[AUTIGER04]

What do these terms mean?

Extreme Spread and Std Deviation?

[Stoker]

Thery are both terms from statistics and both are used to indicate the extent of dispersion in a group of values. So, in terms of shooting, they could be used to indicate how big a group was.

Extreme spread would be the distance between the two most widely separated holes on the target.

This is a pretty crude measure since it uses only two pieces of information. Standard deviation is a much more sophisticated measure of dispersion, since it makes use of all the available data.

It assumes a normal distribution - the well-known 'bell curve' - and with such a distribution we can say with confidence that 95% of the items will lie within slightly less than two standard deviations on either side of the centre.

So, going back to shooting a group, the smaller the standard deviation, the tighter the group. If 95% of the holes fall within a 12 inch circle, you have a standard deviation of roughly 3 inches - a 12 inch circle amounting to 2 standard deviations on either side of the centre. Clearly, if you just looked at extreme dispersion, one flyer could give you quite the wrong impression about an otherwise tight group.

Other measures of dispersion are used - the 50% circular error probability measure was commonly used in assessing the accuracy of free-fall bombing, for instance.

[Thumper]

They don't let 'em tech Statistics at Auburn.

:evil:

GO DAWGS!

[Hal]

Aut,
Both are terms used a lot by handloaders to guage/measure the uniformity of the loads they make.
Short answer:
Think small numbers for either/both as being better.

Longer:
SD is a measure of uniformity. An arbitrary numerical value is assigned (usually by the inner working of a good chronograph) to an arbitrary number of shots fired through the "Screens". The end result is how uniform each shot compares to the "standard", and the deviation is how far off of that "standard" the gross string is.

Extreme spread:
Highest velocity minus lowest velocity for the string fired over the screens.

Little more:
If you read Mr. Camps excellent range reports, you'll note he clocks most of his loads, and lists the SD and Extreme spread for the different ammunition he fires. Since Mr. Camp is an excellent (very very very consistant) shooter, I like to play a little game with his reports to see how the results compare to what I believe they should be,,,before I look at the pictures of his targets.

It's interesting to note, that on paper ie: loads with small numbers, usually show the tightest groups when I scroll down to check the targets.

[Sean Smith]

An arbitrary numerical value is assigned (usually by the inner working of a good chronograph) to an arbitrary number of shots fired through the "Screens". The end result is how uniform each shot compares to the "standard", and the deviation is how far off of that "standard" the gross string is.

Well, it isn't arbitrary and it isn't assigned. It is calculated from the data (in this case, the chronographed velocities). Basically, it is the mean distance of each value from the mean of the values.

So this:

Avg = 1,250 ft/sec
ES (extreme spread) = 50
SD = 25

Would tell you that the difference between the highest and lowest velocities recorded was 50 ft/sec, and that the "average difference" of the individual measurements from the average of those measurements was 25 ft/sec... a darn inconsistent load, in other words.

The formula for calcuating the standard deviation is attached.

[Penforhire]

Curious that "extreme spread" is not simply called "range", the proper statistical term for what people have said it represents.

[Sean Smith]

Beats me, but I'm always seeing ES instead of range when talking about handgun chrono results. Maybe so it isn't confused with range (as in distance)?

[Hal]

Basically, it is the mean distance of each value from the mean of the values. That's what I was meaning to say :D
Thanks for the correction.

[Geezer]

I am a licensed CE, (means nothing), and as a profession, prepare construction cost estimates. Since we are talking about the future cost of something, I frequently have to make up numbers. The only data I have is from past events.

In order to accurately predict the future, I can't use all past events, only those that were "typical." Therefore, as a matter of course, in cost estimating ALL flyers are discarded from the calculation of the ED (range) and the SD.

Example: If we built five stores in various parts of the country, and store C was destroyed by a tornado during the framing stage, which meant we had to redo all of that work, the final cost of store C would be excluded from our calculations.

A bullet can be a flyer for a multitude of reasons, like hitting a bumblebee halfway down range. The reason is irrelevant, flyers are atypical (at least we hope so!!) so they should be excluded from any calculations which are intended to provide guidance for future actions.

On the other hand, if I want to know the probability of any particular shot from my weapon hitting the center of the bullseye, then I would include all the datsa, including the flyers.

So the actual working definition of either term, just like almost everything else in life, should be preceeded by.."It all depends..."

God bless and y'all be careful out there.:cool:

[Penforhire]

Understandable explanation of "flyers" Geezer, but in statistics we call those "outliers" and there are valid rules for dealing with them without the judgement call of just tossing them out.

[AUTIGER04]

Thumper, Im a jr high school drop out. :rolleyes:

ps
Im just a AU fan. But, Those AU can handle them "DAWGS" anyday of the week.:neener:

"AU Auburn University" not to be confused with UA Alabama University.;)

pss Thanks people.:)

[Phantom Warrior]

I hate stats, but have dealt with it. So, for the unsophisticated (like me), standard deviation is basically a measure of the spread (or range) that is designed mathematically to filter out extreme values or "outliers." I you shoot a 2 inch group and then miss the target on your last shot your extreme spread would still be 3 feet or something, but the standard deviation would be much closer to your group.
Specifically, 67% of the values (shots) are within 1 standard deviation of the mean value (the center of your target), 95% within 2 standard deviations, and I believe 99% are within three standard deviations.
So, if you shoot 100 rounds (all generally hitting the target, not many fliers)and your standard deviation is 1 inch, 67 rounds will be within 1 inch of the center (for a 2 inch circle or group), 95 will be within 2 inches of the center (for a 4 inch circle or group), and so on.
Hope this helps a little.

[Geezer]

Penforhire: Of course you are exactly right concerning statistical terms.

I used the term flyer in the sense we use it at the range, a round, that for some unknown reason, did not perform as expected or as could be predicted from the available data.

Construction cost estimating is not a science, it is formally defined by R.S. Means, the standard of construction cost values, as an art. We get to do things with numbers that tend to horrify mathematicians and particularly statisticians. Every quarter I have one or two in my classes, and they spend the quarter just sitting there shaking their heads at what I am putting on the blackboard.

For instance, if three out of four projects had unusual cost factors, say a flood, a strike, and arson, respectively, with only one of the four being constructed as planned. we would conside all three of the exceptional projects as flyers, and only use the data from the "normal" one. I appreciate that is not kosher in statistics, but hey, we're artists!

It is one of the few fields today where the latest knowledge (i.e. brand new college graduate) is essentially useless. Most estimators do not begin to really hit their stride until they have fifteen or twenty years experience, and the really good ones all have over thirty years.

Back to the range. As a hunter, what I care about is how accurate my weapon is when everything goes right. Flyers are irrelevant, group spread, (range), with the flyers excluded, is totally relevant. That allows me to determine the maximum range at which I can call my shot.

God bless and y'all be careful out there.:cool:

[mec]

The statisticians say that standard deviation is much more meaningful in a large sample than in the ten+round groups most of us fire over a chronograph for our average and extreme spread. Generally a small mean is also going to show a small standard deviation in small samples.

Still, our scientific calculators, some chronographs and most ballistics programs show the sigma factor and we have gotten used to it. Smith's discussion of calculated variation of a given round from the mean is a good one. Still, most statistics gives me a headache. If I present data based on 5 rounds i omit S and if I include S with the ten plus round strings I never fail to present extreme spread also.

On a practical level for example, if we see nine rounds with a small extreme spread and one outlier significantly higher or lower, further experimentation might reveal a patern suggestive of poor ignition.

[Guy L Johnson]

Stistical Anylisis for Quality Control
By Feganbaum

spelling is aproximate has been 25 years since colleage
Guy