Rifled vs. Smooth barrel test by PunkWorks

[MondoMor]

If you don't know, "Punkworks" is an organization of several people, most visibly Bryce Larson (who posts here) and "Cockerpunk".

They did a test comparing a non-rifled (smoothbore) CP barrel against some Hammerhead rifled barrels.

http://www.techpb.com/forum/index.php?showtopic=36870

The raw data can be seen here: http://spreadsheets.google.com/pub?key=rcXQ_2IgHXgLAfekTFmxY5Q&output=html



MondoMor's interpretation: Rifling doesn't help accuracy in paintball. In fact, seems to hurt accuracy compared to a good quality smoothbore barrel.

Also, at 120 feet you're depending almost completely on luck to hit something with a paintball, no matter what barrel you're using.


The CP barrel is a bit more accurate than the Hammerheads, comparing the vector length.

The CP is also slightly more consistent. At 120 feet, the 95% of the shots out of the CP would be within about a 31" circle. With the best Hammerhead, shots would be within a 36" circle. **



I need to go back and read how they came up with the X and Y offsets (offsets from an arbitrary origin or offsets from the geometric center of all the impacts?).

Vector is the hypotenuse of a right triangle with the X and Y standard deviations forming the other two sides. I think.




** My statistics knowledge is rusty, somebody please check my assumptions. I have a rule of thumb that in a normal distribution, 95% of the data falls within 3 standard deviations of the mean. :)

[Thalion]

You are correct that 95% of a normal distribution belongs in +/- 3 standard deviations.

While this test comes out to conclude CP as superior to Hammerhead, I personally feel it is too close to be statistically significant. There simply are too many variables with paintball, and while this was a very well-done experiment, there are things outside anyone's control with paintball.

I would say most likely there is "no difference" between rifling and smoothbore in paintball.

[C4 Explosives]

It is minimal improvement within a controlled test, you have to consider windage, elevation to target, speed of target, paint weight, etc. They should conduct some close range tests as well.

I'm wondering the effects either has on lob, mostly because consistent lob is how I get most of my long range kills, which are many, and from prone, oh i've said too much.

[brycelarson]

hey guys.

maybe I can clarify a couple of things.

We shoot the board - but don't need to center the target - since the only calculations we make based on the impact locations is a Standard Deviation - which already includes an average of all the shots. So, if you take the mean of the x and y impact location - that's the center of where those shots hit. The Standard Deviation of the x and y give us a SD for each x and y. We then take the hypotenuse of those - which is easiest to think of as a radius on a circle - that radius being 1 standard deviation.

So, you take a circle twice as large as the vector - that's your %68 hit circle. Double that size (radius of 2 x vector) - and that's your %95 circle.

When you have a vector of 10, 11, 14 - in the range of this test - and you only have a variation of 3 or 4 - that's just not big enough. Maybe verging on have statistical significance - but prob not. We're full on the side of "no difference".

So, yes, at 120 feet hitting stuff with a paintball is pretty much luck. Taking the best barrel in this test - the CP 14" .685. That shot a 10.6 vector - so let's just round up to 11 to make the math easy. 11" radius means that %68 of the paint shot will land inside a 22" circle. That's about chest sized. Which sounds about right. If you have time to line up a shot and are shoting well - hitting a guy 3/4 of the time from 120 is about right. And 44" circle is the 95% target. That feels about right as well - if you are lined up you can put paint on a pallet bunker pretty much every time at 120'.

I always like to make up a little story to gut check our data - once in a while I'll catch a multiplication error or something by doing that :)

First strike and apex accuracy tests are coming up - stay tuned!

[Oldest Sage]

And that would account for why you usually hit me when we're opponents...you do the math.tongue:)

[MondoMor]

We shoot the board - but don't need to center the target - since the only calculations we make based on the impact locations is a Standard Deviation - which already includes an average of all the shots. So, if you take the mean of the x and y impact location - that's the center of where those shots hit. The Standard Deviation of the x and y give us a SD for each x and y. We then take the hypotenuse of those - which is easiest to think of as a radius on a circle - that radius being 1 standard deviation.

Thanks Bryce.

Could you guys include this as a "tab" in the data spreadsheets? For these kinds of tests it's pretty much copy-paste so I don't imagine it would take a lot of writing, and having a test procedure tagged onto your test data makes it easy for "peer review".


I thought I might have gotten hit with a First Strike round yesterday, but it turns out it was only rock-hard Impact. :D

[brycelarson]

yeah, we need to standardize our forms - that really shoud be included - since it's something we made up for our testing. :)

and yes, the FS rounds sting - the welt on my shoulder from Sat at Splat tag is pretty good.

I couldn't make it on Sunday - so CockerPunk, Chris Nathan and I went out for open play. We shot each other across tower - I drilled Chris from about 175' dead center in the chest first shot. He clipped me in the shoulder in return from the same distance. feels like being bunkered - but from distance.