Originally Posted By vicious_cb:
Some long term testing of the Larue stealth 16" upper. Groups were fired over the course of several months.

Methods: 10-shot groups fired at 100 yards. Groups measured end to end the subtracted the bullet diameter (-.224) to find the extreme spread. Fired from a rest with a rear bag using a 2.5-10x NF scope at 10 power.

Ammo used: PPU 75gr MATCH factory loads.
For a review of this ammo see: http://www.ar15.com/forums/topic.html?b=3&f=16&t=372350

Sample size of 16 10-shot groups averaged: 1.575" (avg extreme spread)
Smallest group measured 1.026" (extreme spread)
Largest group measured 1.976 (extreme spread)
Most common measurements were 1.5-1.6" (extreme spread)

The rifle(usually wears a 2.5-10x scope at the range)
http://i232.photobucket.com/albums/ee52/j_rock001/IMG_0072.jpg

The sample size of 16 10-shot groups
http://i232.photobucket.com/albums/ee52/j_rock001/IMG_0073.jpg

And for all you internet commandos a 5-shot group measuring 0.65"
http://i232.photobucket.com/albums/ee52/j_rock001/IMG_0074.jpg

I dont have RSI shooting lab or a crazy testing set up like Molon but I hope this review does it justice

Originally Posted By vicious_cb:
Also something I noticed, alot of my groups tend to string vertically and I'm starting to think theres a barrel harmoics issue going on or is it just me? Anyone have any answers?



http://i232.photobucket.com/albums/ee52/j_rock001/IMG_0075.jpg

Originally Posted By tpelle:


Need to know more about how you were shooting - off of a bench with sandbags, prone or sitting with a sling?

Generally vertical stringing is due to inconsistent breathing - not getting your respiratory pause consistent between shots.

Originally Posted By TACKDRIVER68:



+1 on the breathing.
The best advice I can offer is to create a routine for yourself- mine is inhale, exhale slowly and regulate the length of pause with saying a word in your head and timing your trigger pull to the end of the word; 'pop', 'bang' or any other short word will do. You can also get 3-5 snap caps and do dime drills to work on the trigger control. I don't shoot more than once every two weeks or so, so sometimes I need to re-train myself on the basics. I do this the day before each range trip before I pack up the weapon.

Originally Posted By raffi383:

Today I went to a 100 yard range and shot off sandbags using the iron sights.  The rifle is a Rock River Arms lower with a White Oak upper.  The front post is about twice the width of the black on the target.  I used a bottom hold so I could see the black.



Here's how my best reloads look (10 shots, 0.938MOA):


I used 69gr Sierra bullets and 23.5grs of Alliant R15 powder.  23.5grs worked the best out of the 4 different amounts I shot.

Raffi

Originally Posted By PaoloOrange:
I made a Molon's 10-shot group that is too good to make sense to me and I can not understand it from a scientific point of view.

The first 7 shots averaged less then 0.1 MOA (???) (for the mean radius), then I was so excited that I start laughing loosing my concentration and I shot 3 fliers outliers. The 10 shot group is still an excellent (for me) 0.3 MOA (for the mean radius)

Believe me or not I shot that at 300 meter.

I feel like i rolled the dices 10 times and got "twelve" 7 times out of ten. Is the "10 shot group theory" still valid to evalute the rifle+ammo+ a very lucky shooter combination?

Originally posted by innocent_bystander:

I will continue to argue for 5 shot groups.

With 10 shot groups barrel heat starts to become a factor in accuracy and if you try to slow shooting enough to prevent heat build up then you run the risk of changes in conditions (wind).

Originally Posted By Molon:

...
In the accuracy spectrum involved with semi-automatic AR-15s, barrel heat has no statistically significant effect on the size of groups between 5-shots and 10-shots ...

Originally Posted By 700shooter:
This thread is about the bigest load of crap I have ever heard. There is only one "internet commando" here and that is the original poster molon (or is it moron?). I guess you think that just because large groups are what you shoot then that's what everybody should shoot. I've got news for you my friend, there are many nationaly and internationaly sanctioned bench rest shooting associations that would disagree with you. And by bench rest, I mean competition using equipment that is almost unrecognizable as a firearm. Purpose built firing platforms that are designed for pure accuracy. It doesn't take many shots
to determine if the action/barrel/ammo combination is sub moa with these platforms. But then again, I guess I should cut you some slack. After all, this is AR15.com. Maybe you would like to take your theories and pretty target pictures over to some of the bench rest association web sites. Prepare to get spanked sonny boy.

Originally Posted By 700shooter:
This thread is about the bigest load of crap I have ever heard. There is only one "internet commando" here and that is the original poster molon (or is it moron?). I guess you think that just because large groups are what you shoot then that's what everybody should shoot. I've got news for you my friend, there are many nationaly and internationaly sanctioned bench rest shooting associations that would disagree with you. And by bench rest, I mean competition using equipment that is almost unrecognizable as a firearm. Purpose built firing platforms that are designed for pure accuracy. It doesn't take many shots
to determine if the action/barrel/ammo combination is sub moa with these platforms. But then again, I guess I should cut you some slack. After all, this is AR15.com. Maybe you would like to take your theories and pretty target pictures over to some of the bench rest association web sites. Prepare to get spanked sonny boy.

Originally Posted By Skillshot:


lolz

Originally Posted By FALbert:
Thanks Molon for turning us on to this method.  

I have a crappy old functional laptop (by today's standards) I was planning to use at the range to obtain quick data about groups.  Turns out On Target freeware will work on Windows ME operating system.  Forget scanning th target, just take a digital photo of the target with a scale for reference.  This can be done at the range.  My only roadblock now is my digital camera is a current model with no USB drivers for Windows 98 or ME.  So I guess I have to get a newer laptop for range use.  A quick comparison of my "manual input from dial caliper to spreadsheet" method versus the On Target freeware software.  What is called "extreme spread" is called "maximum spread" and "actual mean spread" is called "average to center" by On Target software.  Temperature was in the low 40s and there was wind.  Targets were out at 100 yards.  Aiming point was dead center of target as best I could.  Shot numbers are for measuring purposes, not in the order they were shot.

For Hornady .223 75 gr TAP using WOA SS 18" upper, KAC 2 stage trigger, and Nightforce 2.5-10x NSX scope:
Maximum spread = 1.90" spreadsheet, 1.915" software
Average to center = 0.58" spreadsheet, 0.585" software
http://img.photobucket.com/albums/v508/FALbert/100%20yard%2010%20shot%20groups/TAP75grBTHP223meanradius58inch.jpg


For Portugese surplus 5.56 62 gr FMJ using Operator 16" CL upper, standard single stage trigger, and Aimpoint M3 with 3x magnifier:
Maximum spread = 6.75" spreadsheet, 6.780" software
Average to center = 2.26" spreadsheet, 2.115" software
http://img.photobucket.com/albums/v508/FALbert/100%20yard%2010%20shot%20groups/Portugese62grFMJ556meanradius226inc.jpg


For SSA 6.8 SPC 115 gr OTM using Kotonics 18" CL SPR, KAC 2 stage trigger, and 16x Super Sniper scope:
Maximum spread = 3.47" spreadsheet, 3.330" software
Average to center = 0.69" spreadsheet, 0.688" software
Maximum spread (excluding flyer) = 1.41" spreadsheet, 1.391" software
Average to center (excluding flyer) = 0.43" spreadsheet, 0.406" software
http://img.photobucket.com/albums/v508/FALbert/100%20yard%2010%20shot%20groups/SSA115grOTM68SPCmeanradius69inch.jpg


The software certainly was way easier and quicker to use and the results vary only because I rounded up to the nearest 100th by eye ball, and manual shot measurements are subjective.  Variations are for the most part less than 1/10 inch.  So the software is good to go.  More powerful optics on the 5.56 uppers would have improved groups as well.

Oh, and since this is AR15.com, gun porn of the hardware:
http://img.photobucket.com/albums/v508/FALbert/100%20yard%2010%20shot%20groups/ARsusedfor10shotgroups.jpg

Originally Posted By Molon:
He has watched the movie Heat 87 times

Originally Posted By 18thcentrury:
At the risk of getting it from all directions I'm going to throw in my two cents.  Since I don't know anyone here, I don't have any personal bias for or against any of the posters.

I'm no expert on AR's, but I am an expert on statistical data analysis and that is at the root of this whole discussion.  (Experience = Ph.D. with, depending on how you count it, about 20 years experience.)

Whether you are concerned with testing loads, or barrel condition, or sighting, or trigger control or whatever, when you use target analysis you are trying to use a statistical sample (sample population) to make a prediction about overall behavior of future shots (parent population).  Assuming the appropriate calculations are done correctly you can then draw some conclusions about the average behavior of future shots within the statistical uncertainties.  However, for this to be valid, there are some key points that have to be satisfied.

1) The sample must be representative.  In practice this means a couple of things.  First you can't arbitrarily throw out shots as "fliers" because they appear way outside the "group".  You can only throw out (and should) shots which you KNOW to be un-representative.  For example if someone slaps you on the shoulder just as you pull the trigger.  But just because a shot "looks" different does not mean that it can be excluded.  Second, the results automatically are representative of all of the statistical variation that goes into the sample.  So if, for example, you want compare the performance of two different loads in the same rifle, then you have to get one or more samples (groups or sets of groups) for each of the two loads, with no other differences.  In practice can you eliminate every single other variable? No, but you should be able to eliminate the statistically significant ones.  Or, at least be aware of and ensure that other variables should contribute equally to the all of the samples for each load.

2) The conclusions are about AVERAGE behavior.  There will always be variations from shot to shot, or group to group, but you can make definite quantitative statements about the average behavior within the statistical uncertainties.

If these requirements are satisfied, then there are a couple more statements which will be absolutely true, and this gets to the main point of this thread and the apparent dispute:

The larger the sample size, the smaller will be the statistical uncertainties in the results.  What this means is that on average a 10-shot group will provide a more accurate estimate of the parent group (all future shots) than will a 3-shot group.  In fact, the uncertainty in statistical results (for example the group center) can be calculated.  It depends on the particulars of the statistic being calculated, but a 10-shot group will result in a decrease in uncertainty over a 3-shot group of roughly a factor of 1.8.  Does this mean a 10-shot group is "the" size to use and a 3-shot group is worthless?  No, it all depends on what level of uncertainty you can live with and how many shots you are willing to make to get the uncertainty down to an acceptable level.  A fundamental, basic result of statistical analysis, whether you are talking about data from a high-energy physics experiment, or data from a set of holes on a target, is that the larger the sample size, the lower the uncertainty in the results.  In terms of the accuracy (how close the results are to some absolute reality) a 10-shot group is better than a 3-shot group, and a 50-shot group is better than a 10-shot group.  In designing a scientific experiment in which statistical analysis will be a major component of the final analysis, you always trade off practical considerations like expense and time needed to increase the sample size, vs. how much accuracy you need.

One other consideration is how much information you want to get out of your target analysis.  If you are just trying to get the group center then it turns out that you don't need nearly as large a sample size as if you are calculating the width or shape of the group.

There has been some discussion here of group sizes for larger sample sizes being "large".  The problem is, that the more shots you fire, the larger the extreme-spread will be.  This is a simple fact.  This is not something related to shooting, but is simply the nature of the underlying probability distribution function.  This is one reason why extreme-spread is not a particularly good statistic for measuring group size.  "Mean radius" (average distance from center is a technically more accurate term) is better, but in true scientific work an even better measure is a statistic generally referred to as "width".  What you really want to know is how densely shots are (or will be) placed near the center of the group.  Extreme-spread is a very crude measure of this.  It's popular because it is easy to measure.  The mathematically best measure is width, but it requires some fairly sophisticated math to do the calculations.

If samples are taken under identical (practically identical in the sense of statistical significance) conditions, then two 3-shot groups, and a 4-shot group are equivalent to a single 10-shot group, provided the raw data is combined BEFORE any calculations are done.  But there is nothing otherwise special about sample size.  A 10-shot group provides the same KIND of information as a 20-shot group but of different quality (that is, with different uncertainties in the results).  The problem is that if you don't know the uncertainties in the results it's impossible to know what the results are telling you without some foreknowledge of what to expect, and even then, without uncertainties your analysis ceases to be purely quantitative and falls back on making what is essentially an educated guess.

In discussions of target analysis (a.k.a. group analysis) uncertainties in the results always seems to be left out.  But without it you can't know whether the differences you are seeing are real, or whether they are just the result of the natural random variation which will occur from one shot (or group) to the next.  In the scientific community, if you don't present the uncertainty component of your results no one would give them any credibility because they would have no way of knowing what your results really mean, if anything.

A discussion of the uncertainties has been left out of the discussion in this thread, but that is probably because the software that has been referred to doesn't do uncertainty calculations.  I'm speculating now, but the software probably doesn't include it because the software author was probably just implementing a set of statistical formulae from a reference.  

Interest in the subject lead me to this thread and I thought my expertise justified a comment, but now I'm getting close to a point where I do have some bias in that I'm involved in selling software which does the uncertainty calculations needed for an actual scientific target analysis, so I'll stop here.

Bottom line –– it is mathematically correct to say that a 10-shot group yields results which on average are more accurate than a 3-shot group, but there is nothing magical about the numbers 3, 10, or 5.

Originally Posted By JoseyWales:
Here are the proper test conditions:

1) Obtain Caldwell shooting rest

http://www.midwayusa.com/mediasvr.dll/image?saleitemid=838651

2) Load 30 rounds in a mag and 5 rounds in a different mag

3) Bore sight upper and optic if not sighted in already

4) Sight in upper with 2 rounds and lock down rifle

5) Attach shoot-n-see target to cover holes

6) Shoot 30 rounds and measure all holes from the same point

7) Load data into Minitab Software

8) Run distribution identification to validate normal distribution or which one is correct

9) Calculate mean and standard deviation using correct distribution

10) Smallest standard deviation rifle is best for same number of rounds.  If you have 30 rounds for 3 different uppers and only 27 for fourth, you must use ANOVA to compare, assuming all normal distributions.

Originally Posted By markm:


I just finally finished watching Heat for the first time 2 weeks ago.  I rented it several years back but couldn't finish it because it was so bad.

Originally Posted By vicious_cb:


Strangely enough I dont have a vertical stringing problem when shooting my bolt gun. Im well of my aware of breath control when shooting for groups

Originally Posted By Muttt:


I have found that I have horizontal stringing when shooting my AR with .223 or 5.56.  But with my bolt action Remington 700 7mm Magnum, most all of my shots are vertically strung.  Any Comments as to why?  I was thinking weight of rifle.  AR much lighter than Remington.  Trigger pull is also much lighter on the Remington (4 to 4.5 lbs) that it is on my AR (new stock trigger between 7 and 8 lbs)  I shoot them both using the same breathing.  Could it be pistol grip vs typical wooden stock grip?  I shoot both using a bipod.  Any input would be interesting.  Sorry to sidebar the string with another sidebar....LOL

Mutt

Originally Posted By StevenH:
Some of you guys cant spot the center of a shot group to save your lives.

Originally Posted By vicious_cb:


I have no idea what causes it. Considering my 16 10-shot group sample size was taken over months and nearly all of them show the same pattern of vertical dispersion(note in the picture some of the targets are sideways) I think the evidence points to barrel. Again I have shot the upper on different lowers with different triggers and different stocks and wound up with the same result. The main trigger I used was a geissle hi-speed set to around 1lb which is pretty darn light! Ive ruled out the ammo as the problem since I use the same brand on several ARs and they do not exhibit a vertical dispersion pattern as wide as the Larue upper did. Perhaps I may have gotten a bad barrel but I doubt it. Its possibly some kind of barrel harmonics issue but Im not quite sure how to test something like that.

Originally Posted By Catman2:
I could shoot those - 1" groups with my Mini 14 but I don't want to

Originally Posted By Katy44:
Very nice group.  
What kind of barrel is on the CTR-02 ?