FPS Boosting.

Great post scenic, will post my reply in detail when I get back to desktop.
Until then heres more testing:
Please note there was only one run at each of the settings and the line I took, while decently similar, vary a bit which on the mischiveous random-boost Spa straight can result in a lot of time gain. Either way, here it is:

i wasnt trying to get as many boosts as possible, just to hold a somewhat similar line

Interesting! And I acknowlegde the effort put into doing and presenting that! So it may be a difference between the fps limiting methods, from your results one would think to avoid in-game vsync (as is the common wisdom), it's faster with unlocked or externally limited. Unless vsync can be set to 144, in which case it's faster than all others. On your pc. 

Since I've got this in Excel already, if your fastest 60hz / unlocked time (1:00:02 / 1:00:27) is comparable to Story's 60fps speed, Story's unlocked setup would do your testrun in 00:56:71 (extrapolating his 5,83% advantage measured over 17 seconds - so assuming he is not still accelerating at that point). 

Looking forward to hearing your full reply/analysis


ps! maybe your result-frame would make a better video-thumbnail. 

bookmark this, it's all you need to know.

Sarah I dont know now but it has always been my belief that the games physics engine is tied to the frame rate and not a set Hz like what more race focus games are. like project cars with a constant, dedicated 200Hz physics engine. So that is why we are seeing the difference as you mentioned and why cars drive and behave differently throughout certain ranges of fps.

The communication between the machine and the screen is only a one-way stream. This means that the machine just hands off a ton of data to the monitor, but never gets anything in return to sync back to. I am not sure if this is different for freesync/Gsync monitors, but for traditional screens what you see on the screen and what's happening inside the machine are two seperate systems.

Layman: the computer spams the screen with images, the monitor just prays it will survive.

So what's Vsync then?
It's the machine's attempt at mimicking the image timing of the monitor, but it's never perfect unless the monitor does some extra work to sync up images as seen with freesync/Gsync. Even when the framerate is exactly equal to the refresh rate you'll most likely still see the images slightly out of sync by a few milliseconds, leaving you with the possibility of a single tear somewhere on the screen. It's for this reason that additional monitor tech was introduced to further improve the image quality by attempting to correct frame pacing if possible.

Layman: Vsync attempts to mimick the monitor, but it doesnt actually communicate back. Freesync/Gsync was introduced to help correcting the frame pacing to further minimize screen tearing if possible. A regular screen will likely have a single tear somewhere even when the framerate is perfectly matched.

Game physics/logic & how your screen is (almost) irrelevant
By default any software ran on a machine will attempt to run as fast as it possibly can, spitting out frame after frame in rapid succession. Additional code is introduced to limit the graphics side, Vsync in this case for games, to help making the image tear less. GTA is one series of games in a massive category where the entire game runs most of its logic/simulation in sync with the graphics pipeline. If the game runs at 30fps, most of the game will calculate movement and such at 30Hz. If the game runs at 60fps it will simulate at 60Hz and so on. This has nothing to do with your screen's refresh rate, the terminology used happens to apply to both. the "Hz" in the simulation is completely seperate from the screen refresh "Hz", pay careful attention to the context. Dont mix them together.

So before you say "ermagerd rockstar sucks at coding" think again. This is more common than you think, way more common. There's a reason why Call of Duty aims for 60fps even on consoles and why CS:GO players will flip their table if their game drops below 200fps (exaggeration, but almost true). There's no way in hell for them to see every single frame, but the game becomes so much more responsive from running faster, therefor if you care about competition you want to run your game as fast as possible no matter what you're doing.

Layman: The game gives no shits what the screen is up to, MOAR FRAMES! RIP screen tearing. Higher framerates improve response, and in the case of GTA along with a LOT of other games improves physics accuracy, so if competition matters to you dont settle for less. Response and accuracy always wins out. PC understeer you say? Console oversteer I say.

---

Lastly related to GTA V specificly...
As mentioned by Grumples_Plox (grimreaper977) and Sarah the physics simulation in games can either be tied to the rendering framerate (as Serenniel showed) or running on its own (as Grim pointed out with pCars). Because of how the world is designed in GTA nowadays Rockstar adds in a lot of wear and tear for the roads to make them feel more natural to drive on, less artificial perfection. Unfortunately for us, because of curb/bump boosting, this means that lower framerates will end up skipping past some of the imperfections in the road surface. Dont mix this up with being "faster by default" by running a high framerate, as it's not that simple. If both cars were to drive on a perfectly flat surface and the framerate is rock solid there should be no variance in their speeds at any point. But because GTA is constantly loading and unloading the world around you (known as streaming) it's impossible to get a truly solid framerate, constantly skewing results slightly. This is why some people say "fps boosting" is bullcrap and others say it's cheating, because both sides are right but also wrong. You're faster because your game is more correct, not because the game is broken when running at a higher fps.

It should be noted, however, that if the framerate drops a certain amount the game will compensate by giving you a temporary speed increase. It just so happens that it's a bit more sensitive to this above 110-120fps. But if the game is relatively stable this wont matter. I have gotten quite significant "frame boosts" by gradually dropping from 60fps to 30-ish before recovering back to 60. The game still ran fast enough to not go slow-mo, meaning I gained some time from this performance drop. Raw framerate doesnt matter, fluctuations in framerate does.

I may be misreading your post, but I don't think the above statement is good guidance, the data suggest differently doesn't it?   
Serenniel got the fastest time with locked framerate, vsync @144, it's faster than unlocked fluctuation it seems (the no-vsync no-limit was a second slower). 
Story is the only one to achieve glitch-like speeds with Nero, plenty of people try with 120-145fps and no dice, Story's 150-175 somehow made the difference. Fluctuation really the key? 
The Vsync 62/63 fps vs. 60 externally limited is night and day, none of them fluctuate (in my case it's obvious from CPU usage that game engine is crippled with in-game vsync 63fps, and similarly from externally 63fps)

It might be the definition of framerate that is a source of confusion. The framerate we see indicated with an overlay, is not necessarily corresponding with the speed the game engine runs. I guess the higher speed of engine, the more boosts can be recorded and factored in. And we don't have a good measure of game engine speed. 

Do you have an actual demonstration of the fluctuation effect?  If I understand you correctly, a stable 60fps might lose out to a fluctuation 30-60fps since the latter is granted speed compensation by the game when it realizes it is missing cycles? 

It almost seems to me as if R* tried to fix the FPS difference and added some limitations on engine speed and/or boost absorption, there was a time when a high fps system would mop the floor with a low fps system, drivers being equal, but not anymore. As the only demonstration of speed difference from high fps vs. 60fps is Story's 160+ (120-145 range fail to get a consistent advantage), it's tempting to think his pc is just overpowering the numbing fps-equalizer put in by Rockstar... (Nero top speed is probably a glitch, like a 811-light)