Let’s talk load development

[tangolima]

Originally Posted by Metal god

When measuring this way you want the mean radius center to be POA ?

No you need to calculate the group center first, which is the average of all the shots. For instance, you fired 5 shots, whose coordinates are

(x1,y1), (X2,y2), (x3,y3), (x4,y4), and (x5,y5)

The group center is at (xc,yc)

xc=(x1+x2+x3+x4+x5)/5
yc=(y1+y2+y3+y4+y5)/5

The center is the centroid or CG of the group if you will.

-TL

Sent from my SM-N960U using Tapatalk

[stagpanther]

Here's a pretty nice article that I found that does a good job of explaining different concepts (by "good", I mean I can actually understand what they are talking about). Something I notice that is inherent in many discussions is how the actual written language used is important in conveying the meaning--and can introduce "grammatical dispersion," if you will.

[Metal god]

Thanks stag very helpfull and maybe I should finish reading it before asking this question but….

Why isn’t POA calculated in this ? Is it because we are constantly changing our POA throughout a string to adjust for wind and other things like where the first shot hit and subsequent shots POI ? Is it assumed your point of aim on a target is the same spot every time .

I’ve almost always done my load development at 100yds so wind has very little influence on the bullet poi . I try to shoot the smallest group and I’ve never cared where that group is as it relates to POA . Once I have a load that shoots tight groups I know I can shoot that with accuracy.

Example :

This 5 shot group at 300yds had at least 3 different POA to get all 5 shots on target , I danced a little jig after that group . That load was worked up all at 100yards to dial it in . Speaking of that , I forgot all about that load from 9 years ago . I think I want to resurrect that and see if it still shoots well. Although I pulled the barrel off of the rifle that shot that group and now it has a different barrel on it so that load may not be as well as it once was .



I guess my overall point is it seems weird that you’re worried about where POI is during load development if your POA is moving around . I think of it like moving POA around when you’re trying to site in a scope. I’ve always found the load that shoots the smallest groups then DOPE my riffle to that load . I usually have my rifle DOPE’d to one specific load then DOPE all other loads around the first loads scope setting which is zero/zero at 100yds

Like this 10 shot group at 100yds , does this group suck cus it’s an inch high and an inch left ? Not in my book I never did use it though because it had way too low of velocity than I was wanting at the time . Not even sure why I was testing such a low charge. There must’ve been some reason, but I didn’t write down what that was. .


Or am I completely misunderstanding this because to me it seems like mean radius shouldn’t be used during load development and only on testing how accurately you can shoot once the load has been developed ?

[stagpanther]

I totally get why you are asking about the reference to POA--I feel the same way to some degree. I think the principal reason it's not given priority in these methodologies is that the POA relationship (accuracy) is assumed to be a function of the mechanics of gear set-up, shooter technique and environmental conditions assessment. Obviously they have influence on the dispersion of a given group--but aren't really useful in predicting probabilities of where your shots will land (consistency) for a given cartridge formulation.

I personally feel (at the risk of getting into another discussion of what "turns a bullet" )that the fluid dynamics of the wind over distance is an area that can be improved upon to provide predictive information; but that's a tough one which would need lots of computational power. On the other hand, we are entering the age of AI--it's conceivable that some of these complex computations may one day--maybe even soon--be as simple as asking a question of your personal device.

[stagpanther]

PS--I did finally decide to subscribe to Ballistic-X's advanced toolkit; I figure they have a good horse in the race.

[tangolima]

When we shoot group to measure consistency, we keep the same POA. Where the POA is doesn't matter as long as it is constant for the group.

However, I have been practicing to put 10 rounds into a designated target. I have to constantly change POA, or even scope's W/E, to counter wind conditions. I record the following info for each shot

POI, POA, MV, W/E, hit/miss

When I process the data, I "normalize" POI to remove effects of POA and W/E. The group I get is as if I have constant POA and W/E. Then I calculate group center and mean radius. That is a measure how good my load is. The hit rate on the designated target is how well I shoot the load.

-TL

Sent from my SM-N960U using Tapatalk

[Metal god]

When I process the data, I "normalize" POI to remove effects of POA and W/E.

How does that work when your wind call can be wrong ( POA ) ?

Example ;

When we were shooting 22 long rifle at 150 yards at that can . When I zeroed my rifle before we started, the wind was blowing 3 to 5 miles an hour right to left. Those last three shots that I took that caused me to lose the match hit consistently high right almost in the exact same spot was likely due to me failing to read the wind . I never looked up at the flag to see if the wind had changed . . It likely slowed or stopped during that three shot string. Causing my poi to change from POA . So that doesn’t mean my load or the rifle wasn’t shooting correctly. That was all on me making a bad wind call . If you’re making bad wind calls how do you calculate your groups? This is assuming you don’t realize you made a bad wind call . You could be compensating for a 4 mile an hour wind, but it’s really eight. Meaning you think you read it right even when calculating the group, but it was wrong the whole time.

[tangolima]

That's exactly what I mean. Normalization removes the effect of POA, right or wrong.

For instance, I felt wind on my right cheek, so wind is from right to left. I shift POA 1" to the right of the target center, coordinates (1,0), and fire. POI is 1" to the right of the target, coordinates (1,0), so the wind call was wrong. Normalization

POI - POA = (1,0)-(1,0)= (0,0)

So if I didn't make the call the shot should have been right on target center.

-TL

Sent from my SM-N960U using Tapatalk

[Metal god]

No, I understand that. But what if you’re calculating a 4 mile an hour wind and it’s really eight . When you’re back at home calculating those numbers or even at the range and you believe your point of aim was accurate for what you perceived to be 4 mile an hour wind was actually wrong because it was an 8 mile an hour wind . That means your point of impact on the target as a relates to your point of aim. Was still wrong. Even though you believed you called it right . My point is I can’t imagine any of us calling the wind exactly right every time.

Maybe you need to describe what normalize means , I might be misunderstanding what that term means because in my head it means you’re changing something from what it was to what it needs to be to calculate ? If so, how do you know what to change the first equation to if you don’t know if the first equation is even accurate ?

[Unclenick]

Originally Posted by Tangolima

Poi distance from group center (radius) follows Rayleigh distribution…

No. It's a normal bivariate distribution, or half a bell curve. A Rayleigh distribution goes to zero at zero, so if the radial distribution followed it, there would be an untouched spot in the center of every group, followed by a rapidly growing shot density spreading away from the center that then fell off much more gradually. Instead, you get a major and a minor axis combined pair of bell curves that cross and peak at the center, corresponding to where the shot density is greatest.

The source of radial standard deviation being the most efficient method of group evaluation is partially described in the first two paragraphs, here.

[tangolima]

Originally Posted by Metal god

No, I understand that. But what if you’re calculating a 4 mile an hour wind and it’s really eight . When you’re back at home calculating those numbers or even at the range and you believe your point of aim was accurate for what you perceived to be 4 mile an hour wind was actually wrong because it was an 8 mile an hour wind . That means your point of impact on the target as a relates to your point of aim. Was still wrong. Even though you believed you called it right . My point is I can’t imagine any of us calling the wind exactly right every time.



Maybe you need to describe what normalize means , I might be misunderstanding what that term means because in my head it means you’re changing something from what it was to what it needs to be to calculate ? If so, how do you know what to change the first equation to if you don’t know if the first equation is even accurate ?

Ok I guess you really mean how to remove the effect of wind on the group. I can't. I have to accept the wind at the time of test firing is part of the load.

150 yd at Escondido is not too bad for center fired. For .22lr, I ignore horizontal dispersion and focus on vertical dispersion only.

-TL

Sent from my SM-N960U using Tapatalk

[tangolima]

Originally Posted by Unclenick

No. It's a normal bivariate distribution, or half a bell curve. A Rayleigh distribution goes to zero at zero, so if the radial distribution followed it, there would be an untouched spot in the center of every group, followed by a rapidly growing shot density spreading away from the center that then fell off much more gradually. Instead, you get a major and a minor axis combined pair of bell curves that cross and peak at the center, corresponding to where the shot density is greatest.



The source of radial standard deviation being the most efficient method of group evaluation is partially described in the first two paragraphs, here.

There are different theories, all of which are just models with different degrees of validity. Rayleigh is surely one of the models. It is based on normal distribution of horizontal and vertical dispersions.

Rayleigh show zero pdf at r=0. But it doesn't necessarily means a distinct hole of discernable radius around the bullseye with no shots, or a safe zone if you will. In this case, pdf can be visualized as number of shots landing on the circumference of a circle with radius r, divided by the circumference. r=0 has zero circumference, so of course number of shots would be zero. It only means zero probability of hitting the dead center of the bullseye. Infinitesimally short distance away from the dead center, the probability is none zero.

In one of the "physics of shooting a rifle" threads, I did a derivation. Suppose the shot distribution is Rayleigh. A fly of finite (none zero) body size. Where is the safest spot to land on the target? Bullseye is the worst. Best to stay as far away from the center as possible.

Thanks for more reading materials!

-TL

Sent from my SM-N960U using Tapatalk

[stagpanther]

After adding the advanced toolkit in Ballistic-X, I redid one of the groups--in this case the 25 PRC shooting 135 gr berger hybrids. I believe the red circular grid overlay is the mean radius--the legend also includes CEP and radial SD measurements.

[stagpanther]

I just compared to the original before I added the advanced calculations; I can't remember for sure what the actual POA was since I don't see it on the first picture; but the mean raduis would still be the same--the offset measurements would not be. Sorry bout that.

[tangolima]

Originally Posted by stagpanther

After adding the advanced toolkit in Ballistic-X, I redid one of the groups--in this case the 25 PRC shooting 135 gr berger hybrids. I believe the red circular grid overlay is the mean radius--the legend also includes CEP and radial SD measurements.

Putting it in Rayleigh distribution model.

50% group diameter is 2.04" or 0.55moa

90% group diameter is 3.7" or 1.00moa.

Top notch at 353yd.

Note that the number the apps gives out is radius. x2 to make diameter. And there is uncertainty (error) due to the number of samples. 10 is not bad but still considered small.

How do you compare different loads?

-TL

Sent from my SM-N960U using Tapatalk

[stagpanther]

How do you compare different loads?

Formally, I used the ES knucklehead approach. But now I'm going to explore the use of these other techniques--keeping in mind that I need to be able to properly interpret the predictive value of whatever measure I use. I like shooting at longer range for the tests because 1) it spreads the impacts out; and 2) I know the closer ranges will be at least as good, if not better. The one drawback at greater distance is that environmental and user influences will have a greater impact on the measurements--harder to control, in other words.

[Metal god]

Ok I guess you really mean how to remove the effect of wind on the group. I can't.

Yes but no . You can totally calculate for wind if you know exactly what it is/was for each individual shot . What you can’t calculate is how far off you read the wind . Unless you have a wind meter in at least two locations down range you have no idea what effect the wind had on the trajectory. When you add in human error and how well the load groups . That all effects POI , ES , dispersion etc . Im just having a hard time understanding if you don’t really know the numbers you are attempting to calculate. How reliable are those calculations ?

Im a visual person , maybe next time we shoot you csn show me what you are doing ? I think l’ll understand it better that way then here trying to use reading comprehension which is not something Im best at .

[akinswi]

Back to the video,

I started a verison of this about a year ago. I stopped weighing brass, sorting bullets etc. I noticed no changes in group size or scores.The main thing I started enjoying reloading again, because I wasnt spending hrs of tedious tasks such as, sorting brass and bullets etc. I tried so many rabbit holes that I would forget which rabbit hole I went down even with my reloading records.

I do keep all my lots the same, Powered, bullets and brass.

I also got too shoot more, which I found to help my groups anyways. From now on I try too keep rabbit holes to a minimum

[stagpanther]

I do keep all my lots the same, Powered, bullets and brass.

I can't do that nowadays--too expensive to buy things in mass quantities of the same lot.

[Metal god]

I started a verison of this about a year ago. I stopped weighing brass, sorting bullets etc. I noticed no changes in group size or scores.The main thing I started enjoying reloading again, because I wasnt spending hrs of tedious sorting brass and bullets etc and i tried so many rabbit holes that I would forget which rabbit hole I went down even with my reloading records.

Amen brother ! I was starting to feel the same way about eight years or so ago . Reloading and load development seemed more like work and was less and less enjoyable . I’d spend 6 to 8 hours at the range testing multiple rifles and loads all day long .

I have been telling people that I haven’t done any load development or shooting for real accuracy in about three years . However, If I really calculate it., it’s more like at least two years before the pandemic so it may be as long as seven years ago that I’ve done any real hard-core, testing or shooting, and the rust shows . Each time I go out I shoot a little better than the last . I’m feeling a little more comfortable behind the rifle each time I go. . I forget half the things you’re supposed to do . Like what I said above about shooting 22 long rifle at 150 yards and completely forgetting or ignoring the wind. I mean, who the hell shoots a 22 long rifle that far and just ignores what the wind could be doing . Total face plant. Haha

-too expensive to buy things in mass quantities of the same lot.

Oh man, check this out, a few weeks ago I bought three or 4 pounds of IMR 4064 and when I got home, I checked the lot numbers and it matched some powder. I bought eight or 10 years ago that I was using at the time. . Wow right ? Now I have like 6 1/2 pounds of the stuff all in the same lot number.

[stagpanther]

Oh man, check this out, a few weeks ago I bought three or 4 pounds of IMR 4064 and when I got home, I checked the lot numbers and it matched some powder. I bought eight or 10 years ago that I was using at the time. . Wow right ? Now I have like 6 1/2 pounds of the stuff all in the same lot number.

LOL I bought tons of 4064 about 10 years ago as well, I think the idea was to use them in 44 mag loads IIRC; but ended up sticking with hotter loads in H110/win 296. Wouldn't surprise me if my 4064 lot numbers matched some of yours.

[stagpanther]

This 5 shot group at 300yds had at least 3 different POA to get all 5 shots on target , I danced a little jig after that group . That load was worked up all at 100yards to dial it in . Speaking of that , I forgot all about that load from 9 years ago . I think I want to resurrect that and see if it still shoots well. Although I pulled the barrel off of the rifle that shot that group and now it has a different barrel on it so that load may not be as well as it once was .

That looks like a competition-winning group (not that I know anything about comps, having never participated in one, but hey, this is the internet!). But your skill in adjusting the windage is a factor independent of the consistency probability of your load.

[tangolima]

Originally Posted by Unclenick

No. It's a normal bivariate distribution, or half a bell curve. A Rayleigh distribution goes to zero at zero, so if the radial distribution followed it, there would be an untouched spot in the center of every group, followed by a rapidly growing shot density spreading away from the center that then fell off much more gradually. Instead, you get a major and a minor axis combined pair of bell curves that cross and peak at the center, corresponding to where the shot density is greatest.



The source of radial standard deviation being the most efficient method of group evaluation is partially described in the first two paragraphs, here.

I went through the cited literatures.

Normal bivariate distribution is a general case of Rician distribution that x- and y-components are correlated. With zero correlation, it becomes Rician distribution, of which Rayleigh distribution is a special case. For our applications, we have little or no reason to have vertical dispersion related to horizontal dispersion. So Rician or Rayleigh distribution model is not invalidated.

The 2nd citation from Kolbe's book is actually discribing Rayleigh distribution. Square root of the sum of x- and y- normal distributions is the parameter sigma of a Rayleigh distribution.

We are actually talking about the same thing.

-TL

Sent from my SM-N960U using Tapatalk

[stagpanther]

Ok--let's have some fun analyzing this group I shot this afternoon. Today was a very gusty day--while the peak gusts were not much over 20 mph, they varied in velocity very quickly, sorta a pulsing type gust, and worse yet would change direction from quartering headwind to quartering tail wind. For this reason I pulled the distance in to get away from some of the big gravel piles which I know would just turbulate the gusts. I can't read the wind in terms of adjusting windage to the left or right--rather, I tried to time the shots to the beginning or end of the gust cycles, but kept the same settings and POA. I even made an cartridge for a 1st shot bore-fouler from a cold clean bore--but it smacked almost dead-center on the POA, so I figured what the heck just include along with the next 10 shots. Today I thought I did a better than usual job of holding the rifle steady through the shots, so I did not call any fliers (for a change ). The group size in terms of CTC ES is not very remarkable, but some of the other measurements look better.

I think what would help me and others would be an expert interpretation of what the different measurements mean in terms of what a shooter could expect in probabilities--and what, if any, decisions would you make had this been a group you shot.

Fire away!

[tangolima]

Originally Posted by stagpanther

Ok--let's have some fun analyzing this group I shot this afternoon. Today was a very gusty day--while the peak gusts were not much over 20 mph, they varied in velocity very quickly, sorta a pulsing type gust, and worse yet would change direction from quartering headwind to quartering tail wind. For this reason I pulled the distance in to get away from some of the big gravel piles which I know would just turbulate the gusts. I can't read the wind in terms of adjusting windage to the left or right--rather, I tried to time the shots to the beginning or end of the gust cycles, but kept the same settings and POA. I even made an cartridge for a 1st shot bore-fouler from a cold clean bore--but it smacked almost dead-center on the POA, so I figured what the heck just include along with the next 10 shots. Today I thought I did a better than usual job of holding the rifle steady through the shots, so I did not call any fliers (for a change ). The group size in terms of CTC ES is not very remarkable, but some of the other measurements look better.



I think what would help me and others would be an expert interpretation of what the different measurements mean in terms of what a shooter could expect in probabilities--and what, if any, decisions would you make had this been a group you shot.



Fire away!

50% group diameter 1.88" or 0.76moa
90% group diameter 3.44" or 1.38moa.

Excellent!

Interpretation. If you fire a lot of shots, half of them falls inside circle of 1.88" diameter. 90% falls inside circles of 3.44" diameter.

-TL

Sent from my SM-N960U using Tapatalk