Discorz

I did this test to see if I could practice my aim in 2D online game like AimBooster so I could just flick shot in 3D space. For that I had to match my in game and desktop sensitivity
....I tried to match horizontal distance mouse travels on fullscreen picture and in game, from point A to B (because of flick shot muscle memory).
As picture above shows there is no correct sens value that perfectly matches desktop and in game sensitivity. Results start to distort because moving around in 2D and 3D space are not the same.
BUT definite most accurate value would be exactly somewhere in between 2.5 - 3.5. Higher or lower than those start to get inaccurate.
3.5 would be for higher and 2.5 for lowersensitivity players.
3.5 is perfectly matching desktop sens only in radius of 250px from center of screen (500px diameter), but this sens feels faster than desktop.
With 2.5 you would be most accurate on edges of the screen and it kinda feels too slow.
Value that would make sense to use is 3.1 (average) and that one is most accurate in the 1st and 3rd quarter of the screen wich is in between center and edge of screen (yes 3.1, not 3.0 because of fov image stretching)
This type of sensitivity matching is connected and depends on FOV (field of view). Any other FOV does not apply to my results anymore, I think.
For test I used: win10 sens 6/11 (default), no accel | CSGO no accel, raw input on | fov_cs_debug 90 (default) | native 1080p/16:9 res | 24.5" monitor | Logitech G102 @200/300/400/800 dpi

I don't think you get it. Distortion will still be present regardless of your monitor match percentage. However choosing 00% may be the most logical cuz you ain't matching to a "random" point on your screen. If you go up or down in the game  world you will still spin instead of going left or right due to how the matrix is designed.