News Nvidia study links FPS to kill-death ratio

So not only are they admitting they spy on customers but also that enabling DXR isn't something you want to do if you care about your K/D ratio.

 

Or in what will come as no surprise to anyone: "Experienced gamers more likely to use powerful GPUs"

Their debunking of this is pitiful.

Face: palmed.

 

Yeah, I did read that, I just mean their counter argument left me entirely unconvinced.

I feel like being "good" at gaming leaves you more likely to spend more on the hardware (or rather, being crap leaves you less likely to spend more) regardless of how many hours per week you log.

"We know that improving at Battle Royale games is a matter of practice" - so that's where the research ended. "We know this" so we won't back it up with anything empirical.

They formed a conclusion (you should buy a new GPU) and sought out "research" to back it up.

 

When games run better, you play better shocker.

Who'd have thunk it?

 

I'm not being blind am I, did they actually state the FPS of the 600 series card in the BR titles?

 

I didn't see that they stated much in the blog post - they had a stack of numbers and picked ones that supported their conclusion IMO.

 

The idea of FPS having a significant impact on K/D in Fartnite is just laughable, while you could obviously harm your K/D by dropping FPS to an extreme low (say for example trying to run it at max settings and 4K on a GTX 1030) the reality of it is that in Fortnite you'll get a far bigger benefit to K/D by learning to fully exploit the treasure trove of info it dumps on you via audio or learning how to build...
As for their other example of PubeG, that runs so poorly due to incompetent devs that no one knows what happens at high FPS

 

*sniggers*

 

I had one of those last night. I found eating more beans increased my farts per second ratio.

 

wasn't some of ForkKnife's or PUGB in-game physics tied to the frame rate at one point?

 

Thought not! I wonder how well the 600 series do at low settings in those games

Wow, even in MP? If so that's crazy.

 

No idea if they patched it since, but yes that was / is a bug in Fartnite.
However the difference was minimal, plus the way the game discourages auto fire in the extreme it didn't really matter.

 

Even odder in PUBG the exact frame rate is/was linked to altering recoil depending on the individual rate of fire of each weapon: https://www.altchar.com/games-news/586276/lower-fps-can-mean-lower-recoil-in-pubg

 

More frames rendered per second means less time spent rendering each frame. Less time spent rendering a frame means less time between an input and the visible result of that input (i.e. latency).

Assuming all other factors are fixed, going from 60fps to 240fps means going from >16.6ms of frame render time to >4.2ms of frame render time. Or in other words, both reacting to the same event (see screen update, perform input, see result) would give a 24.8ms advantage (double the difference in render time) to the person working at 240fps vs. 60fps. Even with framerates capped, using 'too [powerful' a GPU can still reduce rendertimes even with frames above vsync being dropped. Rendering at 1000fps but dispalying at 60fps (dropping 94% of frames) still means you are getting a 1ms render time per frame, as you are dropping all the earlier rendered frames and only displaying the most recent frame.
On top of that, targets are moving: for example, making a 360° spin in a leisurely one second with a FoV of 90° means an object crosses the screen in 0.25 seconds. At 60fps that gives you 15 images across the display, e.g. for a 27" screen that means an object moves across the screen in 4cm jumps at 60Hz, or 1cm for 240Hz. A 1920x1080 27" panel will have 32 pixels/cm (linear), so at 60fps and object smaller than 128 pixels across will judder across the screen rather than moving perceptually smoothly. Then there's the old rule of thumb for maximum pan rates without judder (this is for film with wide shutter angles, games with little to no motion blur have much lower max pan rates before perceptual judder): 7s per whole screen pan at 24fps, 3.5s/screen at 48fps, etc. that gives you roughly a perceptually smooth pan rate of 2.8s at 60fps and 0.7s at 240fps.

tl;dr more frames means visual tracking tasks are easier and more reliable at higher framerates than lower framerates. Rendertimes mean visual response tasks are faster at higher framerates than lower framerates.

Of course, high framerates and low frametimes are no substitute for the need to Git Gud, but once you have Got Gud then there is benefit to be had (as can be seen with spread widening with experience.