? on ES / SD

[goodgrouper]

GG, Thanks for the explanation and I now understand how you can have a high extreme spread and still have a low SD.

You're welcome!

 

[Winchester 69]

Actually, for a large number of shots with uniform ammunition, the ES is described by 6 SD's. Said another way, the SD is 17% of the ES. This is an application of Statistical Process Control.
.

 

[goodgrouper]

Win 69,

How 'bout running a standard deviation on my ten shot string mentioned earlier?

 

[Winchester 69]

How about running a standard deviation on my ten shot string mentioned earlier?

The 3050 fps round contradicts the stipulation of uniform ammo due to the hypothetical tight neck. If I get a chance later, the remaining nine shots could be an interesting study. What SPS does is to identify units outside of process expectations, those having "special cause" influences, such as your tight neck.

The technique doesn't dictate your results, but it does give you an indication of the capability of your process. When the results don't meet expectations, it gives you a place to start looking. You also need more data than a single 10-shot string; otherwise you're in the world of sampling, which is not a corollary to SPC.

The purpose of my commentary is to establish the relationship between ES and SD under conditions of an in-control process. If the results don't follow, then external influences exist, whether they be ambient temperature, component, measuring error, or procedure.
.

 

[Mikecr]

GG you describe a glass half full.
I'll describe a glass half empty.
I think the ammo described producing an ES of 50(because of whatever), sucks. You could generalize the results and accept things in a rosey way..
Or, drop the brass in a can and try again, and again, focusing your efforts until the ES problem is resolved. You might find another offender later, an H20 capacity mismatch, or eventually a die problem..

Regardless of the cause, your ES was 50. That's a cold hard fact

 

[davewilson]

Thanks everyone, i understand now.

 

[Forester]

GG you describe a glass half full.
I'll describe a glass half empty.
I think the ammo described producing an ES of 50(because of whatever), sucks. You could generalize the results and accept things in a rosey way..
Or, drop the brass in a can and try again, and again, focusing your efforts until the ES problem is resolved. You might find another offender later, an H20 capacity mismatch, or eventually a die problem..

Regardless of the cause, your ES was 50. That's a cold hard fact

Exactly, my take is you can no more throw out one unexplainable velocity than you can an uncalled flyer in a group. If you don't rectify the problem then that same shot could occur on a long range shot at the (fill in the blank) of a lifetime.

 

[Pdvdh]

You guys are missing the haystack for the needle that's inside of it. What GG gave us was a hypothetical set of data to explain, and help us understand, the role of ES and SD in objectively determining the quality of a load for long range work. If you guys want to throw his hypothetical load out the window because of one load that caused an ES of 50 when the remaining nine shots yielded an ES = 5, have at it. It would be your loss. A high ES (50) with an associated, unusually low SD of 15, means there are very few outlyers and that by far and away, the bulk of the data in the set are clustered closely together. This tells me that this load is worth investigating in the effort to identify and eliminate the cause of the sole outlyer. If you only pay attention to the ES in this hypothetical example, you'd be dismissing a great load because you either don't understand the significance of SD, or... worse. As GG has already stated, ES and SD are mathematically related and entwined in statistics.

I'd love to have a bunch of loads that match GG's hypothetical data set with an ES of 50. The more the better. Guaranteed I'd put them to good use. If the game animals I hunted knew the significance of an ES of 50 and an associated SD of 15, they'd be headed for the hills well before daylight.

Granted, you are completely correct that if you can obtain an ES of 5, you've always got a dynamite load - you're all set. That would be ideal. The ultimate. Lovely. But you'll work long and hard to accrue many of those loads. If you also include a consideration of SD, you may be able to ID some additional loads that will also perform outstandingly. As in the exaggerated hypothetical data set GG presented.

 

[getsmart]

OK, I will jump into this one.

My rough understanding of standard deviation is how close your data points cluster around the mean (average). Small SD most data are close to the average; large SD most data are further from the average.

I do not really look at SD, but this thread made me think about it some more (always a good thing). I guess now I have something else to fret over. My .300 Wby shoots 210 SMK at 2975 fps average with an ES of 27. A small SD could tell me my average velocity was a good to plug in a ballistics program. A large SD would tell me to beware. Looking at the data set would likely tell me the same thing, I think.

Is there anything else a small SD can tell you?

eric

 

[goodgrouper]

You guys are missing the haystack for the needle that's inside of it. What GG gave us was a hypothetical set of data to explain, and help us understand, the role of ES and SD in objectively determining the quality of a load for long range work. If you guys want to throw his hypothetical load out the window because of one load that caused an ES of 50 when the remaining nine shots yielded an ES = 5, have at it. It would be your loss. A high ES (50) with an associated, unusually low SD of 15, means there are very few outlyers and that by far and away, the bulk of the data in the set are clustered closely together. This tells me that this load is worth investigating in the effort to identify and eliminate the cause of the sole outlyer. If you only pay attention to the ES in this hypothetical example, you'd be dismissing a great load because you either don't understand the significance of SD, or... worse. As GG has already stated, ES and SD are mathematically related and entwined in statistics.

Exactly. I'm glad someone picked up on the idea of my post. It flew right by some others....

 

[Mikecr]

In your example GG you implied that someone would 'miss' a load with great potential, because of it's high ES. I implied just the opposite; that someone might generalize a low SD as good to go, ignoring ES.

Of course, nobody would do either.
In the original post, the shooter had all the data he needs. His ES already indicates clearly that SD would be useless for him.
He needs a better chronograph and/or load..

He's trying to see a half empty glass -as half full.

 

[goodgrouper]

In your example GG you implied that someone would 'miss' a load with great potential, because of it's high ES. I implied just the opposite; that someone might generalize a low SD as good to go, ignoring ES.

Of course, nobody would do either.
In the original post, the shooter had all the data he needs. His ES already indicates clearly that SD would be useless for him.
He needs a better chronograph and/or load..

He's trying to see a half empty glass -as half full.

I'm not sure what you mean here. In my 10 shot hypothetical group, one could and should ignore ES as it was a false indicator that the load was bad when in reality it was quite good and the SD would have shown it. To "generalize the SD as good to go" would be safe here in my opinion if the load gave good accuracy.

 

[Winchester 69]

Men occasionally stumble over the truth, but most of them pick themselves up and hurry off as if nothing had happened.
- Winston Churchill
.

 

[goodgrouper]

Men occasionally stumble over the truth, but most of them pick themselves up and hurry off as if nothing had happened.
- Winston Churchill
.

I wonder who this was directed at?