[Egomaniacal]

cockerpunk, on 24 September 2012 - 03:53 PM, said:

again, scale.

the scale of the testing error reduction is relative to the usage the device will see. if i can't see it with a viced gun and normal field paint that i shoot every day, then as a consumer, it is useless to me. again, as a scientist, i can see value in further testing, as it might reveal more about the nature of the accuracy problem.

Fucking NO! Scale is not an appropriate objection here! High quality paint is WIDELY available, especially in tournament and scenario conditions. If there were an effect, I would EXPECT it to be destroyed by low quality paint!

Egomaniacal, on 24 September 2012 - 03:41 PM, said:

Yes, scale is important. But you're thinking completely linearly, assuming any potential effect size is going to be the same over all conditions. This is almost NEVER the case.

This post has been edited by Egomaniacal: 24 September 2012 - 04:00 PM

[cockerpunk]

going with a cars counter example.

last weekend i dynoed the redhead.

she pushed a health 240 rwph, which i was very happy about it. sadly, after 3 pulls it almost overheated on the dyno. why was this, i never have overheating problems in real life driving, even autocross and hard street driving. its because during a dyno day, they are pushing cars in and out of the dyno at the rate of 2-3 per hour, and setting up fans and ducting and getting realistic airflow over the radiator and intercooler is just not something they are going to fuck with.

does this mean my dyno results are NOT indicative of the performance of my car?

dynos, even different dynos of the same make and model produce as much as 5% different horsepower. heck, a 10 degree reduction is temp of ambient is probably good for 3-5 extra rwhp. should i then claim the dyno results null and void?

THE SCALE OF THE ERROR REDUCTION NEED ONLY BE AS GOOD AS THE USAGE OF THE DEVICE BEING TESTED.

[cockerpunk]

Egomaniacal, on 24 September 2012 - 03:59 PM, said:

cockerpunk, on 24 September 2012 - 03:53 PM, said:

again, scale.

the scale of the testing error reduction is relative to the usage the device will see. if i can't see it with a viced gun and normal field paint that i shoot every day, then as a consumer, it is useless to me. again, as a scientist, i can see value in further testing, as it might reveal more about the nature of the accuracy problem.

Fucking NO! Scale is not an appropriate objection here! High quality paint is WIDELY available, especially in tournament and scenario conditions. If there were an effect, I would EXPECT it to be destroyed by low quality paint!

Egomaniacal, on 24 September 2012 - 03:41 PM, said:

Yes, scale is important. But you're thinking completely linearly, assuming any potential effect size is going to be the same over all conditions. This is almost NEVER the case.

i have not been to BYOP field in ..... 4 years?

every scenario event (both nationally produced and locally produced) and national tournament (both NPPL and PSP) i have been to has been FPO. on a good day that is 2 choices of paint.

sorry, the vast majority of paintball play is done with very few options in paint. and most of the time, that paint is mediocre at best.

i wish that too wasn't the case, but it is.

and if the performance gain is destroyed by using widely available paint, then its not useful to me. i hate to bring it out as an example, but its the sweep system all over again. we did find it worked on some balls, some balls did have backspin ... but that doesn't make it a successful product. could an idea like the sweep system be refined into a useful product? maybe ... idk. someone should try and we'd be happy to test it out.

This post has been edited by cockerpunk: 24 September 2012 - 04:09 PM

[Troy]

cockerpunk, on 24 September 2012 - 03:53 PM, said:

what i am saying is the barrel isn't a useful product for the improvement of accuracy during typical paintball play.

I think this is understood, what he would like to see tested is the barrel under atypical conditions. If we come to grips with the fact that using premium tourney level paint isn't what most of the populace does, then we can find if there are some performance gains that are interesting to US (and not the general population).

The best speedball field in my area allows off site paint... so, I am interested in tests using better than "normal" paint. However, I recognize that my situation is, certainly, an edge case.

[andrewthewookie]

I'll throw in my $.02: Now that we've seen the barrel under "normal" conditions, I'd like to see it pushed to the limits; to see what it's capable of in those scenarios where it may have an affect that's not hidden in the noise.

[Snipez4664]

They did; the apex puts a pretty consistent spin on things, but it didn't produce more accuracy in your tests, right?

Some of which show a slight improvement, and all of which outperformed your control in the x dimension. (I get a P value of 0.49 though, so such a statement is completely worthless) https://spreadsheets...NvaGotX1E&gid=0

I don't have a good sense whether that vector represents better paint than in the 'cuda testing, since they were at different lengths. But, controlling spin didn't produce a big change either is my point

[Egomaniacal]

Alright. Pause. Breathe..... okay.

Here's what I hear you saying. Noise in a measurement need only be reduced to the desired resolution of the effect size you're looking for (actually not even that far, but that's besides the point). I hear you, and I probably understand your point better than you realize.

What I'm saying is you need to be careful in generalizing your conclusions. Tournament grade paint is widely available. It is within the realm of day-to-day play to be shooting very good paint (as far as paint goes), and is desirable to most people on this forum. Furthermore, if an effect that might cause an increase in accuracy is dependent on the paint quality - meaning relative changes in accuracy increase or decrease as the paint quality changes, you may have missed it by using the grade of paint you seem to have used. You cannot know the strength of this dependence well enough to say there are no circumstances where you might see an improvement in accuracy on the field. To be using high quality fresh paint is not outside of the realm of possibility for daily play, but that condition has not been covered by this test. And in fact, because the rails make contact with the ball in a limited number of small areas, I would fully expect any possible gains in accuracy to quickly disappear as paint roundness decreases.

The bottom line is we already know bad paint shoots poorly. I think it's fair to say most of us buy the best paint we can afford/find. So in my opinion it makes sense to use the best quality of paint you can for tests like these.

Again I appreciate the time, effort and money you have put into these tests. Thank you, you have certainly put a constraint on accuracy improvements within the conditions they were tested.

This post has been edited by Egomaniacal: 24 September 2012 - 04:32 PM

[Egomaniacal]

On a side note:

cockerpunk, on 24 September 2012 - 04:08 PM, said:

i hate to bring it out as an example, but its the sweep system all over again. we did find it worked on some balls, some balls did have backspin ... but that doesn't make it a successful product. could an idea like the sweep system be refined into a useful product? maybe ... idk. someone should try and we'd be happy to test it out.

How is someone who designs products supposed to do that if you're not more specific in the constraints of your test? Saying "this works on some paint [but we don't know which paint or why, and we can't tell you all the constraints in the system]" isn't useful information for someone trying to refine an idea.

Along those same lines, there may have been a breakdown in communication here. I was not going into this expecting much of anything. At least from my perspective this was not intended to be product testing, but rather proof of concept and prototype testing... as this is a prototype.

cockerpunk, on 24 September 2012 - 04:08 PM, said:

i have not been to BYOP field in ..... 4 years?

The field I frequent is BYOP, and if I play at PEV's they almost always have quality paint in stock (for a price). Your experiences are not the rule.

This post has been edited by Egomaniacal: 24 September 2012 - 04:45 PM

[UV Halo]

Here's something I'm wondering given this whole discussion of paint quality effects on barrel testing.

How does this paint compare to the DXS silver that most of the previous barrel testing was conducted with? I'm under the impression that the answer is 'comparable'.

If this is the case, why should we strive to introduce a higher grade of paint? It seems to me, that if we're to compare barrels (or more specifically barrel characteristics like under/overbore, control length, porting, etc) then we should stay consistent. Unless that is in my opinion, we plan on revisiting all of the previous concepts with higher quality paint.

[Egomaniacal]

UV Halo, on 24 September 2012 - 05:08 PM, said:

Here's something I'm wondering given this whole discussion of paint quality effects on barrel testing.

How does this paint compare to the DXS silver that most of the previous barrel testing was conducted with? I'm under the impression that the answer is 'comparable'.

If this is the case, why should we strive to introduce a higher grade of paint? It seems to me, that if we're to compare barrels (or more specifically barrel characteristics like under/overbore, control length, porting, etc) then we should stay consistent. Unless that is in my opinion, we plan on revisiting all of the previous concepts with higher quality paint.

Since you can't really compare the two sets of data due to differing and unrecorded conditions, that's not really an issue. The two data sets stand on their own, you really shouldn't compare them.

cockerpunk, on 24 September 2012 - 01:32 PM, said:

he pointed out that our vectors are larger then we typically get at 100 feet, which is true. which is why we shoot benchmarks like the 14 .685 CP to give us a known value to compare against.

It is my guess that the increase in vector is due to lower quality paint being used in this test.

This post has been edited by Egomaniacal: 24 September 2012 - 05:17 PM

[brycelarson]

Egomaniacal, on 24 September 2012 - 05:17 PM, said:

Since you can't really compare the two sets of data due to differing and unrecorded conditions, that's not really an issue. The two data sets stand on their own, you really shouldn't compare them.

It is my guess that the increase in vector is due to lower quality paint being used in this test.

I agree with both points. Each accuracy test is a separate set of data - other than comparing ranking in each test they should all be considered separate. Additionally I think it's fair to say that we have shot paint with lower vectors at the same range - indicating that this paint wasn't as good.

I'm torn. I totally agree that there is a possibility that out paint was obscuring some result in this test. To say otherwise is silly. We know that this paint shot more poorly than other paints we have shot. That said, I will continue to extol the idea that for a system to be useful to a player, any player, it needs to work in sub-optimal conditions.

I'm a low hanging fruit guy. Give me three options to improve something and I'll ask which one matters most. In this case, the low hanging fruit is the paint. If a barrel is going to provide useful performance improvement then the barrel needs to be the low hanging fruit. It needs to be the first move, the obvious choice to improve a system.

[cockerpunk]

UV Halo, on 24 September 2012 - 05:08 PM, said:

Here's something I'm wondering given this whole discussion of paint quality effects on barrel testing.

How does this paint compare to the DXS silver that most of the previous barrel testing was conducted with? I'm under the impression that the answer is 'comparable'.

If this is the case, why should we strive to introduce a higher grade of paint? It seems to me, that if we're to compare barrels (or more specifically barrel characteristics like under/overbore, control length, porting, etc) then we should stay consistent. Unless that is in my opinion, we plan on revisiting all of the previous concepts with higher quality paint.

i would rank the paint as worse then that silver. probably more like the bronze we had back then. which, bronze is a field paint at many fields ...

Egomaniacal, on 24 September 2012 - 04:26 PM, said:

What I'm saying is you need to be careful in generalizing your conclusions.

tell me where i said more then:

cockerpunk, on 24 September 2012 - 03:53 PM, said:

what i am saying is the barrel isn't a useful product for the improvement of accuracy during typical paintball play.

don't put words in my mouth and try to make me defend them. nowhere have i said anything more then this.

[Demon]

ARMGFGNBIODFSGVOAIBSVSVIBS


science.... something jargin,


i don't even know what's going on can i just see the video now?

[upriver]

If you have time between all this bickering about paint could someone please tell me what "vec." means on the spreadsheets? I'd be happy with a link to a page discussing the use of the pythagorean theorem on the standard deviations of the X and Y dimensions of the impacts.

Personally, I think you are doing it wrong. The equation you've used treats the X's and the Y's as independent numbers and I don't think that is the proper way to do it. Years ago I was using image analysis to do similar measurements and used the X and Y coords to calculate the distance from each shot to the center of the target or better yet the distance of each shot to the average center of the shots. This way you treat each shot as an individual data point and can run the statistics on the shot placement and not just on the X or Y coordinates. Using this method and a one way ANOVA shows there to be little statistical difference between the ten test combinations (i combined data from a couple of your spreadsheets).

I'm still interested in seeing the test setup and a description of how the tests were run.

[brycelarson]

upriver, on 24 September 2012 - 06:54 PM, said:

If you have time between all this bickering about paint could someone please tell me what "vec." means on the spreadsheets? I'd be happy with a link to a page discussing the use of the pythagorean theorem on the standard deviations of the X and Y dimensions of the impacts.

Personally, I think you are doing it wrong. The equation you've used treats the X's and the Y's as independent numbers and I don't think that is the proper way to do it. Years ago I was using image analysis to do similar measurements and used the X and Y coords to calculate the distance from each shot to the center of the target or better yet the distance of each shot to the average center of the shots. This way you treat each shot as an individual data point and can run the statistics on the shot placement and not just on the X or Y coordinates. Using this method and a one way ANOVA shows there to be little statistical difference between the ten test combinations (i combined data from a couple of your spreadsheets).

I'm still interested in seeing the test setup and a description of how the tests were run.

welcome!

I'm a bit busy right now but I'll track down our previous discussions about the "vector" measurement. We found that it's a great quick and dirty way to compare accuracy.

Since we're using the standard deviation of x and y we are effectively comparing each to the average center of the shots. There are slightly more accurate ways to analyze the data but we have found no instances thus far where our method has failed to produce the same results as a fuller analysis.

here's my thread on vector. you're obviously a stats guy - so feel free to start another thread with questions if you have them.

http://www.techpb.co...=1

[rntlee]

rntlee, on 22 September 2012 - 06:52 AM, said:

The best bet for creating a statistically measurable, better-shooting apparatus is via spin-reduction. This would require a completely redesigning the paintball marker from the ground up, and still the result would be measurable only with the best performing paints.

I was wrong here, there is one other way...impart axial spin to the ball. When I fooling with the sabots, I also shot sabots from a hammerhead barrel. Interesting results, with the smoothbore sabots, only seamless paints would show improved groupings. With the hammerhead I could get very dramatic improvement from even bafly dimpled paint.
Here's a couple of examples I recorded:
My link
My link

Not that this is any easier than achieving zero spins

This post has been edited by rntlee: 24 September 2012 - 09:37 PM

[Egomaniacal]

cockerpunk, on 24 September 2012 - 06:20 PM, said:

Egomaniacal, on 24 September 2012 - 04:26 PM, said:

What I'm saying is you need to be careful in generalizing your conclusions.

tell me where i said more then:

cockerpunk, on 24 September 2012 - 03:53 PM, said:

what i am saying is the barrel isn't a useful product for the improvement of accuracy during typical paintball play.

don't put words in my mouth and try to make me defend them. nowhere have i said anything more then this.

If you can't see that you're making a generalization from the conditions of your test in that very statement, I don't know how to help you.

This post has been edited by Egomaniacal: 24 September 2012 - 10:06 PM

[cockerpunk]

Egomaniacal, on 24 September 2012 - 10:05 PM, said:

cockerpunk, on 24 September 2012 - 06:20 PM, said:

Egomaniacal, on 24 September 2012 - 04:26 PM, said:

What I'm saying is you need to be careful in generalizing your conclusions.

tell me where i said more then:

cockerpunk, on 24 September 2012 - 03:53 PM, said:

what i am saying is the barrel isn't a useful product for the improvement of accuracy during typical paintball play.

don't put words in my mouth and try to make me defend them. nowhere have i said anything more then this.

If you can't see that you're making a generalization from the conditions of your test in that very statement, I don't know how to help you.

again, if it doesn't work with paint of normal quality, that people shoot every day, then its not a useful product for me.

sorry, thats just the cold hard truth. its the standard we apply to everything we test here in punkworks. the backspin bolt doesn't work because it doesn't put reliable spin on balls, the hammerhead doesn't work because supposedly it only works when the paint is perfect. this is by no means a new standard that we hold things to. as an accuracy improvement device, this barrel does not work.

as a research tool? very useful, and i look forward to trying it again. i do fear that no amount of design can mitigate muzzle effects, or that muzzle effects are to minor to fundamentally improve accuracy, and i'd love to give this barrel (or a next gen?) another shot at it.



and again, the fascinating thing about this testing is that a .680 barrel, with massive porting, somehow kept up at the chrono with the v1 and v2 lurker barrels. this is the true mystery of this test. i'd love to see the specs or at least a half decent description of the specs for the barrel, because then we could focus on figuring out what design components really did work well. this is a much more useful conversation the over and over repeating that if it doesn't work with normal paint, then it isn't a useful product.

[Troy]

cockerpunk, on 24 September 2012 - 10:19 PM, said:

if it doesn't work with normal paint, then it isn't a useful product.

*whistle* Logical fallacy: begging the question... ten yard penalty.

[cockerpunk]

Troy, on 24 September 2012 - 04:15 PM, said:

cockerpunk, on 24 September 2012 - 03:53 PM, said:

what i am saying is the barrel isn't a useful product for the improvement of accuracy during typical paintball play.

I think this is understood, what he would like to see tested is the barrel under atypical conditions.

Troy, on 25 September 2012 - 09:16 AM, said:

cockerpunk, on 24 September 2012 - 10:19 PM, said:

if it doesn't work with normal paint, then it isn't a useful product.

*whistle* Logical fallacy: begging the question... ten yard penalty.

mmmmmm self conflicting statements are a logical fallacy as well .....

This post has been edited by cockerpunk: 25 September 2012 - 09:43 AM