mcarterbrown.com  

The Dead Zone Paintball Related Chat

Reply
 
Thread Tools Search this Thread
Old 11-04-2009, 09:30 PM   #21 (permalink)
Inch of Gold
 
Mayvik's Avatar
 
Join Date: May 2007
Location: Denver

Fan of EMR
CCM Fan
AKA Fan
Brass N Wood Fan
Palmers Fan
PBN Fan
Again..why do we care when the real desired result is X is better than Y or Y is better than X?
__________________
My Feedback
Mayvik is offline   Reply With Quote
Old 11-04-2009, 09:35 PM   #22 (permalink)
MCB Member
 
Fubarius's Avatar
 
Join Date: Apr 2006
Location: New Richmond, WI
Send a message via AIM to Fubarius Send a message via Yahoo to Fubarius

AKA Fan
Brass N Wood Fan
Palmers Fan
Quote:
Originally Posted by Mayvik View Post
Again..why do we care when the real desired result is X is better than Y or Y is better than X?
Because this is how we figure out WHICH is the better one. We can end up with two groupings of identical size, but one with the greater shot density in the middle is the better one. Marginally better depending on the overall size, but better none the less.
__________________
David Johnson
aka Fubarius

000110 200 11202 10 000020
012211 021 22110 22 121101
222001 222 10220 00 022212
Fubarius is offline   Reply With Quote
Old 11-04-2009, 09:35 PM   #23 (permalink)
Mod/mod
 
Drum's Avatar
 
Join Date: May 2006
Location: Trenchtown

Fan of EMR
Quote:
Originally Posted by Cockerpunk View Post

an analogy that most poeple should get is this one -

say two classes take identical tests.

in one class there are

5 in the 90s
10 in the 80s
15 in the 70s
10 in the 60s
5 in the 50s

in the other class there are

1 in the 90s
3 in the 80's
38 in the 70's
2 in the 60's
1 in the 50s

which class scored closer to 75% on the test?

you could argue the first one did, because the mean is 75%. you could also argue that the second one did too, because it is also 75% (close enough). then you could look at the range of the data, and both are over a 50% range. so you can't use the range either. this is like carter/x3/gimilsim/tibarms method. with only those tools, its possible to not be able to tell the difference between these two data sets.

so who scored closer to 75% on the test?

while, since we know the distribution of the scores, we can compute the standard deviation, which just figured from my butt is probably about 12-15% for the top one, and probably 7% or so for the bottom one.

so the second class is significantly closer to 75%.

Okay, I think I get the distinction you guys are seeking. That last analogy from CP helped.

Essentially, a paintball that impacts with a higher percentage in a given area, but also with a wider spread of deviant impacts would otherwise "score" the same as another paintball type that had a lower percentage of impacts in a given area but much with fewer wide-ranging deviations.

Now that I understand it (or think I do) I am beginning to find myself siding with those who have suggested that this is a distinction without a difference.

Thanks for breaking it all down for me.

D

Edited to include the eureka analogy.

Last edited by Drum; 11-04-2009 at 09:49 PM.
Drum is offline   Reply With Quote
Old 11-04-2009, 09:46 PM   #24 (permalink)
Inch of Gold
 
Mayvik's Avatar
 
Join Date: May 2007
Location: Denver

Fan of EMR
CCM Fan
AKA Fan
Brass N Wood Fan
Palmers Fan
PBN Fan
Quote:
Originally Posted by Fubarius View Post
Because this is how we figure out WHICH is the better one. We can end up with two groupings of identical size, but one with the greater shot density in the middle is the better one. Marginally better depending on the overall size, but better none the less.
But based on the test Carter and Paul did..you don't end up with two groupings of identical size. You end up with one ginormous one, and one less ginormous one. Seems pretty straightforward.
__________________
My Feedback
Mayvik is offline   Reply With Quote
Old 11-04-2009, 09:49 PM   #25 (permalink)
MCB Member
 
brycelarson's Avatar
 
Join Date: Apr 2007
Location: Minnesota

Drum, the reason that the statistical analysis is helpful can be seen exaggerated in Poe's opening post.

I don't know if you've read this or not - but Tom Kaye has another good write up on why it's important:

Automags Online - Tom's Tech Tips - Statistics Without Math, Tech Tip #5

his examples 14 and 15 are a great example.
__________________
-Bryce Larson
brycecollinlarson@gmail.com
Feedback
brycelarson is offline   Reply With Quote
Old 11-04-2009, 09:53 PM   #26 (permalink)
MCB Member
 
brycelarson's Avatar
 
Join Date: Apr 2007
Location: Minnesota

Quote:
Originally Posted by Mayvik View Post
But based on the test Carter and Paul did..you don't end up with two groupings of identical size. You end up with one ginormous one, and one less ginormous one. Seems pretty straightforward.
yes, and I'll say it again - my assumption is that their test is good and the future tests will show the same thing. BUT, using statistics we will know exactly how much difference there is in accuracy AND if it happens that the groups start to look very similar - we will have a solid way to determine the actual winner.
__________________
-Bryce Larson
brycecollinlarson@gmail.com
Feedback
brycelarson is offline   Reply With Quote
Old 11-04-2009, 10:00 PM   #27 (permalink)
A Secret Police Member
 
DPrekel's Avatar
 
Join Date: Mar 2006

Fan of EMR
CCM Fan
So, how does your test prove that other tests were not accurate?
__________________
Proud member of POOP!
DPrekel is offline   Reply With Quote
Old 11-04-2009, 10:05 PM   #28 (permalink)
Mod/mod
 
Drum's Avatar
 
Join Date: May 2006
Location: Trenchtown

Fan of EMR
Yeah, I read TK's and I just read it again.

I definitely understand the issue you raise.

I also understand the folks who are saying that the "circle size method", would seem to be enough for paintball.

From what I have been able to see of the testing .68 vs. .50 so far, I really think the difference between this version of .50 cal and typical .68 is probably already fairly clear to most observers.

I hope that they can use all this to make their next version better.

Good luck with your testing.


D

Last edited by Drum; 11-04-2009 at 10:15 PM.
Drum is offline   Reply With Quote
Old 11-04-2009, 10:09 PM   #29 (permalink)
MCB Member
 
brycelarson's Avatar
 
Join Date: Apr 2007
Location: Minnesota

Quote:
Originally Posted by DPrekel View Post
So, how does your test prove that other tests were not accurate?
It doesn't. ours just tells you more.
__________________
-Bryce Larson
brycecollinlarson@gmail.com
Feedback
brycelarson is offline   Reply With Quote
Old 11-04-2009, 10:15 PM   #30 (permalink)
MCB Member
 
Fubarius's Avatar
 
Join Date: Apr 2006
Location: New Richmond, WI
Send a message via AIM to Fubarius Send a message via Yahoo to Fubarius

AKA Fan
Brass N Wood Fan
Palmers Fan
Personally, I like both methods used together. Give me both the outer pattern size and the average distance from center. That way I can have a clear mental image of the pattern and what to expect.
__________________
David Johnson
aka Fubarius

000110 200 11202 10 000020
012211 021 22110 22 121101
222001 222 10220 00 022212
Fubarius is offline   Reply With Quote
Reply

  mcarterbrown.com » Paintball » The Dead Zone

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On

Forum Jump


All times are GMT -4. The time now is 06:35 PM.


Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Search Engine Optimization by vBSEO
© MCB Network LLC