smaller always better? , .684 vs. .688

[djinnform]

Thanks for all the great info. guys.

I shot 3/4 a case of paint yesterday through the .684 back, (no breaks) and the needle on my air tank barely moved. I used the air from last week! So, I'm pretty sure that the smaller bore increased my air efficiency to a noticeable difference at this point. But, my gut feeling is that the .688 is still slightly more accurate, but slightly less air efficient. Next time I play, I will switch back to the .688 and report back.

[GoatBoy]

i disagree on the grounds that we have been testing barrels for 5 years ... and EVERY SINGLE TIME we end up with the same conclusion. that repeatability is a sure sign of data convergence. anyone can do it themselves at there own field too

Then what you are talking about is a much larger aggregate of five years worth of data which represents effectively more shots. Which is fine, that is a way to tackle the problem. That data should be better collated or something -- you can't just point to some strings of 20 shots and expect a rational person to accept a conclusion regarding the standard deviation effectively. Based on that immediate data. I'm not disagreeing with your possible trend. I have no data either way.

Your five years of testing, however, has not proven that 20 shots alone is enough. Look at your data, look at my data, look at anyone else's data where you can pick a random run of 20 and compare it against the aggregate longer run. The data strongly shows that if you're looking for differences like what were indicated in that bore test, 20 shots is not enough.

i like that your testing things though, please come by punkworks and post up what you are working on! more data = more better ALWAYS.

Likewise, I appreciate your testing, I just... man, I wish I had the resources you guys have. I get to do everything myself, without proper facilities, without good equipment access, and without pretty much any assistance because quite frankly nobody quite understands what I'm doing anyways. That's why I can't tolerate errors in my own testing. It's literally too painful to go and redo it.

Trying to limit my forum exposure though. I'm actually probably going to start contracting it...

[cockerpunk]

Cockerpunks tests do not account for several important factors that he refuses to acknowledge.

Please elaborate, or link me to past discussions.

Frizzle can only ever be that vague.

is our testing incomplete? you bet. all testing, of everything is incomplete. thats why we still don't know everything there is to know.

but that doesn't mean its not useful, or we cannot learn to improve performance from it. i've been underboring for 4 or 5 years, as have many many other folks. and as i said, esp on our automags ... the extra efficiency is a godsend.

[nak81783]

Cockerpunks tests do not account for several important factors that he refuses to acknowledge.

Please elaborate, or link me to past discussions.

[Frizzle Fry]

Cockerpunks tests do not account for several important factors that he refuses to acknowledge.

[cockerpunk]

Why not? I'll give it a shot.

Blow test with seams engaging the barrel ID. Then downsize .003-.005" (or closest increment down).

But if I get bunkered in the back while swabbing out a messy barrel...

you'd be cleaning out that barrel either way. most barrel breaks are a function of the loading system and the ball/bolt/breach interaction, not the ball/barrel interaction.

[nak81783]

Why not? I'll give it a shot.

Blow test with seams engaging the barrel ID. Then downsize .003-.005" (or closest increment down).

But if I get bunkered in the back while swabbing out a messy barrel...

[cockerpunk]

I can't agree to anything that isn't backed by data, and that includes efficiency.

For average velocity, 20 is probably fine, judging by the data.

For standard deviation in velocity, 20 does not appear to be fine when the possible error in the standard deviation resulting from a sampling of 20 shots is potentially on the same scale as the trend that you were trying to indicate.

I ran some numbers with my own data, I will plug yours in once I figure out which data of yours you're willing to accept. Like I said originally, I looked at the last 100 of your first 500 because the previous shots were probably going crazy. Sure, you'll get an errant shot every now and then. That's what a larger sample is supposed to smooth over. Errant shots, and the probability of hitting a particularly bad run which is not reflective of the overall deviation, either on the + or - side. Shorter runs have a higher probability of showing up.

I did a run of 100 shots, same paint, same barrel, two different guns. (My testing was actually targeting efficiency.) I also generated random numbers with a similar FPS variance. Here's what it looks like:

Automag:
Window size 10, stdev variance 8.871928-2.936362=5.935566, mean variance 8.400000
Window size 20, stdev variance 6.990783-4.148240=2.842543, mean variance 5.700000
Window size 30, stdev variance 7.515195-4.657092=2.858103, mean variance 2.666667
Window size 40, stdev variance 6.978171-5.118844=1.859327, mean variance 1.750000
Window size 50, stdev variance 6.476457-5.346599=1.129857, mean variance 1.460000
100 shot (mean, stdev) = (284.24, 6.171161343153523)
Spyder:
Window size 10, stdev variance 8.875059-3.335000=5.540059, mean variance 9.800000
Window size 20, stdev variance 8.041177-4.073018=3.968159, mean variance 6.250000
Window size 30, stdev variance 7.793381-4.508485=3.284896, mean variance 5.600000
Window size 40, stdev variance 7.330302-5.196399=2.133903, mean variance 3.600000
Window size 50, stdev variance 6.867136-5.502838=1.364298, mean variance 2.400000
100 shot (mean, stdev) = (291.15, 6.189556208695906)
Random (100):
Window size 10, stdev variance 7.241854-3.326660=3.915194, mean variance 9.000000
Window size 20, stdev variance 6.863327-5.369505=1.493823, mean variance 4.600000
Window size 30, stdev variance 6.662987-5.476806=1.186182, mean variance 2.933333
Window size 40, stdev variance 6.587478-5.858098=0.729381, mean variance 2.475000
Window size 50, stdev variance 6.610320-5.842979=0.767340, mean variance 1.840000
100 shot (mean, stdev) = (285.48, 6.179928409902099)

Full data set is here:

https://docs.google.com/spreadsheet/ccc?key=0AtUb05cFvbYmdGpYZWFPOFJrci1Za2FEdFFxSU5vYkE&usp=sharing

And the python script I used for the calcs I can make public too; it's pretty trivial though.



The upshot is kind of obvious: the larger a window you take, the better the variance in the resulting standard deviation is.

How large a sampling is large enough? It's a matter of scale between the possible error in standard deviation, and how big the trend that you're trying to spot is.

If the possible swing in your results is larger than your trend, it's hard to say your results are conclusive.

i disagree on the grounds that we have been testing barrels for 5 years ... and EVERY SINGLE TIME we end up with the same conclusion. that repeatability is a sure sign of data convergence. anyone can do it themselves at there own field too

i like that your testing things though, please come by punkworks and post up what you are working on! more data = more better ALWAYS.

[cockerpunk]

So I'm trying to get on board with this. I took five paintballs and rolled them around at the top of various barrel inserts until they dropped in and I could blow them through. Using that same barrel insert that all five paintballs blew through, I then paid close attention to orientation. Since seams are typically the widest point of a paintball, I tried that first. I could only get three of the five to blow through with the seams acting as seals. Should I use the roll orientation method that got the five to go through, or the widest point method and size up until they all blow through. Then size down the .003-.005" you recommend?

The problem I have with this is chambering orientation is completely random. So no matter what method I use for the blow test, I could swing from underboring to overboring (or extreme conditions of either) given the paintball inconsistencies. This is why I stated earlier to get it relatively close, and fire away, hoping the paintball chambering orientation gods are on your side that day.

Thoughts?

then .003 to .005 works well because you don't have to be that close in your blowtesting. because every ball will stick in a .005 even if your blow test was off by .002. in reality though, we have a limited number of bores to test, so that becomes the problem with the blow test.

athomas: Any given barrel size essentially acts like a go/no-go gage. Fit testing at the seams seems most logical. However, orientation changes could lead to extreme overboring, and paintball inconsistencies could lead to extreme underboring. I would think you would not want to throw out the out of round paint unless you are going to sort your entire supply. As you stated earlier, if you don't account for those few fliers, they could become an issue.

cockerpunk: Do you see any validity to repeating your test, performing your blow test with properly oriented seams, noting that bore size, and then gradually decreasing bore size until you do start to see obvious evidence of underboring causing barrel breaks? Perhaps you could do this with economy, mid, and high level paints from a few different companies. This could lead to a very repeatable procedure of what the average player should do and expect given a certain quality of paint, and the degree of underboring that could be safely used. Do whatever sample size you feel is appropriate for average and standard deviation (I see n=30 tossed around a lot), but I would be more concerned with quality X paint breaks Y balls out of a pod/hopper/bag/case/etc. with Z level of underboring. You may want to show barrel breaks going two or three inserts in the overbore direction as well just to cover everything. I think all this in a single chart, would let each individual decide on consistency, efficiency, and relative risk for their personal preference.

I didn't get a chance to look through all the other data you linked earlier. Maybe you already have this.

yup:

http://punkworkspaintball.com/index.php?p=7&id=6

in this test we tested over, under, and bore matched, and found nothing of statistical significance in our breakage. sure, our sample sizes were not ideal, but we shot literally all day, out of that gun over 9000 rounds, and what we found was all barrels will break paint.

this pointed us to barrels not really being the cause of barrel breaks.

now, i'd love to shoot until we had something like 20 breaks out of each barrel too, but using those numbers, that would be something on the order of 6 cases from each barrel, and something like 4 days of shooting. sorry, we can't do that.

https://spreadsheets.google.com/pub?key=t1P1KlZM-B7_CqEMklBz7xQ&output=html

in this test we looked at insane underbores ... up to .018 or so. we did not start seeing breakage until .012+ so a minor .005" like above just isn't dangerous.

[GoatBoy]

alright, we can argue about st dev all day, and im happy to do that. but lets establish then a common ground ... we both agree 20 shots is effective at establishing the average of a sample. yes? this means then we can agree from the data set about the efficiency of barrels, yes? we can agree that under boring is more efficient. cause that is half the battle right there.

alright ....

ill give you consistency is a tricky one, if you note in the first test data, two .687 freak inserts shot quite different consistencies ....

I can't agree to anything that isn't backed by data, and that includes efficiency.

For average velocity, 20 is probably fine, judging by the data.

For standard deviation in velocity, 20 does not appear to be fine when the possible error in the standard deviation resulting from a sampling of 20 shots is potentially on the same scale as the trend that you were trying to indicate.

I ran some numbers with my own data, I will plug yours in once I figure out which data of yours you're willing to accept. Like I said originally, I looked at the last 100 of your first 500 because the previous shots were probably going crazy. Sure, you'll get an errant shot every now and then. That's what a larger sample is supposed to smooth over. Errant shots, and the probability of hitting a particularly bad run which is not reflective of the overall deviation, either on the + or - side. Shorter runs have a higher probability of showing up.

I did a run of 100 shots, same paint, same barrel, two different guns. (My testing was actually targeting efficiency.) I also generated random numbers with a similar FPS variance. Here's what it looks like:

Automag:
Window size 10, stdev variance 8.871928-2.936362=5.935566, mean variance 8.400000
Window size 20, stdev variance 6.990783-4.148240=2.842543, mean variance 5.700000
Window size 30, stdev variance 7.515195-4.657092=2.858103, mean variance 2.666667
Window size 40, stdev variance 6.978171-5.118844=1.859327, mean variance 1.750000
Window size 50, stdev variance 6.476457-5.346599=1.129857, mean variance 1.460000
100 shot (mean, stdev) = (284.24, 6.171161343153523)
Spyder:
Window size 10, stdev variance 8.875059-3.335000=5.540059, mean variance 9.800000
Window size 20, stdev variance 8.041177-4.073018=3.968159, mean variance 6.250000
Window size 30, stdev variance 7.793381-4.508485=3.284896, mean variance 5.600000
Window size 40, stdev variance 7.330302-5.196399=2.133903, mean variance 3.600000
Window size 50, stdev variance 6.867136-5.502838=1.364298, mean variance 2.400000
100 shot (mean, stdev) = (291.15, 6.189556208695906)
Random (100):
Window size 10, stdev variance 7.241854-3.326660=3.915194, mean variance 9.000000
Window size 20, stdev variance 6.863327-5.369505=1.493823, mean variance 4.600000
Window size 30, stdev variance 6.662987-5.476806=1.186182, mean variance 2.933333
Window size 40, stdev variance 6.587478-5.858098=0.729381, mean variance 2.475000
Window size 50, stdev variance 6.610320-5.842979=0.767340, mean variance 1.840000
100 shot (mean, stdev) = (285.48, 6.179928409902099)

Full data set is here:

https://docs.google.com/spreadsheet/ccc?key=0AtUb05cFvbYmdGpYZWFPOFJrci1Za2FEdFFxSU5vYkE&usp=sharing

And the python script I used for the calcs I can make public too; it's pretty trivial though.



The upshot is kind of obvious: the larger a window you take, the better the variance in the resulting standard deviation is.

How large a sampling is large enough? It's a matter of scale between the possible error in standard deviation, and how big the trend that you're trying to spot is.

If the possible swing in your results is larger than your trend, it's hard to say your results are conclusive.

[nak81783]

Always having two points of contact is agreed upon, until you start compressing the ball, at which point those two points become larger "point" surfaces, and eventually "line" surfaces around the entire circumference.

Given the previous statement that accuracy is unaffected by consistency up to around 15fps (although a relative statement, I tend to agree with it, because "aiming" inconsistency is a much larger factor), are we really just talking about efficiency? If I recall correctly, any insert in cockerpunk's data was below a 7.5fps st dev. And if consistency leads to nothing more than accuracy, but all inserts produce an acceptable consistency, should we really just be comparing efficiency vs. potential ball breaks?

[athomas]

athomas: Any given barrel size essentially acts like a go/no-go gage. Fit testing at the seams seems most logical. However, orientation changes could lead to extreme overboring, and paintball inconsistencies could lead to extreme underboring. I would think you would not want to throw out the out of round paint unless you are going to sort your entire supply. As you stated earlier, if you don't account for those few fliers, they could become an issue.

No matter what the orientation, you'll always have two points of contact if the paint is properly sized. Don't throw out the paint that is not round. Just don't use it for testing purposes. It throws off the results.

Other than that, if you know you have imperfect paint, you need to expect inconsistent shooting. Now you have to deal with it using a method you are comfortable with. Perhaps you should measure a large sample of balls and determine the smallest diameter of contact and use that as your barrel size. Then, every other ball will be using an underbore. Or, do the same using the largest measured diameter, except use an overbore. Personally, I use an "extreme" overbore. I sacrifice efficiency, but my shot consistency is good.

[nak81783]

athomas: Any given barrel size essentially acts like a go/no-go gage. Fit testing at the seams seems most logical. However, orientation changes could lead to extreme overboring, and paintball inconsistencies could lead to extreme underboring. I would think you would not want to throw out the out of round paint unless you are going to sort your entire supply. As you stated earlier, if you don't account for those few fliers, they could become an issue.

cockerpunk: Do you see any validity to repeating your test, performing your blow test with properly oriented seams, noting that bore size, and then gradually decreasing bore size until you do start to see obvious evidence of underboring causing barrel breaks? Perhaps you could do this with economy, mid, and high level paints from a few different companies. This could lead to a very repeatable procedure of what the average player should do and expect given a certain quality of paint, and the degree of underboring that could be safely used. Do whatever sample size you feel is appropriate for average and standard deviation (I see n=30 tossed around a lot), but I would be more concerned with quality X paint breaks Y balls out of a pod/hopper/bag/case/etc. with Z level of underboring. You may want to show barrel breaks going two or three inserts in the overbore direction as well just to cover everything. I think all this in a single chart, would let each individual decide on consistency, efficiency, and relative risk for their personal preference.

I didn't get a chance to look through all the other data you linked earlier. Maybe you already have this.

[athomas]

When you measure the size of a paintball for the purpose of determining a bore size, always measure the seam. That will always be the largest diameter. Any barrel you select at that size will always touch on two points. So, if you physically measure each paintball, you can accurately select ones that are consistently sized. Of course, this depends on the paint being in good condition and round. Any paint that is not round won't work like that but should not be used for testing anyway. You can determine that by measuring the balls at two points on the seam at 90 degree angles from each other. Any measured values that are noticeably different should have that paintball tossed out of the test.

[nak81783]

what i am saying is the actual measurement of a paintballs size is not a very good determination for what bore size it is.

this is why i most often refer to paint in "blow test size" not actual physical size. because the blow test functional size is the only one that actually matters.

So I'm trying to get on board with this. I took five paintballs and rolled them around at the top of various barrel inserts until they dropped in and I could blow them through. Using that same barrel insert that all five paintballs blew through, I then paid close attention to orientation. Since seams are typically the widest point of a paintball, I tried that first. I could only get three of the five to blow through with the seams acting as seals. Should I use the roll orientation method that got the five to go through, or the widest point method and size up until they all blow through. Then size down the .003-.005" you recommend?

The problem I have with this is chambering orientation is completely random. So no matter what method I use for the blow test, I could swing from underboring to overboring (or extreme conditions of either) given the paintball inconsistencies. This is why I stated earlier to get it relatively close, and fire away, hoping the paintball chambering orientation gods are on your side that day.

Thoughts?