ES What the heck?

[Dudedog]

https://www.ssusa.org/articles/2018/11/13/review-imr-target-powder/

The results of five 10-shot strings fired over a chronograph worked out like this: Average velocity: 728.6 fps, with a low of 750.6 and a high of 831.7. That doesn’t tell us much; more informative is the average extreme spread (ES) of 50.62 and average standard deviation (SD) of 16.48. The latter especially interests us because it is our best indicator of consistency―the lower the number, the more consistent the shot-to-shot performance, and that equates to precision (accuracy). ES and SD both indicate differences in bullet velocities, but while ES is the average difference between only the slowest and the fastest bullet fired over the chronograph, SD takes into account the variance of all the individual shot velocities in the string, so it’s more precise.

OK, am I nuts or he he confused on what ES is. (Maybe I am, and my chrono is as well)
ES-> MAX Vel - MIN Vel = extreme velocity spread in the string.
?? No average involved. How the heck is he coming up with average Extreme Spread
SD = (for a sample of the population)



Looks to me like the ES is 81.1 (831.7-750.6 but then the average can't be right either)
of course my math is failing me attempting to somehow figure out how he got an average of 728.6 with a low of 750.6 and a high of 831.7
(only way I can see this would work is a shot with negative velocity, came out the back of the gun ????) But that would still be a positive vel I guess just a negative direction. .
Ok maybe a low of 728 and a high of 831 giving an average of 750 but that would make the ES about 103.
To bad we don't have shot string numbers....
Of course maybe he was having a can't type day on the average.
I have those myself, more often than I like.


Anyway I was looking to see how people liked IMR Target.
Getting ready to order some powder and was thinking about another jug of something I haven't used before just for fun.

I would appreciate any info from people who have tried it.

 

[fxvr5]

https://www.ssusa.org/articles/2018/11/13/review-imr-target-powder/




OK, am I nuts or he he confused on what ES is. (Maybe I am, and my chrono is as well)
ES-> MAX Vel - MIN Vel = extreme velocity spread in the string.
?? No average involved. How the heck is he coming up with average Extreme Spread
SD = (for a sample of the population)
View attachment 839744


Looks to me like the ES is 81.1 (831.7-750.6 but then the average can't be right either)
of course my math is failing me attempting to somehow figure out how he got an average of 728.6 with a low of 750.6 and a high of 831.7
(only way I can see this would work is a shot with negative velocity, came out the back of the gun ????) But that would still be a positive vel I guess just a negative direction. .
Ok maybe a low of 728 and a high of 831 giving an average of 750 but that would make the ES about 103.
To bad we don't have shot string numbers....
Of course maybe he was having a can't type day on the average.
I have those myself, more often than I like.


Anyway I was looking to see how people liked IMR Target.
Getting ready to order some powder and was thinking about another jug of something I haven't used before just for fun.

I would appreciate any info from people who have tried it.

Read it again.

He says he shot five 10-shot strings. My guess is the 750.6 and 831.7 are apparently the low and hi from all the shots (50). Then he notes the AVERAGE ES of 50.62. The use of the word AVERAGE suggests that he used more than one ES for his calculations - probably five ESs from his five strings of 10 shots. Also note the AVERAGE SD, and he notes in the next paragraph that one SD was 11.04, so it was likely one of the five SDs he used to get the average of 16.48.

Make sense?

 

[Rule3]

To early for this. I was told their would be no math.

I agree that it probably the AVERAGE ES but without all the numbers just a WAG

 

[lordpaxman]

Make sense?

Your interpretation of the article makes sense. The article does not. I would recommend everyone read "How to lie with statistics", it's enlightening.
10 shot strings with a 6 shot revolver, hmmmm, they're all the same load so there's one ES number, one average, and one SD with a sample size of 50. No averages of averages allowed!

of course my math is failing me attempting to somehow figure out how he got an average of 728.6 with a low of 750.6 and a high of 831.7

We learned from the same book. IF Vmin=750.6, you can't have an average less than that.

Anyway I was looking to see how people liked IMR Target.

I've done some workups for 9, 38 and 45. I was trying to find a good alternative to Titegroup, and PV had IMR Red and Target in stock, sadly. I disagree with Art Merrill on the metering aspects of Target. In my LNL powder measure, the consistency of drops was poor until I modified the measure to better scrape the cavity. Target also tends to be "clumpy", at least through the winter months and it may have been a static related but even dumping from the 4# jug it seemed to clump together. Felt recoil was similar to Titegroup, but it burned cooler, cases and barrel were not as hot as Titegroup. Smoke and cleanliness were ok from my notes - meaning acceptable to me. I'm glad I started at the low end of the data as I could make minor or major, depending on the caliber, quite easily. You uber short COL people take heed. In 38 I did not do position firing, it was just load and shoot, and even then the SD numbers were in the high teens to low 20s - something that does agree with the article in question! I do plan on shooting more of Target, but SP kind of stole the show.

 

[higgite]

The average of high and low isn’t the same as the average of all 50 shots, nor even the average of the 5 string averages. I’m also not sure why he would do average ES and SD over 5 x 10 shot strings instead of the true ES and SD over the entire 50 shots.

 

[Walkalong]

. I’m also not sure why he would do average ES and SD over 5 x 10 shot strings instead of the true ES and SD over the entire 50 shots.

Correct.

If you fire 50 rounds and the highest velocity is 720 and the lowest velocity is 660, your ES is 60.

If you shoot another 50 rounds and get a high/low of 700 & 670, the ES is 30.

The ES for all 100 shots is still 60, not 60 + 30 = 90/2 =45

 

[Walkalong]

Or maybe the statisticians will tell me I'm wrong.

 

[Dudedog]

The average of high and low isn’t the same as the average of all 50 shots, nor even the average of the 5 string averages. I’m also not sure why he would do average ES and SD over 5 x 10 shot strings instead of the true ES and SD over the entire 50 shots.

Ah, OK I was sleepy last night, that makes sense on how he came up with his numbers, even if the reason he did it that way doesn't. We are going to average averages
My brain sort of locked up attempting to figure out how the average was lower than the low....
It never occurred to me to take 5 strings of the same load and average the low, high ES and SD.

hmmmm, they're all the same load so there's one ES number, one average, and one SD with a sample size of 50. No averages of averages allowed!

That's what I was thinking.
5 tests but one sample, 1 low, 1 high, 1 average, 1 ES 1 SD.

I would recommend everyone read "How to lie with statistics", it's enlightening.

Might be a fun read, something along this line maybe...
Hmmm so instead of one 10 shot sample I could do 3 3 shot strings, that way if I had a high ES on one the lower ESs would give me a better average ES so my numbers look better.
but wait a minute better still if I did 10 one shot strings my ES and SD for each string would be 0 and my average low would = my average high!
Great meaningless numbers


Anyway looking at Target being flakes I would guess metering would not be as good as small sticks/ball powder.
Thinking I will pick something else new to try for fun, happy with WSF and SP for 9mm but was just thinking of trying something new.
There is an adventure in every bottle/jug of powder

PS: Sorry if my OP didn't sound quite High Road, was just scratching my head about the numbers.

 

[higgite]

My brain sort of locked up attempting to figure out how the average was lower than the low....

It can't be. You're not the one that had brain lock. The author of the article apparently wasn't a math major in school. Or isn't a good proofreader. Or both.

 

[lordpaxman]

The author of the article apparently wasn't a math major in school.

Perhaps they’re using that new fangled Common Core Math?

 

[fxvr5]

Clearly the average (728.6) is wrong. I suspect a typo. The editor didn't catch it either.

I'm also not sure why he broke the analysis into 5 separate 10-shot strings with the same ammo instead of lumping them into one 50-shot group unless he had a particular point he was trying to make, but he should have explained that so readers would understand why.

 

[fxvr5]

I would recommend everyone read "How to lie with statistics", it's enlightening.

I don't think he's trying to lie. But the reason for his method of analysis is not clear.

 

[Rule3]

It's only one persons evaluation without all the data. I would not loose any sleep over it.

 

[Varminterror]

The answer to the OP question has been revealed above - the article talked about averages of independent data sets, not discrete stats for a single data set.

I’m also not sure why he would do average ES and SD over 5 x 10 shot strings instead of the true ES and SD over the entire 50 shots.

Never for a single shooting session, but I often consider average ES and SD for various strings, for example, fired over different days, or with different lots of powder or bullets, or with different firing count on my brass. I do that to give me an idea of how significant the batch-to-batch variation might be, without examining a huge data set. Not terribly useful most of the time, but do that particular math sometimes.

 

[Hondo 60]

In my book of knowledge (which is extremely short lol) ...

ES is extreme spread.
It's calculated by taking the fastest reading & subtracting the lowest.
If my fastest reading is 895, & the slowest is 795 then ...
My ES or extreme spread is 100.
(.38 Special)

895-795= 100.

Now the SD or standard deviation, that is too complicated for me.
I just let my chronograph figure that one.
And I don't care for a math lesson please.

 

[Walkalong]

but I often consider average ES and SD for various strings, for example, fired over different days, or with different lots of powder or bullets, or with different firing count on my brass.

Yes.

 

[jmorris]

I would recommend everyone read "How to lie with statistics"

Not to mention 88.63% of all statistics are made up on the spot.

 

[Dudedog]

but I often consider average ES and SD for various strings, for example, fired over different days, or with different lots of powder or bullets, or with different firing count on my brass

That makes more sense, different lots of powder/primer/bullets, different weather, etc. (for that matter maybe a different pistol since we are talking about ES and SD not the MV itself)
I can see where working up number for say all my 9mm loads with different lots of WSF, S+B SPs, RMR 124 MPRs in my 9mms might give a useful number to say hey this combo in all my guns
(no more S+B SPs BTW , I liked those at $20/k on sale)
is this consistent(or not if that is the case, but usually on the nots we move on to a different combo) so an average ES of X and and average SD of Y for this load over time.
I still think I would be more interested in the worst case #s
Hmmm if we average them then maybe a SD of ESs and SDs as well.

But lets not forget the important thing no matter what the numbers show, Does it shoot well on paper.
(this ignores the can I shoot it well part, huge ES there, SD isn't very good either)

I didn't see a reason for him breaking it in to 5 stings of 10 when it sounded like one batch of everything on the same day, I'm sure he had one.

 

[denton]

I am constantly amazed at the pseudo statistical inventions that shooters come up with....

 

[Varminterror]

You can poke fingers at the validity and confidence interval of a small dataset all you want, but “pseudo statistical inventions” is a long reach. Average, ES (Range or Spread depending upon whose book you’re reading), and SD are about the most basic statistical functions possible. While it’s true, a lot of guys take far too few shots to establish a valid dataset, it’s pretty difficult to mess up a Range or Standard Deviation function.

When handloaders start talking about revolutionary new stochastic algorithms for load optimization, then I’ll start criticizing the math.

 

[fxvr5]

Looking at statistical parameters can be useful, however:

It's important to keep in mind that most of the velocity parameters can vary even when comparing the same ammo shot at the same time, which is what the author in the article did. He doesn't give the range of the SDs or ESs, but it's not unusual to see a high and low SD (or ES) differ by a factor of 2 or more. That's just the nature of this sport, where velocity can vary considerably from one shot to another even with carful weighing of powder charge. That's why there are 'match' primers which, when their mixture weight is more uniform, the ES tends to be reduced - indicating that primers are part of the problem, too.

Having more shots in the sample size can reduce the variation, but it can still be higher that one might expect.

It's good to keep in mind that a SD or ES that is twice the size with ammo A than it is with ammo B might be meaningless because one can find that much variation in the measurements when comparing two separate strings from just ammo A.

 

[Walkalong]

It's good to keep in mind that a SD or ES that is twice the size with ammo A than it is with ammo B might be meaningless because one can find that much variation in the measurements when comparing two separate strings from just ammo A.

I have seen this when chronoing loads. Shoot 6, then 6 more, then 6 more, all of the same load, getting the Hi/Lo/Avg/ES/SD each time, and sometimes they are a good bit apart. I take all small sample chrono results with a certain grain of salt.

It is, however, encouraging when a load always shows similar numbers in small samples.

 

[ATLDave]

When someone gives an average of averages, the overwhelmingly likelihood is that they don’t have a good grasp of statistics.

 

[Varminterror]

Presenting the average of averages can easily make sense if the data sets compiled in the comparison set are independent but related.

Let’s say I want to know if Hornady brass is as good as Lapua brass. So I buy .30-06, 284 win, 7-08, and 6 creed brass.

I shoot 2 or 3 different brands of bullets from each cartridge in each set, over a couple different powders. Now I have hundreds of not thousands of shots in this multifactorial matrix.

I can simplify presentation of the data set by presenting average SD and average ES’s. If a guy only compared the singular ES for one cartridge loaded in the 2 types of brass, with one bullet, the data set would not be very conclusive. But presenting a comparison of the two matrices of different bullets and powder combinations in each cartridge would be exhaustive. Instead, a comparison of average ES’s and average SD’s might be very meaningful. Even then, taking the ES and SD of the ES’s, and the ES and SD of the SD’s might be very meaningful - in a very concise and palatable results set.

 

[vaalpens]

He says he shot five 10-shot strings. My guess is the 750.6 and 831.7 are apparently the low and hi from all the shots (50). Then he notes the AVERAGE ES of 50.62. The use of the word AVERAGE suggests that he used more than one ES for his calculations - probably five ESs from his five strings of 10 shots. Also note the AVERAGE SD, and he notes in the next paragraph that one SD was 11.04, so it was likely one of the five SDs he used to get the average of 16.48.

++1

Sometimes we read what we think should be there instead of what is actually written. I sometimes have a difficulty finding errors in my own requirements/technical documents when I proofread them.

It would be interesting to find out how IMR Target compares with Bullseye and 700-X, especially when it comes to 45acp and 38spl.