How do you evaluate precision - eTargets

[Keith Glasscock]

Having seen several discussions, and participated in many of them regarding eTargets, I noticed that none of us seem to be on the same page on what to measure in regards to their accuracy.

The NRA rules require:
• 10.17.1 Electronic Target Requirements- Any electronic target system
that meets the requirements of these rules may be used for NRA Sanctioned
matches.
(a) Electronic scoring targets must be capable of scoring shots to within
.25 inches of accuracy.<snip>

What I've seen:

There are some that want to use the "average" error - I have to assume that means they want to use the mean (statistical average) error. If the error is truly random, the mean will be zero once the acoustic center is found.

There are others that want to use the Standard Deviation (SD). That has resulted in quite the discussion about normal distributions in another thread, and doesn't bear repeating here.

Yet others want to use the extreme spread. By using the Range (statistical term again), they are wanting to capture every error.

While that seems plenty to discuss, there is more:

How do we determine the acoustic center of a target in order to have a place to measure from?

Some have been using the average error of all shots fired in a string to determine the acoustic center. This is great, except it tends to hide the shifting of the acoustic center during a string.

I have suggested looking at it from the perspective of a shooter that fires 2 sighters. In my case, I suggest using the average error of those two shots to determine the acoustic center for the entire string.

OK, I turn it over to you. I do have some of the previously presented data at hand so I can show the effects of each method, but want to hear from the community at large before torturing the numbers with statistics.

 

[dcali]

This is one of the bigger problems with the fairness and randomness of E targets. You need an engineering degree to figure it out, and codifying the necessary detail into the rulebooks is going to be unintelligible to laymen, and prone to gross misinterpretation. "Within .25 inches of accuracy" is simple, but meaningless, and any interpretation requires a healthy dose of assumption. At the very (very!) least you'd want something closer to "mean error of less than x, with a standard deviation of less than y". Even if you could agree on a standard that was logically sound, how would you check and verify targets without making a spec that would make the ASTM blush? Who's going to do that work, pay for it, and what do you do if a target fails inspection?

Until they are so good that everyone just laughs and says "shut up, they're fine.", this is a problem. But they'll never get that good if nobody uses them (which is fine by me).

In practice, the targets are what they are, and they're not throw away cheap, so people will just use them as is once they are installed. Errors will vary between location and manufacturer. I don't see a practical way around that.

 

[Bob3700]

We shoot on a Silver Mountain target system. To check the accuracy of the etarget, we simply shoot our 20 shots on a new target face and then take a screen capture of the etarget. Now go down to the target and compare the actual bullet holes with the screen capture. That will pretty well tell you if your etarget is calibrated properly.
Now to the question of does it record your shots within .25". Well I can tell you when we are preparing for a match, we use a live target with a live puller. No question about your zeros then.

Bob

 

[Bob3700]

We shoot on a Silver Mountain target system. To check the accuracy of the etarget, we simply shoot our 20 shots on a new target face and then take a screen capture of the etarget. Now go down to the target and compare the actual bullet holes with the screen capture. That will pretty well tell you if your etarget is calibrated properly.
Now to the question of does it record your shots within .25". Well I can tell you when we are preparing for a match, we use a live target with a live puller. No question about your zeros then.

At a match with etargets. My question is: Have all the targets calibration been verified as equal? Is one target accurate to .125 while the target next to it is calibrated to .25. I would venture that no one actually knows and it is "ASSUMED" that all targets are calibrated the same. I am betting that they are not.

Bob

 

[Dos XX]

If you calibrate a target with the calibrating shooter lined up dead center on target, would it change anything if the next shooter was lined up left or right of where the calibrating shooter was lined up? I have seen discussions about angle of entry of the bullet.

 

[MattPeetz]

The interesting thing is that on all these tests on e targets the accuracy has been greater than the .25” requirement but no one seems to be able to grasp that.

 

[dcali]

The interesting thing is that on all these tests on e targets the accuracy has been greater than the .25” requirement but no one seems to be able to grasp that.

I'd believe that, but first someone has to tell me what the .25" requirement actually means. As written, it means whatever the reader wants it to.

 

[sdean]

Having seen several discussions, and participated in many of them regarding eTargets, I noticed that none of us seem to be on the same page on what to measure in regards to their accuracy.

The NRA rules require:
• 10.17.1 Electronic Target Requirements- Any electronic target system
that meets the requirements of these rules may be used for NRA Sanctioned
matches.
(a) Electronic scoring targets must be capable of scoring shots to within
.25 inches of accuracy.<snip>

What I've seen:

There are some that want to use the "average" error - I have to assume that means they want to use the mean (statistical average) error. If the error is truly random, the mean will be zero once the acoustic center is found.

There are others that want to use the Standard Deviation (SD). That has resulted in quite the discussion about normal distributions in another thread, and doesn't bear repeating here.

Yet others want to use the extreme spread. By using the Range (statistical term again), they are wanting to capture every error.

While that seems plenty to discuss, there is more:

How do we determine the acoustic center of a target in order to have a place to measure from?

Some have been using the average error of all shots fired in a string to determine the acoustic center. This is great, except it tends to hide the shifting of the acoustic center during a string.

I have suggested looking at it from the perspective of a shooter that fires 2 sighters. In my case, I suggest using the average error of those two shots to determine the acoustic center for the entire string.

OK, I turn it over to you. I do have some of the previously presented data at hand so I can show the effects of each method, but want to hear from the community at large before torturing the numbers with statistics.

Shouldn't you be talking to the NRA about this? I really don't see what anyone thinks on this forum has anything to do with that. What do you expect out of this? For every place that has invested in E targets to trash them. Can't you find enough pull target places to shoot until they get good enough for you? I just don't understand the same arguments back and forth from the 10 or so same people. I shoot at Bar3 on Hexta targets and happy with them.

 

[dcali]

Have all the targets calibration been verified as equal? Is one target accurate to .125 while the target next to it is calibrated to .25. I would venture that no one actually knows and it is "ASSUMED" that all targets are calibrated the same. I am betting that they are not.

To be fair, paper targets aren't all the same size either. Try measuring them.

 

[243Mendoza]

Having seen several discussions, and participated in many of them regarding eTargets, I noticed that none of us seem to be on the same page on what to measure in regards to their accuracy.

The NRA rules require:
• 10.17.1 Electronic Target Requirements- Any electronic target system
that meets the requirements of these rules may be used for NRA Sanctioned
matches.
(a) Electronic scoring targets must be capable of scoring shots to within
.25 inches of accuracy.<snip>

What I've seen:

There are some that want to use the "average" error - I have to assume that means they want to use the mean (statistical average) error. If the error is truly random, the mean will be zero once the acoustic center is found.

There are others that want to use the Standard Deviation (SD). That has resulted in quite the discussion about normal distributions in another thread, and doesn't bear repeating here.

Yet others want to use the extreme spread. By using the Range (statistical term again), they are wanting to capture every error.

While that seems plenty to discuss, there is more:

How do we determine the acoustic center of a target in order to have a place to measure from?

Some have been using the average error of all shots fired in a string to determine the acoustic center. This is great, except it tends to hide the shifting of the acoustic center during a string.

I have suggested looking at it from the perspective of a shooter that fires 2 sighters. In my case, I suggest using the average error of those two shots to determine the acoustic center for the entire string.

OK, I turn it over to you. I do have some of the previously presented data at hand so I can show the effects of each method, but want to hear from the community at large before torturing the numbers with statistics.

The average error is not zero even when the acoustical center is the paper target center. The average x and y error may be zero after proper calibration but the average radius error (aka linear error) is not zero since this error is random and positive in value. A proper criterion for error is the mean error plus some multiple of the standard deviation.

Why should the acoustic center move during a string unless the system is failing.

 

[MattPeetz]

I'd believe that, but first someone has to tell me what the .25" requirement actually means. As written, it means whatever the reader wants it to.

Why make it so difficult. .25” in all directions from actual bullet center.
.25” is way to big

 

[dcali]

Why make it so difficult. .25” in all directions from actual bullet center.
.25” is way to big

That's not what it says, but that would be the simple way to go about it (and probably what they meant and what it should say). But even then, it allows for some significant wiggle room. Is it acceptable if most of those shots are on one side of the hole (that is, that there is a consistent shift in one direction or another) ? What if the holes are shifted .25" for one target and not at all for another? It's easy to go down a rat hole of eggheadedness on this.

And I agree. 1/4" is too big no matter how you define it, especially at 300, 500, and 600.

All of these issues go away once targets get good enough. At some point you don't even need a rule, just like there's no rule for paper - it's not like they say "9 out of 10 people must be able discern a bullet hole within .010" from the line" or something silly like that. Unfortunately, confidence in the e targets isn't at that level yet.

The more practical thing here is that the rule won't matter. Nobody is going to check the targets meaningfully, and if they do, they won't agree on what "capable of scoring shots to within.25 inches of accuracy" actually means. I just don't think this rule has any practical importance. It's not like any clubs are going to throw away their targets if they don't measure up completely.

 

[MattPeetz]

That's not what it says, but that would be the simple way to go about it (and probably what they meant and what it should say). But even then, it allows for some significant wiggle room. Is it acceptable if most of those shots are on one side of the hole (that is, that there is a consistent shift in one direction or another) ? What if the holes are shifted .25" for one target and not at all for another? It's easy to go down a rat hole of eggheadedness on this.

And I agree. 1/4" is too big no matter how you define it, especially at 300, 500, and 600.

All of these issues go away once targets get good enough. At some point you don't even need a rule, just like there's no rule for paper - it's not like they say "9 out of 10 people must be able discern a bullet hole within .010" from the line" or something silly like that. Unfortunately, confidence in the e targets isn't at that level yet.

The more practical thing here is that the rule won't matter. Nobody is going to check the targets meaningfully, and if they do, they won't agree on what "capable of scoring shots to within.25 inches of accuracy" actually means. I just don't think this rule has any practical importance. It's not like any clubs are going to throw away their targets if they don't measure up completely.

Use a little common sense.
Even if it was a consistent .25” away from the line of the 10 ring, its still unacceptable.
We are talking accuracy here. The whole key component to ease shooters minds. It doesn’t matter if 9 shots are .25” away from actual bullet center and 1 is spot on. They ALL need to consistent.
You have a .25” circle from actual bullet center.
The current ET’s have failed at that. At least the SM’s have.

 

[Keith Glasscock]

The average error is not zero even when the acoustical center is the paper target center. The average x and y error may be zero after proper calibration but the average radius error (aka linear error) is not zero since this error is random and positive in value. A proper criterion for error is the mean error plus some multiple of the standard deviation.

Why should the acoustic center move during a string unless the system is failing.

I see what you are saying. Makes perfect sense statistically. What confidence level do you think would be appropriate for the standard deviation addition?

On an open system, the acoustic center moves with the wind across the target face. Unfortunately, testing data where the noise of the random error doesn't overwhelm the shift is hard to come by.

 

[JohnKielly]

I guess the concerns I have with ET use is:

• Whether a given range range has done due diligence on the chosen equipment before making the selection
• Whether they have a suitable maintenance regeime to ensure continuity of the the original acceptable level of performance
• Whether they have developed processes to ensure that physical & acoustic centres coincide

 

[Keith Glasscock]

Use a little common sense.
Even if it was a consistent .25” away from the line of the 10 ring, its still unacceptable.
We are talking accuracy here. The whole key component to ease shooters minds. It doesn’t matter if 9 shots are .25” away from actual bullet center and 1 is spot on. They ALL need to consistent.
You have a .25” circle from actual bullet center.
The current ET’s have failed at that. At least the SM’s have.

I find it interesting that some targets have proven errors in excess of my group sizes at 600 yards.

 

[243Mendoza]

I see what you are saying. Makes perfect sense statistically. What confidence level do you think would be appropriate for the standard deviation addition?

On an open system, the acoustic center moves with the wind across the target face. Unfortunately, testing data where the noise of the random error doesn't overwhelm the shift is hard to come by.

I think the standard 95% confidence would be attainable with today's systems but 99% would be nice.

The new 8-mic system should fixed errors due to wind.

 

[243Mendoza]

Use a little common sense.
Even if it was a consistent .25” away from the line of the 10 ring, its still unacceptable.
We are talking accuracy here. The whole key component to ease shooters minds. It doesn’t matter if 9 shots are .25” away from actual bullet center and 1 is spot on. They ALL need to consistent.
You have a .25” circle from actual bullet center.
The current ET’s have failed at that. At least the SM’s have.

You cannot have consistency with random error. You can only ask for a specific statistical performance like any other product quality control.

 

[Bindi2]

The .250 error is a red herring. The repeatable .250 is the fact. Example the shot hole is fact, the recorded position is in the same position next to each hole .250 away maximum . Testing done here is the error is way less than .250 and repeatable at the same distance. One shot on one target at 800m showed an error of 8mm but in the same place as others just further out. 8mm is just larger than a 30cal projectile. I own and use a SMT. I shoot competitions on Kongsberg which I also maintain for my club. The sound chamber condition is critical to good operation. The targets can be shot out at between 300 rounds minimum to 1200 rounds maximum. So a means of in shoot maintenance is required that is fast and keeps accuracy at the premium.
Random error is showing the target is shot out no ifs no buts shot out.
The E target saga has also shown how inaccurate paper targets are. The movement in the frames either up or down forwards or backwards affects the accuracy of e targets, this was never seen or worried about before e targets. Another issue is the squareness of the range and the targets to the mounds or the vertical aspect of the frame to the firing line.

 

[CharlieNC]

Using either the average error or the SD of the error is not adequate, both must be specified for completeness. The same applies to characterizing shots on paper..Also just because the average error might be 0 at the center, that might not assure the same if shots are not centered if the overall calibration is not homogenous for all the audio devices.