
 
In my experience, when testing different drivers for both Nvidia and AMD, its rare to have a larger than 2% variance in average FPS and 1% low FPS for both games and synthetic benchmarks between driver versions.  Yes there are a few rare exceptions, but those are very few and far between and never as large as you are claiming.  The biggest difference I've seen between driver versions for port royal can be considered to be within margin of error.  I saw a consistent ~200 point difference between two driver versions in Port Royal after multiple benchmark runs with all the variables controlled as much as possible and 15+ minutes of cool down between benchmark runs.  However that ws 200 points in a system that scores 29,000+ points in ambient conditions.  That's less than a 0.7% variation between drivers after testing multiple drivers.  I've seen variations between drivers and Time Spy and Fire Strike that were between 1% and 2%, but often that can be counted as within margin of error.
 
For CS:GO, Rainbow Six Siege, and WoT at 1080p, small variations in other components of the system usually have a much larger effect than Nvidia driver versions.  With an air cooled GPU, I've seen bigger variations in those games with a 5C difference in GPU temperature(allowing a slightly faster GPU clock for a short time) than with different driver versions.  I've also seen much bigger variations than different driver versions in those games if you allow the CPU to control its own clock speed than locking the CPU to one clock speed for all cores.