Again, it's only possible in controlled environment where 100% of your focus and concentration is dedicated to reacting to the projectile coming at you, when it's clearly visible and you know it's coming. This is not applicable at to battlefield circumstances at all.
It's like saying that just because you can shoot bullseye with a rifle at 400 yards 10 times out of 10, that you should be able to do the same while sitting in a muddy trench with bullets and shrapnel flying around, while half starved and badly constipated. Ridiculous, isn't it?
Maximum997 and darksoulshin cover relevant points quite nicely.
Though to address your point, it seems the core of your argument is “battle is too hectic to perform optimally”. Yes this is true for most, hence why experience has been a relevant factor throughout human military history. Because when they gain experience, they learn how to perform under battle conditions.
The primary determining factor in its realism is not its mental limitations, but physical. It’s been shown that some are able to split bullets with a sword. Do I expect that of all swordsman? Obviously not. But clearly there’s a world of difference between an arrow and a bullet, and that’s the evaluation of the physical mechanics.
As I’ve highlighted to you, we have professional sports that deal with projectiles nearing or equal that level of reaction and projectile speed. Many enthusiasts are able to marginally perform deflections with less than a day of practice. It culminates in what the skill system is supposed to represent—proficiency and expertise. Hence why I stated in my initial point “To be fair”, as with current information and knowledge of its physical mechanics, there’s a fair possibility
P.S. I realize there is a very simple rebuttal that I and others have mentioned. If we understand the skill description clearly, it requires switching the sword block to deflect the arrow. If you have a physical sword intercept an arrow in flight, does it act upon the arrow? Yes obviously