When comparing accuracy, I prefer to use just the X-component deviation. Vector is one of those numbers that’s great for the internet age. Short, sweet, and easily grasped. For the average TL;DR guy it’s a great number to have, Punkworks knocked it out of the park here. So why don’t I use it?
Half because of shot to shot velocity variation*. The problem is that you will never be able to eliminate it, and it will always influence the strike of the round higher or lower. It stops being a pure measurement of barrel performance and starts being a measurement of barrel performance with a little regulator performance thrown in. The best you can do here is mitigate it by getting a good reg, keeping it constant throughout your tests, and keeping the test range short enough that variations wont manifest in a big way.
In conventional paintball ammunition the ammo is already so inaccurate and short ranged that we can reasonably ignore the effect of velocity variation. Anything it adds at such short ranges will pale in comparison to the ammo’s natural dispersion. First strike rounds, on the other hand, are so much more accurate and (more importantly) so much longer ranged, that these variations have a much higher chance of manifesting in some meaningfully confounding way. Addendum: we use the same reasoning to explore the effects of a markers action type on FS accuracy.
The other half is because the y-deviation tends to matter less in terms of hitting a human target. Barring extreme physiologies or proned out targets, you will generally being shooting at a macro-shape that is taller than it is wide (a standing or kneeling human). As such, the target is more forgiving of high and low strikes than it will be of left or right. The average human male is six feet tall but at best, (i.e. the shoulders) under two feet wide. It gets worse as you travel up and down from the shoulders but to keep the math simple well exaggerate and call it an even two feet from top to bottom. That means your target is at least 66% more forgiving of lateral deviance than it is of longitudinal deviance, with the actual percentage being much greater. A realistic percentage would make an interesting calculus problem…
In essence, the Y-deviation is a less pure measurement of a barrels performance and its less critical to getting rounds on target. For these reasons, when I compare accuracy metrics with FS, I’m more interested in how the x-component behaves.
*There are other things that influence the strike up or down, but they are generally appropriate influences (like round to round variations in mass or caliber) or easily controlled (like the impact angle of the rounds relative to the target).
You can find the community discussion of this post at the following forums:
Whiskey Hammer Paintball 