Statistics- what does and doesnt matter.

[BradY]

A few people seem to agree or disagree, or even disagree to agree or agree to disagree, on this topic.

What sort of statistics really back the findings of our load testing and what ones do we just work out to make ourselves feel better?

Im talking ES and SD here. So who uses what? What purpose does it serve or what does it tell us?

My opinion is that SD isnt all that relevant. To have a small SD you are going to have a small ES anyway... is it just a figure that helps us show off how good our loads are? I mean to hold X ring at 1000yd f class or shoot tiny groups at 600yd or 1000ydBR you need to be shooting a load that is ballistically going to be travelling all similar speeds to reduce vertical on target. I sure as heck would like single figure ES rather than double figure ES and single figure SD...

Thoughts?

 

[jcampbellsmith]

Statistics do matter, but I don't see much use being made of them. ES is for the birds, all it tells you is the difference between the fastest and slowest shot, but nothing about the distribution of the other shots. Just because it is easy to measure doesn't mean that it's useful.

If you looked at 7 loads with the same ES, SD would tell you which was actually the load to pursue further. I've certainly used SD to drive my approach to brass prep.

Here's a couple of very good specific articles, but our sport is ripe for someone to write a good book on how to apply statistics across all that we do.

http://riflemansjournal.blogspot.co.uk/2010/07/statistics-for-rifle-shooters.html

http://www.the-long-family.com/group_size_analysis.htm

I watch folk using electronic targets, but taking no electronic copies of the target data and that's weird in my book. In an ideal world, I would chrono every shot I take, but I don't see any kit around to help me do that, especially when I'm in the field hunting.

I think making good use of statistics is a huge opportunity for us. I get the sense sometimes that just as folk aren't keen to pay for rifle coaching or training, they wouldn't be keen to pay for good statistical advice. Natural talent gets some folk an awfully long way, but to be really good at what we do, no stone should be left unturned.

Regards JCS

 

[CatShooter]

Do a search... I think this same topic was up a short weeks ago.

No one agreed, and there was a lot of "Faux science" thrown about.

There will be a lot of faux science thrown about in this one too.

 

[markr]

If.....it were as simple as using stats to develop loads and or loading tecniques, everyone would be using them to win. The biggest winner is the component manufacturers.

 

[GSPV]

Information is valuable. SD provides information. ES provides information.

 

[jcampbellsmith]

If.....it were as simple as using stats to develop loads and or loading tecniques, everyone would be using them to win. The biggest winner is the component manufacturers.

I don't agree. Most folk have no concept about how to calculate standard deviation or what it might tell them once they have calculated it.

The trial and error approach certainly favours component manufacturers and barrel manufacturers.

Regards JCS

 

[jcampbellsmith]

... ES provides information.

about two shots only, nothing about the intermediate velocities. Regards JCS

 

[GSPV]

Sometimes one is all that it takes to lose.

For example, H4198 in the 30 BR is known for dropping shots out the bottom on cool mornings. In that situation, you could have a truly great SD over a statistically relevant sample size and still lose to a load with worse SD but without that bad Min value.

So, I tend to look at Min and Max and where they occur in the string in addition to SD.

BTW, I was a decent statistician before I sold out and started doing computer engineering.

 

[WyleWD]

Do a search... I think this same topic was up a short weeks ago.

No one agreed, and there was a lot of "Faux science" thrown about.

There will be a lot of faux science thrown about in this one too.

+1 on the above post.

The statistics that you (might) rely on are only as accurate as the quality, and tenacity of the testing methodology. The quality and repeatability of the chronograph needs to be questioned as well as well as the accuracy of the individual bullet, primer, case, powder charge that you're testing. Then couple all that with wind drift, drafts, brain farts by the shooter, and subjectivity of the shooter.

There are more questions than anyone has answers for on the topic IMHO.
1) Is that 60+fps ES real or the result of cloud cover vs sunlight over the chrony, or just within the accuracy level of the chrony itself? How do you 100% KNOW that reading is correct? Really?? :
2) How do you know exactly that the bullet/load in the chamber is going to go 3248fps +/- 5fps AND land in the same hole as the last one? (Keep in mind there is the shooter variability here too)
3) What time of day is the testing taking place? Between 10am and 3pm or something else?
4)What is the exact sun angle, cloud density (or lack thereof)?
5)Distance of the chrony from the muzzle, and is the chrony over grass, blacktop, gravel, sand, or elevated?
6) How can I explain the two bullets that went into the same hole at 200yds with a 63fps speed difference according to the chrony last week? :-\

There are a host of other questions one needs to ask themselves, but to save wear and tear on the keyboard here.... one should be open minded, unbiased, use conditions as near to perfect and identical as possible, keep in mind that even the best equipment has an error range and that error range can be expanded by improper or inconsistent application and use. In the end, if the shot hits where it is supposed to, then you harvest the animal, shoot a tight group or what ever the gratification..... and if that itch is scratched the ES & SD are just a footnote. Both are tools, and are only as good as the conditions and shooter that uses them for information to make a decision. Are they gospel... no. The size of the group at the distance you're shooting is. JMHO. WD

 

[CatShooter]

Well said, Wyle...

For some reason, people think that chrono's are 110% accurate, and if it says you have a 3 SD and a 4 ES, then you believe it, even though, at 400 yards, you have a group that is 5 times as tall as it is wide.

A fellow on this site and I were shooting together a few years ago, and we put our chrono's in line, and ran rounds through them... you might expect that one might have a constant error because the crystal clocks were not exactly in tune... but the results we got were all over the place - one shot would have the first reading high, and the next shot, it would be reading lower. Even with clock error, the ES and SD should be the same... but Noooo. They also varied.

I no longer think of chrono's as being an accurate tool, but just an approximation of what to expect - and all tests MUST be repeated - the 5 shot load that gave you an SD of 6 and an ES of 12, might give you an SD fo 21 and an ES of 93 five minutes later... but both will make the same sized group at long range.

I have also had the same experience as you... a ES of 160 and a 1/2" group at 200 yards.

But the faux theories will pile up as the day passes.

 

[jcampbellsmith]

... put our chrono's in line, and ran rounds through them... you might expect that one might have a constant error because the crystal clocks were not exactly in tune... but the results we got were all over the place - one shot would have the first reading high, and the next shot, it would be reading lower. Even with clock error, the ES and SD should be the same... but Noooo. They also varied.

I no longer think of chrono's as being an accurate tool, but just an approximation of what to expect ...

Agree. I replicated this test recently and had the same kind of results. Thanks JCS

 

[XTR]

The equipment that nearly all of us use is neither accurate or precise enough to get really reliable info. For that reason I never look at what the Chrony says for ES/SD. I let the target tell me that. I use the chrony to get a good feel for where my average MV is running and what I can expect for come ups and to compare windage on different bullets. For example, if I know I'm getting MV of X with one bullet then I know I need to find an accurate node at MV Y± in order for it to be as good or better at 1000 yards. One other use I have found is identifying max pressures when MV stops increasing as the powder charge is increased (though primers falling out tends to confirm this finding :-[ )

I couldn't tell you what my beta chrony has given me for an SD/ES on any load in the last two yrs. Shooting at an F-Class target at 1000 yards will tell you if your ES is too high in just a few shots.

 

[BoydAllen]

Many shooters do not have nearby range facilities to test for vertical at long ranges. They have to use chronographs to prepare for matches that they go preloaded to, so I doubt that the use of chronographs is going to diminish. One thing that my old Oehler 33 has is variable screen spacing, out to something like 10'. It has been a long time since I opened up a case to replace the D cells so I could be off here. Anyway, my screens are the latest versions, and I would expect that the greater the screen separation, the more reliable the data. Do any of you have any information on this?

 

[BradY]

Ive tried testing through a magnetospeed and found the results to be much more consistant. We didnt get to fire many rounds through it but what we did see was consistency. Ive never seen that through the F1 than I own or a CED that a friend owns. Im more inclined to believe it than other chronys.

Saying that I have done my load testing at club shoots at 1000yds. Using a 3 shot replicated ladder test I can get an idea of what load ends up holding best waterline. To the point where I got one load running consistently within the X ring and often half that. Its shoots beautifully (shame about the person steering it in the wind). I will be running those rounds over a magnetospeed this weekend. End of the day I wont change the load as I know it shoots, no matter what ES and SD tell me. I believe the target more than the chrony.

Its good to see most people are on the same page about this. What a wonderful thing the internet is- we can learn so much from reading then discussing our thoughts and experiences.

 

[Lapua40X]

I'm talking ES and SD here. So who uses what? What purpose does it serve or what does it tell us?

If I allow target print ES to influence my decisions I run the risk of factoring in shooter errors (like inconsistent hold, etc.) so I just don't go there.
I much prefer MV SD data. Even when it's all over the place it's at least a fixed value if I use the median factor from a series of sample MV test results.
Taking the median from a series of samples to determine SD makes me feel good about my decisions. I'm no math guru so I have to admit I may not be focusing on the best approach but it's gud 'nuf fer me.
For those who don't have the high dollar chronographs, take your numbers home and open a spread sheet on your computer. Somewhere in the help section you'll find standard deviation formulas (something like =STDEV(cell:cell) that'll do the hard work for you.

 

[BoydAllen]

This is what I think happened. Back in the dim past of chronographs, some manufacturer (probably Oehler) decided to include the computation of SD in the functions that their then new model, probably as a convenience to those who were testing large lots of ammunition. Reloaders saw the information, and liked the technical sound of "standard deviation" and that it was a significantly smaller number (small good...big bad) than extreme spread, a term that lacked the mystery of SD because it is self explanatory, or at least easily figure out, if you subtract the velocity of the lowest shot from that of the highest. I should add that not one out of fifty chronograph users could give a proper definition of SD or has ever manually calculated one, but it became a fixture, that is commonly referred to no matter how small the sample. Its continued use is virtually guaranteed because it performs two important functions. It sounds really technical, and it tends to soften the truth of the extreme spread, which unfortunately, targets do not. So there you have it. The bald faced truth about SD in the world of chronographs.

 

[Nomad47]

Very well said, Boyd. I used to pay a lot of attention to SD, but Steve, Erik, and now, you, have convinced me that ES is much more important.

 

[BradY]

Boyd- you think the same way I have been thinking. SD does just sound like its there to either confuse or talk up the data. I still like to think simple- if you have a big ES then your going to have vertical on target. I remember a thread that German Salazaar posted referring to I think was 14 shots at 2850fps and one at 2900fps. Stating that the ES was crap but the SD was still good. Honestly if I saw 14 shots at 2850 and one at 2900 I would be totally discounting that one. A round not loaded as it should have been. Single digit spreads will be what I chase over the chrono but only ever as a side to working out a mean muzzle velocity and to back up what performs on target.

 

[stubbicatt]

Sometimes I read this stuff and fell like shouting, "you mean there isn't a Santa Claus?"

 

[jlow]

I would expect that the greater the screen separation, the more reliable the data. Do any of you have any information on this?

Boyd - yes, the greater the screen separation, the more reliable the data. The caveat is the rails are not bent and have been accurately placed in terms of position. The way you check is to check the difference between the two numbers, the lower the difference, the better. Mine are usually 10 or less.