AR Sponsor
Posted: 6/10/2010 1:38:42 PM EDT
[Last Edit: Molon][Edited]
|
ONE-SHOT “GROUPS”
For this simple, yet effective and illustrative test I compared the results of firing three consecutive 10-shot groups to that of firing thirty 1-shot “groups” in a row. By 1-shot groups, I mean that I fired one single shot on each of thirty different targets. As with my usual method, testing was conducted from a concrete bench-rest at a distance of 100 yards. The test vehicle was my Krieger barreled AR-15. The handguard of the rifle rested in a Sinclair Windage Benchrest and the buttstock rested in a Protektor bunny-ear bag. For sighting, I used a Leupold Competition scope (45X magnification) with a mirage shade attached.The scope was adjusted to be parallax-free at 100 yards. Wind conditions were monitored using a Wind Probe. The ambient temperature during testing was a balmy 72 degrees F. Just prior to the beginning of testing, I fired off a 10-shot control group from 100 yards using handloaded 55 grain Sierra BlitzKings that had an extreme spread of 0.74”. 
Testing was conducted using hand-loaded Hornady 55 grain FMJ bullets. The thirty, 1-shot groups were fired first. In order to mitigate possible effects of barrel heat on the points of impact of the 1-shot groups (and therefore the overall mean radius of the 30-shot composite group that would later be formed from the thirty individual shots) I allowed the barrel to cool between each individual shot. The barrel was allowed to cool for approximately 1.5 minutes between each shot. The barrel remained only slightly warm to the touch during the firing of all thirty 1-shot groups. I was able to wrap my hand around the barrel and hold it before each new shot was fired without the slightest bit of discomfort. Also, to mitigate possible effects of the so-called “heat soak” effect (the cartridge “soaking” heat from hot chamber walls while sitting in the chamber) on the points of impact of the individual shots, I single-loaded each shot just before firing it and left the bolt locked open after each shot to allow the chamber to cool. The thirty, 1-shot groups were overlayed on each other using RSI Shooting Lab to form a 30-shot composite group that had a mean radius of 0.46”. The 1-shot groups 
The three 10-shot groups were fired following the 1-shot groups. I fired at a slow steady pace, but no time was allotted between each shot for the barrel to cool nor was any time allowed for the chamber to cool. Each round was immediately single-loaded into the chamber after firing the previous shot. The three 10-shot groups had extreme spreads of 1.42", 1.22" and 1.27" for an average extreme spread of 1.30”. The three 10-shot groups were also over-layed on each other using RSI Shooting Lab. The mean radius for that 30-shot composite group was 0.42”. One of the 10-shot groups. 
Next, I measured (by hand) the individual radius of each shot from the statistical center of the 30-shot composite group formed from the 1-shot groups. I repeated the same procedure with the 30-shot composite group formed from the three 10-shot groups. These measurements gave me two sets of data for further comparison, with each set containing 30 data points on the accuracy/precision level of the two different methods of firing a total of 30 shots; that is the three 10-shot groups fired at a steady pace versus the thirty 1-shot groups fired slowly with time for the barrel to cool between each shot. Lastly, I ran a statistical analysis comparing the two data sets obtained in the manner described above using an unpaired, two-tailed t-test. The results showed that there was no “statistically significant” difference in the mean radii (and thus the accuracy/precision) of the two composite groups! As an aside, this test further reinforced the point that 55 grain FMJ bullets fired from a semiautomatic AR-15 are simply not capable of consitent MOA accuracy. Also, not forgetting our Internet Commandos in the audience, I fired a 3-shot group using the control load of 55 grain BlitzKings that had an extreme spread of 0.298" 
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By Zhukov:
Did you run a paried-sample t-test? What was your confidence interval and null and alternate hypotheses? That's on The List. 
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Molon, what is your barrel twist?
I have a Colt with a 1:13 twist Krieger SS barrel; an Alexander Arms billet receiver (really massive) and am limited to 52 grs. max.and get .250 to .373/.400 groups. But that's with either 52 gr. smk hpbt or custom 52/53 gr. hpbt bullets. But what I'm really asking about is your gas block. What type block, the diameter of your barrel and the bore of your gas block? I'm trying to find a .850 adj. gas block to shoot single shot with no "sproing", but no luck so far. |
I have honestly thought about sending Molon a MRP upper to do some proper accuracy tests with some of the most popular loads given I would never have the method, instuments and lastly and most importantly the SKILLS Keep doing what your doing pal, I always enjoy reading your accuracy and loading threads very much, and have learned quite a bit
|
To the OG gangsters of "BADNESS".."Mr. Clip", "Micro Man", and to the "Godfather" himself, THANKS for fueling my dreams!
The "Unicorn Man" and The "Sir"....Fuel accepted!
The "Unicorn Man" and The "Sir"....Fuel accepted!
|
I was kinda surprised when you said you were using Hornady 55gr FMJ's for this, since as you pointed out they're not sub-moa bullets. Guess it really didn't matter given the objective of the testing, though.
Nice job, as always. 
|
"I would rather ask a Roto Rooter scuba diver for advice on cranial micro surgery than ask for female advice here." - Piccolo
*The content above may or may not reflect my own views, but through random chance may express the views of other
*The content above may or may not reflect my own views, but through random chance may express the views of other
|
Originally Posted By Eric802:
I was kinda surprised when you said you were using Hornady 55gr FMJ's for this, since as you pointed out they're not sub-moa bullets. Guess it really didn't matter given the objective of the testing, though. Nice job, as always. The Hornady 55 grain FMJ bullets were chosen on purpose. Being less accurate bullets than say, Sierra MatchKings, they will naturally display a greater variation in shot-group dispersion. Since the two 30-shot composite groups had nearly identical mean radii, using the 55 grain FMJ bullets, this only serves to reinforce the validity of the results and conclusions. Remember, the goal here wasn't to shoot the smallest groups possible, it was to compare the shot-group dispersion of the groups obtained using the different firing methods. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By Molon: Originally Posted By Zhukov: Did you run a paried-sample t-test? What was your confidence interval and null and alternate hypotheses? That's on The List. Oh, what the heck - I'll mark this post "don't archive" anyway...  |
Daddy loves you. Now go away.
|
The results showed that there was no “statistically significant” difference in the mean radii (and thus the accuracy) of the two composite groups! (P= 0.49)
Excellent news for those of us too lazy to single feed our rifles when accuraccy testing!
OR do the testing for ourselves...... |
|
Originally Posted By PR361:
The results showed that there was no “statistically significant” difference in the mean radii (and thus the accuracy) of the two composite groups! (P= 0.49) Excellent news for those of us too lazy to single feed our rifles when accuraccy testing!
OR do the testing for ourselves...... The single-feeding was only used in order to allow the bolt to lock-back after each single shot and allow the chamber to cool during the "one-shot group" portion of the testing to eliminate the so-called "heat soak effect" on accuracy. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
| That 10-shot group you pictured looked pretty ugly considering the range and your equipment. Am I not getting something? With a 45X scope no less? Why not pretty much one hole? I am not trying to be a wise guy, but twenty years ago when I could see, I did better than that with a Bushmaster 20" 'National Match' and A2 upper. |
|
Perhaps I misunderstand; my take was that the composite groups showed that single feeding and allowing the chamber to cool between shots made no statistical difference in accuracy when compared to a slowfired 10 shot group. So, an accuracy test of single fed 75 gn Berger VLDs against 77 grn magazine fed sierras would not suffer significantly from the different feeding methods.
I'm a bit slow sometimes as I have a bad case of childrens, so what am I missing? |
|
Originally Posted By PR361:
Perhaps I misunderstand; my take was that the composite groups showed that single feeding and allowing the chamber to cool between shots made no statistical difference in accuracy when compared to a slowfired 10 shot group. So, an accuracy test of single fed 75 gn Berger VLDs against 77 grn magazine fed sierras would not suffer significantly from the different feeding methods. I'm a bit slow sometimes as I have a bad case of childrens, so what am I missing? Sarcasm, perhaps...
|
|
Originally Posted By 1saxman:
That 10-shot group you pictured looked pretty ugly considering the range and your equipment. Am I not getting something? With a 45X scope no less? Why not pretty much one hole? I am not trying to be a wise guy, but twenty years ago when I could see, I did better than that with a Bushmaster 20" 'National Match' and A2 upper. Your memory and reality don't match. That's what you're not getting. 'One hole groups', other than needed a a great rifle & shooter, require the most precise and CONSISTENT of ammo (brass, primers, powder, and bullets. Not some FMJ 'core stuffed from the rear' plinking bullets. If also look at his targets you'll see why it's always recommended for accuracy testing you need a minimum of 10 shots, 3 round groups just don't tell the full story. |
'98 Jeep XJ Owner
6.8 > 6.5
6.8 > 6.5
|
Originally Posted By Forest:
Originally Posted By 1saxman:
That 10-shot group you pictured looked pretty ugly considering the range and your equipment. Am I not getting something? With a 45X scope no less? Why not pretty much one hole? I am not trying to be a wise guy, but twenty years ago when I could see, I did better than that with a Bushmaster 20" 'National Match' and A2 upper. Your memory and reality don't match. That's what you're not getting. 'One hole groups', other than needed a a great rifle & shooter, require the most precise and CONSISTENT of ammo (brass, primers, powder, and bullets. Not some FMJ 'core stuffed from the rear' plinking bullets. If also look at his targets you'll see why it's always recommended for accuracy testing you need a minimum of 10 shots, 3 round groups just don't tell the full story. +1..I did ONE SHOT BULLSEYE stuff with .22lr. it realy comes down to "fliers" and ammo. In .22lr 50yard bullseye, a flier is just a lead pencil width away. |
There is no level playing field in life ~ Para069
|
Originally Posted By 1saxman:
That 10-shot group you pictured looked pretty ugly considering the range and your equipment. Am I not getting something? With a 45X scope no less? Why not pretty much one hole? I am not trying to be a wise guy, but twenty years ago when I could see, I did better than that with a Bushmaster 20" 'National Match' and A2 upper. Hint - look at the ammo he's using. I reload using the same bullets and the BEST group I've gotten with them is about 1.25" with a WOA upper. |
"I would rather ask a Roto Rooter scuba diver for advice on cranial micro surgery than ask for female advice here." - Piccolo
*The content above may or may not reflect my own views, but through random chance may express the views of other
*The content above may or may not reflect my own views, but through random chance may express the views of other
|
Originally Posted By PR361:
Perhaps I misunderstand; my take was that the composite groups showed that single feeding and allowing the chamber to cool between shots made no statistical difference in accuracy when compared to a slowfired 10 shot group. So, an accuracy test of single fed 75 gn Berger VLDs against 77 grn magazine fed sierras would not suffer significantly from the different feeding methods. I'm a bit slow sometimes as I have a bad case of childrens, so what am I missing? 
Yes, you’re missing something. This test had nothing to do with comparing the precision of single-loading versus magazine loading. Both composite groups were obtained via single-loading. From my original post: one-shot groups “Also, to mitigate possible effects of the so-called “heat soak” effect (the cartridge “soaking” heat from hot chamber walls while sitting in the chamber) on the points of impact of the individual shots, I single-loaded each shot just before firing it and left the bolt locked open after each shot to allow the chamber to cool.” 10-shot groups “I fired at a slow steady pace, but no time was allotted between each shot for the barrel to cool nor was any time allowed for the chamber to cool. Each round was immediately single-loaded into the chamber after firing the previous shot.” |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
Thank you for your patience with the slow learners in the audience.....So the test was to compare the accuracy of a cool barrel vs slow fired and eliminate the effects of heat sink from the chamber. I read right past the single loading BOTH groups
Learning is hard...... |
|
Originally Posted By PR361:
Thank you for your patience with the slow learners in the audience.....So the test was to compare the accuracy of a cool barrel vs slow fired and eliminate the effects of heat sink from the chamber. I read right past the single loading BOTH groups
Learning is hard...... The test showed that when fired in a slow steady manner from a semi-automatic AR-15 from a distance of 100 yards, three 10-shot groups of Hornady 55 grain FMJ bullets did not have any increase in shot-group dispersion when compared to “groups” of smaller sample size over-layed on each other. This was demonstrated by the two-dimensional analysis of the issue. I’ve demonstrated these concepts previously comparing 6-shot, 5-shot and 3-shot groups to 10-shot groups and took the testing to its logical culmination using 1-shot “groups.” The test also showed, that “barrel heat” did not induce any greater amount of shot-group dispersion in three 10-shot groups of Hornady 55 grain FMJ bullets fired in a slow steady manner from a semiautomatic AR-15 than thirty 1-shot “groups" fired with the barrel allowed to cool in between shots. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By Molon:
I think you used the most accurate 55gr FMJs around for this experiment. Your groups would have been MUCH bigger with Winchester bulk 55s; comparing a sample from 100 Winchester bulk 55s with a same-size sample from 100 Hornady 55s was an eye-opener. I measured mass and length of 10 randomly selected bullets from each box, and found that there was very little variation in the Hornady bullets, but wide variation in the Winchester bullets. In fact, the variation in the Win bullets was visible without measurement; their ogives weren't even consistent, tips were angled, and the amount of jacket closing the bases was even visibly different.
Originally Posted By Eric802:
I was kinda surprised when you said you were using Hornady 55gr FMJ's for this, since as you pointed out they're not sub-moa bullets. Guess it really didn't matter given the objective of the testing, though. Nice job, as always. The Hornady 55 grain FMJ bullets were chosen on purpose. Being less accurate bullets than say, Sierra MatchKings, they will naturally display a greater variation in shot-group dispersion. Since the two 30-shot composite groups had nearly identical mean radii, using the 55 grain FMJ bullets, this only serves to reinforce the validity of the results and conclusions. Remember, the goal here wasn't to shoot the smallest groups possible, it was to compare the shot-group dispersion of the groups obtained using the different firing methods. So with a fancy Krieger barrel and (what I assume was) appropriate diligence in loading the test ammo, even what are probably the best available bulk 55gr FMJs couldn't manage MOA groups. That's pretty much done, as far as I'm concerned. |
"--you can't conquer a free man; the most you can do is kill him."
Heinlein
Heinlein
|
Originally Posted By Nichonator:
You even managed to bring out Zhukovs lighter side. Lighter side, pfffft. Closest he comes to having a lighter side is smirking before he chokes a bitch. |
"I would rather ask a Roto Rooter scuba diver for advice on cranial micro surgery than ask for female advice here." - Piccolo
*The content above may or may not reflect my own views, but through random chance may express the views of other
*The content above may or may not reflect my own views, but through random chance may express the views of other
|
Can you post the raw data for each shot in the three 10 shot groups as well as the 30 1 shot groups?
Also, with regard to your methodology, how are you defining and measuring "extreme spread"? There are also some flaws in your methodology given that difference you are seeking to observe between a single 30 shot group and 30 1 shot groups may well be lost in the noise created by using a less accurate bullet, In other words, if the variance due to dispersion of the bullet itself is greater than the dispersion due to different methods in shooting the groups, you will have no significant difference in the groups. This is to be expected any time the observed variance is attributed to a limiting variable or component prior to reaching the level of discrimination needed to observe the effect you are trying to observe. For those you may not understand the concern. Let me provide a more extreme and obvious example. I'll take a shot out pakistani made copy of an AK 47 to the range and shoot it off a concrete bench versus off hand. If I get the same 6 foot' groups at 100 yards, can I conclude that the is no significant accuracy difference between shooting offhand and shooting off the bench? Most of us would agree that is not the case. The only difference between this "experiment" and the one above is a matter of degree. In effect, all you CAN say is that under the specified test conditions (including you as the shooter, and including the limitations imposed by a single trial) that there is no observable difference in the the two methods using 55 gr Hornady bullets. However you cannot generalize this conclusion to more accurate rounds and, with only one trial, you can't say much about the reliability of the results either. In short you have clearly not confirmed the hypothesis you are putting forward that there is no difference between the two methods. You have made a serious methodological error in thinking that the use of a less accurate bullet would tend to inflate and clarify the effect you are seeking to measure when instead it can logically be expected to obscure the effect. Your reputation here suggests you can do better. The tendency of most of the members here to buy into authoritarian arguments suggests and even demands you have a greater responsibility to either do a better job on the research or be a lot more honest about stating the limitations. To be less kind, the difference between your claims and those of the "internet commandos" you slam is only in a matter of degree. If anything, there is a higher degree of honesty in claiming 1 MOA within the stated statistical limits of it only being a 5 shot group. And as an aside, your results don't add any further support to the hypothesis you are also promoting that 55 gr FMJ ammuniiton is not capable of MOA accuracy. |
|
Originally Posted By rallain:
Molon, what is your barrel twist? I have a Colt with a 1:13 twist Krieger SS barrel; an Alexander Arms billet receiver (really massive) and am limited to 52 grs. max.and get .250 to .373/.400 groups. But that's with either 52 gr. smk hpbt or custom 52/53 gr. hpbt bullets. But what I'm really asking about is your gas block. What type block, the diameter of your barrel and the bore of your gas block? I'm trying to find a .850 adj. gas block to shoot single shot with no "sproing", but no luck so far. The barrel has a 1"7.7" twist. The barrel has a diameter of 0.937" at the gas block. I used a modified LaRue gas bock. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By GHPorter:
Originally Posted By Molon:
I think you used the most accurate 55gr FMJs around for this experiment. Your groups would have been MUCH bigger with Winchester bulk 55s; comparing a sample from 100 Winchester bulk 55s with a same-size sample from 100 Hornady 55s was an eye-opener. I measured mass and length of 10 randomly selected bullets from each box, and found that there was very little variation in the Hornady bullets, but wide variation in the Winchester bullets. In fact, the variation in the Win bullets was visible without measurement; their ogives weren't even consistent, tips were angled, and the amount of jacket closing the bases was even visibly different.
Originally Posted By Eric802:
I was kinda surprised when you said you were using Hornady 55gr FMJ's for this, since as you pointed out they're not sub-moa bullets. Guess it really didn't matter given the objective of the testing, though. Nice job, as always. The Hornady 55 grain FMJ bullets were chosen on purpose. Being less accurate bullets than say, Sierra MatchKings, they will naturally display a greater variation in shot-group dispersion. Since the two 30-shot composite groups had nearly identical mean radii, using the 55 grain FMJ bullets, this only serves to reinforce the validity of the results and conclusions. Remember, the goal here wasn't to shoot the smallest groups possible, it was to compare the shot-group dispersion of the groups obtained using the different firing methods. So with a fancy Krieger barrel and (what I assume was) appropriate diligence in loading the test ammo, even what are probably the best available bulk 55gr FMJs couldn't manage MOA groups. That's pretty much done, as far as I'm concerned. Yup, that was a "side-conclusion" of the test. The test load used weight-sorted, match-prepped brass, Federal Gold Medal primers and Vihtavuori powder. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By DakotaFAL:
Can you post the raw data for each shot in the three 10 shot groups as well as the 30 1 shot groups? Also, with regard to your methodology, how are you defining and measuring "extreme spread"? There are also some flaws in your methodology given that difference you are seeking to observe between a single 30 shot group and 30 1 shot groups may well be lost in the noise created by using a less accurate bullet, In other words, if the variance due to dispersion of the bullet itself is greater than the dispersion due to different methods in shooting the groups, you will have no significant difference in the groups. This is to be expected any time the observed variance is attributed to a limiting variable or component prior to reaching the level of discrimination needed to observe the effect you are trying to observe. For those you may not understand the concern. Let me provide a more extreme and obvious example. I'll take a shot out pakistani made copy of an AK 47 to the range and shoot it off a concrete bench versus off hand. If I get the same 6 foot' groups at 100 yards, can I conclude that the is no significant accuracy difference between shooting offhand and shooting off the bench? Most of us would agree that is not the case. The only difference between this "experiment" and the one above is a matter of degree. In effect, all you CAN say is that under the specified test conditions (including you as the shooter, and including the limitations imposed by a single trial) that there is no observable difference in the the two methods using 55 gr Hornady bullets. However you cannot generalize this conclusion to more accurate rounds and, with only one trial, you can't say much about the reliability of the results either. In short you have clearly not confirmed the hypothesis you are putting forward that there is no difference between the two methods. You have made a serious methodological error in thinking that the use of a less accurate bullet would tend to inflate and clarify the effect you are seeking to measure when instead it can logically be expected to obscure the effect. Your reputation here suggests you can do better. The tendency of most of the members here to buy into authoritarian arguments suggests and even demands you have a greater responsibility to either do a better job on the research or be a lot more honest about stating the limitations. To be less kind, the difference between your claims and those of the "internet commandos" you slam is only in a matter of degree. If anything, there is a higher degree of honesty in claiming 1 MOA within the stated statistical limits of it only being a 5 shot group. And as an aside, your results don't add any further support to the hypothesis you are also promoting that 55 gr FMJ ammuniiton is not capable of MOA accuracy. While you're at it, can you quantify just how many angels can in fact dance on the head of a pin? The man isn't trying to disprove some religiously-held concept here using the Scientific Method. He clearly laid out the methodology and materials he utilized, and the conditions under which he conducted his tests. It's one of those "take it for what it is" studies, and I suspect most of us would obtain similar results (with adjustments for equipment and shooter skill), and as such it's meaningful to the majority of us. I think Molon made a good case for the point he's illustrating ... if you disagree, that's cool. Just don't over-analyze it and make it into something it isn't. . . |
|
That bit of sarcasm at the end flew right over my head at first. Then I got the joke... Your posts are so technical that the humor sounded like an error at first - then I realized that the 'error' was so huge that it was statistically impossible for the author to have made it. "Scotty, you're confined to quarters for the duration of this mission..." "Aye, Captain. And thank you - that'll give me some time to catch up on my technical journals." |
Social de-evolution appears complete. Typical human behavior is now entirely absent.
|
Originally Posted By Molon:
So your test ammo was some of the most expensive 55gr FMJ loads ever made. The test load used weight-sorted, match-prepped brass, Federal Gold Medal primers and Vihtavuori powder. 
Which VV powder was it? |
"--you can't conquer a free man; the most you can do is kill him."
Heinlein
Heinlein
|
Originally Posted By GHPorter:
Originally Posted By Molon:
So your test ammo was some of the most expensive 55gr FMJ loads ever made. The test load used weight-sorted, match-prepped brass, Federal Gold Medal primers and Vihtavuori powder.
Which VV powder was it? This one. 
N135
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By MikefromTX:
The point is that he was not using the "scientific method" as it is my no means a legitamate experiment or even quasi experiment. But the use of data and statistics makes it appear that way to many, if not a majority of the members here. And sadly, it obscures the point that the conclusions are not in fact useful to most people. as they are not valid.
While you're at it, can you quantify just how many angels can in fact dance on the head of a pin? The man isn't trying to disprove some religiously-held concept here using the Scientific Method. He clearly laid out the methodology and materials he utilized, and the conditions under which he conducted his tests. It's one of those "take it for what it is" studies, and I suspect most of us would obtain similar results (with adjustments for equipment and shooter skill), and as such it's meaningful to the majority of us. I think Molon made a good case for the point he's illustrating ... if you disagree, that's cool. Just don't over-analyze it and make it into something it isn't. LIke it or not, there is a level of dishonesty in taking that approach, especially when you start throwing around conclusions in a couple areas where the data is no better than the data that the researcher frequently critisizes. Stating the test conditions without being up front about the obvious and not so obvious limitations is just bad form. |
|
Originally Posted By DakotaFAL:
Can you post the raw data for each shot in the three 10 shot groups as well as the 30 1 shot groups? Also, with regard to your methodology, how are you defining and measuring "extreme spread"? There are also some flaws in your methodology given that difference you are seeking to observe between a single 30 shot group and 30 1 shot groups may well be lost in the noise created by using a less accurate bullet, In other words, if the variance due to dispersion of the bullet itself is greater than the dispersion due to different methods in shooting the groups, you will have no significant difference in the groups. This is to be expected any time the observed variance is attributed to a limiting variable or component prior to reaching the level of discrimination needed to observe the effect you are trying to observe. For those you may not understand the concern. Let me provide a more extreme and obvious example. I'll take a shot out pakistani made copy of an AK 47 to the range and shoot it off a concrete bench versus off hand. If I get the same 6 foot' groups at 100 yards, can I conclude that the is no significant accuracy difference between shooting offhand and shooting off the bench? Most of us would agree that is not the case. The only difference between this "experiment" and the one above is a matter of degree. In effect, all you CAN say is that under the specified test conditions (including you as the shooter, and including the limitations imposed by a single trial) that there is no observable difference in the the two methods using 55 gr Hornady bullets. However you cannot generalize this conclusion to more accurate rounds and, with only one trial, you can't say much about the reliability of the results either. In short you have clearly not confirmed the hypothesis you are putting forward that there is no difference between the two methods. You have made a serious methodological error in thinking that the use of a less accurate bullet would tend to inflate and clarify the effect you are seeking to measure when instead it can logically be expected to obscure the effect. Your reputation here suggests you can do better. The tendency of most of the members here to buy into authoritarian arguments suggests and even demands you have a greater responsibility to either do a better job on the research or be a lot more honest about stating the limitations. To be less kind, the difference between your claims and those of the "internet commandos" you slam is only in a matter of degree. If anything, there is a higher degree of honesty in claiming 1 MOA within the stated statistical limits of it only being a 5 shot group. And as an aside, your results don't add any further support to the hypothesis you are also promoting that 55 gr FMJ ammuniiton is not capable of MOA accuracy. +1 It appears that this is 2 "experiments" jammed in to one. The variables, bullets and shooting style, are not isolated. That said, it still makes me feel good about my shooting with 55gr FMJs |
|
Originally Posted By Zhukov:
There appear to be obvious "fliers" in your groups. Just throw 'em out and recalculate the results! "Fliers" are so passe. The Internet Commando has evolved and has added new tools to his box to discount even more shots to shrink his group size. These new tools include "target movement" and (my personal favorite) "The Butterfly Shot." Rather than admit that your rifle/ammunition aren't as accurate as you believe, claim that just before the bullet impacted with the target, it collided with a butterfly which deflected the shot!  Seriously, I couldn't make this stuff up.
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By Zhukov:
Originally Posted By Molon:
Originally Posted By Zhukov:
Did you run a paried-sample t-test? What was your confidence interval and null and alternate hypotheses? That's on The List.
Oh, what the heck - I'll mark this post "don't archive" anyway...
Exellent. I added this thread to the Reference Links. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By Molon:
. Rather than admit that your rifle/ammunition aren't as accurate as you believe, claim that just before the bullet impacted with the target, it collided with a butterfly which deflected the shot! http://www.box.net/shared/static/qnki7q98ug.gifSeriously, I couldn't make this stuff up.  That is one I have not heard before. How did you keep from laughing?
|
'98 Jeep XJ Owner
6.8 > 6.5
6.8 > 6.5
|
Originally Posted By Forest:
Originally Posted By Molon:
. Rather than admit that your rifle/ammunition aren't as accurate as you believe, claim that just before the bullet impacted with the target, it collided with a butterfly which deflected the shot! http://www.box.net/shared/static/qnki7q98ug.gifSeriously, I couldn't make this stuff up. That is one I have not heard before. How did you keep from laughing?
87GN pointed that one out to me. It was from a thread on this very website. |
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
Originally Posted By PR361:
I gotta get a better spotting scope, I'm pretty sure my range is INFESTED with butterflies! Just think of the small groups that you'll be able to claim after you discount all The Butterfly Shots.
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
|
"The results showed that there was no “statistically significant” difference in the mean radii (and thus the accuracy/precision) of the two composite groups! (P= 0.49) "
.42" mean radius with 1.5min wait time - .38" mean radius with zero wait time = .04" difference. .04 as a percentage of / .38 = .105 = 10.5% Reverse Proof: .38 * 1.105 = .4199 The group radius increased by 10.5% –– At what point would your statistical measuring declare "yes, there was a significant change?" I know I would find a 10% annual salary increase as very significant, versus the insignificant actual increases of 2 or 3% I've been getting last several years.
|
"A man with one AR is like a man with one arm." -me (pun intended)
|
Originally Posted By chris65:
"The results showed that there was no “statistically significant” difference in the mean radii (and thus the accuracy/precision) of the two composite groups! (P= 0.49) " .42" mean radius with 1.5min wait time - .38" mean radius with zero wait time = .04" difference. .04 as a percentage of / .38 = .105 = 10.5% Reverse Proof: .38 * 1.105 = .4199 The group radius increased by 10.5% –– At what point would your statistical measuring declare "yes, there was a significant change?" I know I would find a 10% annual salary increase as very significant, versus the insignificant actual increases of 2 or 3% I've been getting last several years.
Statistical significance is not merely "boy that's a big difference." In statistics, a result is called statistically significant if it is unlikely to have occurred by chance. There is a huge amount of calculation behind what constitutes "significance" in any statistical analysis. I do not know Molon's "confidence level," but typically 0.95 is used. If he used that confidence level, that means that there is a 95% probability that the differences between those two groups were caused strictly by chance - and only a 5% probability that the differences were caused by something else. |
"--you can't conquer a free man; the most you can do is kill him."
Heinlein
Heinlein
|
Originally Posted By PR361:
With thanks to Molon for his insight into the cause of my shooting deficiencies, I now conclude that I am, after all, a great marksman! Photo 1http://i933.photobucket.com/albums/ad172/sureshot31/Varget.jpg Improved Version! http://i933.photobucket.com/albums/ad172/sureshot31/Butterflyimproved.jpg Sorry, I couldn't help myself! 
|
All that is necessary for Trolls to flourish, is for good men to do nothing.
In God We Trust. Everyone else needs to post data.
In God We Trust. Everyone else needs to post data.
AR Sponsor