5 shot groups "kinda worthless"?

[LiveLife]

I don't understand the obsession with groups out of a firearm meant for self-defense

Actually, OP is shooting 22LR and doing comparison range test at 75 yards (So we are talking about rifle) of 20+ brands to determine which is better.

comparing 20-some brands of .22 in ... three different guns ... 75yd test)

And the question posed of "5 shot groups 'kinda worthless'?" is based on majority of groups not being readily distinguishable.

Two or three brands were "best" and duplicate-ably better by a modest margin...and two or three brands sucked. But the VAST majority were simply average and tough to distinguish.

I was trying to do the same with my 10,000+ round comparison test of 25+ brands/weights/lots of 22LR out of 10/22 and T/CR22.

And just like OP, some ammunition such as CCI SV, Blazer and Aguila 40 gr boxed LRN/CPRN consistently produced smaller sub 1" groups at 50 yards and Federal 36 gr CPHP loose bulk 525/550 packs consistently produced larger 2"+ groups. But the majority varied in group size with flyers that made small sample size of 5/10 groups hard to distinguish which was better.

In answering OP's question of "5 shot groups 'kinda worthless'?", it wasn't until large enough sample size of groups were shot (with confirming reference groups of Aguila 40 gr CPRN that consistently produced 3/4"-1" groups with cold/hot barrel verified at the start of range session, during and at the end of target sets of 8-12 groups) that 5/10 shot groups were simply subset of 50/100 shot groups.

So my takeaway from the 10,000+ round comparison testing is small sample size of 5 shot groups may not provide enough information to determine the actual group size of particular ammunition as random samples of 5 rounds may not represent the full extremes of barrel harmonics/swing until large enough sample groups are shot.

What I learned while accurizing both rifles by free floating the barrel, lightening the trigger, better securing the scope ring/mount bolts with Loctite/torque wrench, securing the receiver to stock, using machined/pinned firing pin bolts, etc. to reduce shooting variables (particularly when shooting longer 100 yard groups); is there are other shooting variables such as barrel heating up, bullet drop, wind pushing bullet and shooter fatigue that need to be considered also and why I suggested to OP to test at 50 yards instead of 75 yards to better eliminate the bullet drop factor (At 22LR velocities, bullet drop at 75 yards could be significant to vertically elongate group size due to variation in muzzle velocities).

Excluding these shooting variables, box of 50 rounds may provide enough information from ten 5-shot groups if the ammunition does not produce too much barrel harmonics/swing. But if particular ammunition produce larger swing of muzzle, 50 rounds may not represent full extremes of muzzle swing we see as "flyers" on target until more rounds are shot. (4 rounds could exit the muzzle near bore axis and 1 round at outer extreme of muzzle swing to produce "flyer" on target away from tight 4 round holes)

Regarding whether to include or exclude flyers, unless other shooting variables could be isolated and shot enough sample size to know the extremes of muzzle swing (Let's say particular ammunition consistently produce 3/4"-1" at 50 yards regardless of range session and cold/hot barrel) and you could be certain that it wasn't shooter induced input on trigger/stock, IMO all flyers need to be included in group size, especially if the sample size is small.

 

[ev780]

Thought about your comment for a bit.... Here are some practical ways to use the d2 constant:

1. Ever wonder how repeatable your chronograph is? Set up two identical chronographs in series, and fire several cartridges across both of them, in random order. Adjust the second one's readings for the expected velocity drop. Now you have matched pairs of data. Take the positive difference in each pair. Average the differences. Divide the answer by the d2 constant for 2 data, and that gives you the standard deviation of the random error in the chronograph. That's your repeatability. I did that with a pair of Shooting Chronys, and they performed well.

2. Working up a load? Make matched pairs of cartridges at 1/4 or 1/2 grain increments. Chronograph them in random order. Again, you have matched pairs of data. Do a scatterplot of MV as a function of charge to see the result. Then take the positive difference in each data pair, average them, and divide by the d2 constant. Now you have the variation of your MV and a nice picture of how much MV the loads produce.

3. If you're shooting 5 shots to get the mean and SD of MV, just skip the calculation of SD. It's not very intuitive for most folks, and the highest speed minus the lowest speed (range) has all the information you're likely to get about variation, and with 5 shots it's a better estimate than SD.

4. If you really want to analyze a target, shoot 5 shots at each of a few targets. For each target, find the center of the group and measure the distance of each shot to it. Subtract the farthest distance from group center from the nearest distance to group center. That gives you a range. Average the ranges of the targets. Divide the average by the d2 constant for 5 data. This gives you a standard deviation for your shot placement. 68% of your shots will fall within plus or minus that number. 95% will fall within twice that number, and 99.7% will fall within three times that number. Those numbers will apply pretty well whether the POIs are normally distributed or not.

All of the above assume that "special cause" has been removed.

So it does lead to some results that would be difficult to obtain otherwise.

Scary how much of that I actually got!

Thanks Professor! That was cool.

 

[JFrank]

Actually, OP is shooting 22LR and doing comparison range test at 75 yards (So we are talking about rifle) of 20+ brands to determine which is better.

Isnt the op just commenting on a video ?

 

[Varminterror]

Reiterating here:

There is a BIG difference between “apparently different” and “actually different.”

The “reloader’s problem” isn’t that they aren’t shooting statistically valid sample sets, it is that reloaders are making inferences, and making decisions, based on statistically invalid sample sets.

As mentioned a few pages ago, when a reloader throws together a matrix of powders and bullets, then shoots a small number of rounds of each combination, IN MOST CASES, the results are not statistically, significantly different. Even when one group might be twice as large as the other, or three times, it’s MORE COMMON THAN NOT for those groups to NOT actually represent differentiated results.

This is really the basis for the origin of statistics as a body of knowledge. Folks needed to understand if populations were actually different, or only apparently different.

So it’s really quite simple for many of us - we pick a combination we want to work, we make it work with VERY little effort or expense, and then we do deep dives into statistically valid sample sets if we need to actually differentiate data. Shooting a handful of different bullets and powders is “worthless” in almost every method it is applied by reloaders, but it’s popular - folks assume it works because they end up with a few small groups at the end, especially if they discount a few “called flyers” from the sample set. In general, those reloaders would be better off - firing less rounds, spending less time, money, and energy - if they just picked ONE combination and made it work, and would end up with just as small of groups, or even smaller for having done actually relevant load development.

So it’s really not so much a problem that reloaders aren’t doing statistically valid work - the problem is that reloaders assume they ARE doing valid work, and making misguided inferences, misguided differentiations, based on observations which don’t actually reflect differentiated data.

 

[LiveLife]

Isnt the op just commenting on a video ?

That's what I thought on post #1

5 shot groups "kinda worthless"?

say, one brand of ammo is consistantly tight-grouped after 5, and another ammo is consistantly all over the place after 5 then duh, obvious...but if different ammos are sorta kinda similar, then 1000's of hits need to be averaged to determine grouping.

But when member Riomouse911 replied preference for 10 shot groups over 5 shot groups, OP specified on post #3 to shooting 22LR comparing 20 brands to determine which was better but having trouble with majority with groups that weren't obvious -

i spent countless hours of comparing 20-some brands of .22 in a marathon weekend, for each of three different guns (10 shot groups and i walked miles setting up new targets for each 75yd test.)

Two or three brands were "best" and duplicate-ably better by a modest margin...and two or three brands sucked. But the VAST majority were simply average and tough to distinguish.

Then when 5 pages later, I noticed OP had not responded back at all, I mentioned on post #122 that OP is trying to determine better shooting 22LR out of 20 or so brands - https://www.thehighroad.org/index.p...s-kinda-worthless.920275/page-5#post-12658780

... post #3 is the last post from OP and it is specific to 22LR tested at 75 yards.

 

[Chief TC]

Really, you need a 50 shot group to produce any real conclusions. And you need to repeat this about 100 times. You need more and more data. Seriously, your rifle will tell you what it can do with what you are feeding with minimal shots. May take 1, may take 3, and so on. My hunting rifle only needs to be fired once for me to know what it can do at each distance. My trapdoor needs 10 shots so I can see what it can realistically do at each distance. My experience has been, the better the rifle the less shots you need to determine you and your rifle's capability. If you are not a regular shooter with consistent techniques, then determining groups becomes much more difficult because now you don't know if it is you or the rifle. Note that I am discussing general shooting and reloading. I'm not talking about competitive shooting. Practice and testing should be quite extensive.

 

[mcb]

Allot of this comes down to how much fidelity are you after. If I have two ammo batches that when fired in one particular gun that will, with the first batch, produce a mean radius of .25 MOA on a arbitrarily large number of rounds fired and a second batch that will produce a mean radius of 2.0 MOA on a similarly arbitrarily larger group I can probably make a meaningful distinction between the two batches of ammo in a single low round count group (3-5 rounds) though to say definitively how different they are with precision is going to require a few more groups. On the other hand if batch one produces mean radius groups of .25 MOA on an arbitrarily larger group and batch two produce .26 MOA. I am going to need a whole bunch shooting before I can tell you which is which with any meaningful confidence.

IE you load two magazines for my 10/22 one with 5 rounds of Thunderbolts and one with 5 rds of Lapua Center-X and let me shoot them I could probably correctly tell 99 times out of one-hundred which magazine held which ammo only by looking the resulting two 5-round groups. I bet I could do nearly as good with only 3-shot groups given the QC difference between the two ammos. On the other hand I shot ten 5 rd groups with both SK Standard+ and SK Flatnose (FLGS actually comped me a box of Flatnose when I bought a brick of Std+ just to try) and could not statically tell you one shot better than the other. The mean average radius was practically identical and the difference was well into the noise due the measurement capability (Range Buddy is so handy). I stayed with the Standard+ because I already had a brick, it was slightly cheaper, and in the ten group testing I shot the smallest single extreme spread group with it. I have no doubt the Flatnose would have performed the same in the subsequently matches I shot after that testing.

-rambling

 

[Cityevader]

Sorry if i seemed absent in my own thread, but feel i contributed when appropriate.

The following is purposely meant to be an opposite extreme example...
I believe most every person here would believe it to be BAD ADVICE to tell somebody to take one shot and adjust the scope accordingly; then take a second shot and make a second adjustment; on and on and on.
Ridiculous, right?

Variables of all sorts...
Stack up of tolerances...
Statistical probabilities...
all make a small sample size "worthless".

Thank you @LiveLife for "picking up what i was putting down"...for "smelling what i was stepping in"... for "know what i mean, jellybean".

 

[LiveLife]

load two magazines for my 10/22 one with 5 rounds of Thunderbolts and one with 5 rds of Lapua Center-X and let me shoot them I could probably correctly tell 99 times out of one-hundred which magazine held which ammo only by looking the resulting two 5-round groups.

While that is true, that's not what the OP is asking

spent countless hours of comparing 20-some brands ... Two or three brands were "best" and duplicate-ably better by a modest margin...and two or three brands sucked. But the VAST majority were simply average and tough to distinguish.

During testing, OP already identified obviously accurate brands from less accurate brands. But the "VAST majority" were harder to "distinguish".

What the OP is asking is how do we tell the difference which brand is more accurate when average group size is less "distinguishable"

On post #122, I mentioned one method I used during the 10,000+ round testing of 25+ brands/weights/lots of 22LR and that is using known reference ammunition (Aguila 40 gr CPRN that consistently produced 3/4"-1" at 50 yards) to verify changing group size (CCI SV produced smaller 1/2"-3/4" but Aguila was much cheaper as I paid less than half price for Aguila).

If group size got larger and I wanted to rule out effects of cold/hot barrel, scope ring/mount bolts getting loose, shooter fatigue, etc.; I would shoot a verifying group. If verifying reference group was 3/4"-1", I would deem larger group size to ammunition. If verifying reference group was larger than 1", then I would deem larger group size to rifle/shooter and check bolts for looseness and/or take a break.

It got to where I just started all of 8-12 5/10 shot group targets with a reference group and ended with a reference group to compare group size of different brands. If reference Aguila groups remained the same at 3/4"-1", I would deem group sizes of different brands true.

three different guns ... 75yd test

And since OP used "different guns", I also illustrated that 10/22 and T/CR22 groups produced different size groups using same brand ammunition and this needs to be considered as well.

Since I saw effects of bullet drop as vertical elongation during my 75/100 yard testing, I also mentioned that OP consider testing at 50 yards as bullet drop from slower muzzle velocities could be viewed as "flyer".

 

[DMW1116]

After reading this thread a couple time and particularly paying attention to @Varminterror ’s posts, I took two of my best shooting loads and went to the range to really see what was going on.

Load 1 is a modest charge of CFE 223 pushing a 77 grain SMK. This load was tested in my 20” PSA Premium FN upper using iron sights and pretty reliably shot under 1.5 MOA when I was testing it.

Load 2 was a modest charge of Shooters World Tactical Rifle pushing a 69 grain RMR BTHPCT ( boat tail hollow point cob tipped) bullet. This load hovered around 1” during testing in two different rifles. Testing for both loads used 4 shot groups.

I swapped my scope back to my 20” rifle and did a one-shot zero as a fouling shot with the RMR load. Then I took 9 shots with each load on the same target at 100 yards. After that I swapped the scope back to my 16” rifle and shot a 9-shot group with the RMR bullet. I would have shot the SMK load too but the range got crowded and I had to leave anyway. I took about 15 to 30 seconds between shots while making the groups.

The RMR group with each rifle was 2.1-2.2 inches. The 16” rifle group was slightly larger. The group with the SMKs was about 2.4 inches. Now all these are bigger than during testing by about double. I expected them to grow but not this much.

I have 200 52-grain Barnes Match Burners and an array of powders to use. How would I set up my tests to make sure I don’t fool myself?

 

[Pudge]

Back in the olden days, they would mock up designs and test them in labs for things like intakes, headers, body effects and such. Now there are computer models that give much better data.

How hard would it be to take some small sample data, and feed it into a computer that could provide a statistically relevant model for an individual rifle and load?

 

[LiveLife]

How hard would it be to take some small sample data, and feed it into a computer that could provide a statistically relevant model for an individual rifle and load?

Not sure how that would apply to OP's quest for identifying more accurate 22LR ammunition

spent countless hours of comparing 20-some brands of .22 ... 75yd test ... Two or three brands were "best" and duplicate-ably better by a modest margin...and two or three brands sucked. But the VAST majority were simply average and tough to distinguish.

5 shot groups "kinda worthless"?

Load 1 ... shot under 1.5 MOA ... Load 2 ... around 1” ... used 4 shot groups.

20"/16" rifle ... I have 200 52-grain Barnes Match Burners and an array of powders to use. How would I set up my tests to make sure I don’t fool myself?

I think you still would have to do some powder work up to identify some sense of accuracy node for each barrel.

But I think this is off-topic in need of a new thread discussion to not hijack OP as you are asking for load development and OP is trying to identify more accurate 22LR ammunition when group sizes are comparable at 75 yards and difficult to identify which is more accurate.

 

[lionking]

I like 5 shot groups for testing it is a acceptable average in between 3 and 10 shots
My preferred rules for testing ammo and rifles are
3 shot group for expensive match ammo in a precision rifle
3 shot group for a hunting rifle
5 shot group for basic center fire ammo in a military surplus rifle or basic rifle
10 shot group for semi auto defense rifles like AR, AK , MINI 14, M1 carbine ect......

 

[Jim Watson]

I am backwards on that. When I was shooting F class, the match was 20 consecutive shots.
When I was shooting BPCR, the silhouette match was 5 or 10 critters, the target match was 10 on paper.
Would my rifle and ammo do well?

 

[Slamfire]

The trash cans at my Gun Club are filled with boxes of Eley Tennex and Lapua Midas plus and Center X.

We have a monthly precision 22lr match, and members are out testing their lots of ammunition, trying to find the lot that shoots the smallest group.

Take your inexpensive 22 lr ammunition and go shoot in a match. When you compare your targets against those of the guys with lot tested match ammunition, you will see a difference. Your groups will be larger. And, you will have flyers.

The trade off: you paid less to shoot lousy groups.

 

[Jim Watson]

My target shooting friend is frugal to the point that he considers SK Standard Plus to be acceptable. It will usually shoot with the high priced spread but he says he can tell from the report that he has found an odd one and expect a shot out of the group.

With Champions Choice out of SK, he is going to try Wolf Match Extra, now made by Eley.