Statistics

[M-61]

Numbers lie. Targets don’t.

Yes.
Statistics tend to support the point the statistician is trying to make.

 

[dcali]

Numbers lie. Targets don’t.

oh, but they do... That's where those lyin' numbers come from.

 

[Chris Edmondson]

I'm a cause-effect person by nature, it's taken me a while to get over myself, lol. When I bought my first "real" rifle it was a solid MOA performer at 200 yards, I spent a great deal of time trying to figure out why it didn't shoot better. A buddy at the range suggested I let "Mikey" shoot it. Sure, why not I said. Mikey put all 5 rounds of factory ammo into one hole and announced, "It's not the rifle". He offered to do it again, that wasn't necessary.

I've come a ways since then

 

[jelenko]

you KNOW what works and what doesnt. There is no substitute for experience.
Only experience allows this..

Exactly.

Statistics is just a way to gain information on how something performs/reacts. It's great when there isn't much known about the system/object under consideration.

In making and tuning precision rifles, there is a TON on information gained from past experience. Anacnotes confirmed through various tests done by many different people. All done on a large population of rifles.
Personally, I find no surprise that, with all this accumulated knowledge, full statistical proof on a new member to the population is not needed.

 

[INTJ]

Thanks for tagging Mr Blaine, this ought to be a shit show! Lol

Mike, your question is actually quite valuable to the thread. So what you ask, answers itself is what I would say. Sometimes I run groups, but still all 12 at once. I really only round Robin when roughing in, or extremely fine tuning trying to decide between 2 "loads" that are extremely close in statistics to date. By comparing in this way, you are equalizing the conditions, good or bad you still see a winner and direction to proceed. When you're getting close, these 3 shot groups start to become 6 or even 9 shot groups as the increments become finer, which helps the confidence in the results as well. Yes, I've seen the occasional barrel not hold up for 5 or 10 shots, but the odds are against it from everything I've seen. Usually in those cases, those barrels never were capable in todays world.

Tom

For those wondering who to listen to, Tom should be at the top of your list for LRBR

 

[INTJ]

The good news is I only have to convince myself as to what I gather from a given ladder or not regardless of “depth” of data collected… what seems to get missed as Alex makes his point is the assumption over Short “time” with a fresh situation or barrel that it won’t support 100’s of ladders from previous ventures to a common result or trend.. the personal Data collection for guys like Glenn Kulzer Tom Mousel Alex wheeler or my Shawn Williams and many other Competitors in Long range bench rest actually is very “Deep” regardless the depth of collection a guy has to eventually pick a path that may or may not compete but eventually guys need to leave the comfort of there couch and Fuxx up some targets Lol all in good fun

Shawn Williams

Here is another guy to get info from for LRBR......

 

[RickyK]

I want to stir the pot a little on this discussion. To ad another point of view to the discussion around statistics. Lately theres been a lot of talk about what is statistically relevant when it comes to testing things or tuning a rifle. Most of the stuff I have seen comes from people with a scientific background. I wanted to ad a point of view from a background of building rifles that have to win matches and shoot the smallest aggs possible.
In my opinion, we want to tune a barrel in the least statistically relevant way possible SO LONG as it produces results. If I could tune a barrel in 5 shots and go break a record and win matches consistently with that method it would be the best tuning method there is. But also the least likely to stand up to scientific scrutiny.
The reality is that none of my tuning methods, or any of the methods used by my customers to win or break records would be considered statistically relevant. It blows my mind when I read some of this stuff. We have to test powders, primers, neck tension, seating depth, tuner test, ext. There are ways to do this with only a small sampling. When I am starting with a new rifle and I want to try 3 or 4 different powders, it only takes about 10 shots to know if that barrel likes that powder or not. Same goes for primers. Group size is another thing. 3-5 shot groups are not relevant? I dont have enough fingers and toes to count the records that say they are. What are we trying to do? Tune a rifle or create data?
What I am trying to stop is the guy that calls and says he is struggling to tune his rifle, hes shot a ton of groups. Piled up plenty of statistically relevant data, but has only tried one powder, one bullet, and one primer, and now has his barrel 1/3 shot out because he would not make decisions based on a small sampling.

Best post i have seen in a while !!!

 

[Ray Gross]

It has always been a pet peeve of mine when people quote SD's from 3 or 5 or even 10 shots.

 

[Coyotefurharvester]

I'm quite sure many matches have been won with loads that WERE NOT the best for the rifle, but that day the rifle/shooter/load combination topped all others. What the shooter choose to shoot that day was most likely based on actual previous targets.

 

[cdgaydos]

Equating "statistics" with requiring enormous amounts of data, thereby creating an approach to tuning that is mutually exclusive with "traditional", well established approaches is a pet peeve of mine...

Statistics is the analysis of data for the purpose of making relevant inferences. The amount of data necessary to make a practically significant inference is based on:
1. The normal variation in the system (how precise/accurate your system is...)
2. The amount of "difference" you decide is practically significant... (ex. how much improvement in group size, POI, etc is "important")
3. How certain you want to be of the results (how much risk are you willing to accept).

For an experienced BR shooter who knows his/her equipment and can honestly call shots, very little data is required to find the diamond in the turds... for less precise systems, more data will absolutely be necessary, unless the size of the differences is great or one is willing to accept more risk... note the times the "golden load developed with 3 shot groups turned out to be "ordinary" or a turd in disguise...

 

[Alex Wheeler]

A few of you are misunderstanding my point. I am not talking about shooters at all, I am talking about the people giving the advice to them that they need to shoot X amount of groups or X amount of shots in a group for it to mean anything. Yes, experience does lead to getting a tune quicker, but I think it applies to the new shooter as well. Even though I am involved in BR and F-Class, most of my customers are hunters. Most of the people I deal with are newer and less experienced shooters. Theres no way Im going to tell them to shoot 3 10 shot groups of something. First, you can cover a lot more ground with 3, 3 shot groups. Second, what happens as you shoot more? Barrel heat, shooter fatigue, mirage, more wind changes. You bring a lot more variables into the group the more you shoot. Even the top shooters rarely would shoot a 10 shot group outside of a match. And if they do, its to verify a load that they are pretty sure is the right one. Its never used to find a load.
Even a 3 shot group is a waste of time if your shooting a powder, bullet, primer, ext. the barrel does not like. Thats the point of smaller sample sizes that cover ground.
This whole statistics thing seems to have popped up recently. 5 years ago when a guy posted up a target you didnt see people posting how it meant nothing because there was not enough shots fired. I dont think everyone became statistic experts all of a sudden, I think they are just following some of this info being putout. And I think its also good to get input from the other side as well.
I do agree that we dont shoot enough to come up with solid es or sd numbers. But es and sd dont effect accuracy until they get extreme, so I dont really care.

 

[dcali]

A few of you are misunderstanding my point. I am not talking about shooters at all, I am talking about the people giving the advice to them that they need to shoot X amount of groups or X amount of shots in a group for it to mean anything. Yes, experience does lead to getting a tune quicker, but I think it applies to the new shooter as well. Even though I am involved in BR and F-Class, most of my customers are hunters. Most of the people I deal with are newer and less experienced shooters. Theres no way Im going to tell them to shoot 3 10 shot groups of something. First, you can cover a lot more ground with 3, 3 shot groups. Second, what happens as you shoot more? Barrel heat, shooter fatigue, mirage, more wind changes. You bring a lot more variables into the group the more you shoot. Even the top shooters rarely would shoot a 10 shot group outside of a match. And if they do, its to verify a load that they are pretty sure is the right one. Its never used to find a load.
Even a 3 shot group is a waste of time if your shooting a powder, bullet, primer, ext. the barrel does not like. Thats the point of smaller sample sizes that cover ground.
This whole statistics thing seems to have popped up recently. 5 years ago when a guy posted up a target you didnt see people posting how it meant nothing because there was not enough shots fired. I dont think everyone became statistic experts all of a sudden, I think they are just following some of this info being putout. And I think its also good to get input from the other side as well.
I do agree that we dont shoot enough to come up with solid es or sd numbers. But es and sd dont effect accuracy until they get extreme, so I dont really care.

I think what you're getting at is context. And you're right. Blowing through 100 rounds just because it's going to give you a text book confidence of 90% isn't usually going to be what you want. I'm going to go out on a limb and say zero shots is the right number for a hunting rifle - load development is just not needed. If you want to tell whether neck turning matters with high confidence, you're going to need a LOT of shots. Probably too many to bother with because you'll burn out the barrel before you get to that goal.

What's going to waste your time and money is not recognizing context. Shooting 200 rounds to develop a hunting load is a waste of 200 rounds of ammo and barrel. Shooting 10 shots and declaring that neck turning is significant is a waste of 10 shots and barrel (and you very well could be wrong, which is also a big waste in the future).

It's all about knowing what you can get away with, and relying on other shooters as much as you can. You don't need to try too many combos if you're shooting F TR with a .308 - putting a Berger 200.20X or Juggernaut on top of Varget is almost certainly going to work. So you're starting out WAY ahead. It's all a matter of what you're trying to do. If you want to try a different bullet, you're still in the right ballpark.

One thing I will say is that sometimes shooting a lot just to get data is interesting and illuminating. If you find that sort of thing fun, you can learn a thing or two. It's just not necessary to develop a good load, and shooters would be wise to not confuse the two.

 

[Bottom Fodder]

I keep an file using Ecel to tracks data for each gun that keeps up with load data, round count, starting scope settings for the matches I shoot along with any setting changes I made during the match along with My score compared to others shooters along with a few other details . Mainly just to help me remember lol

 

[Alex Wheeler]

I think what you're getting at is context. And you're right. Blowing through 100 rounds just because it's going to give you a text book confidence of 90% isn't usually going to be what you want. I'm going to go out on a limb and say zero shots is the right number for a hunting rifle - load development is just not needed. If you want to tell whether neck turning matters with high confidence, you're going to need a LOT of shots. Probably too many to bother with because you'll burn out the barrel before you get to that goal.

What's going to waste your time and money is not recognizing context. Shooting 200 rounds to develop a hunting load is a waste of 200 rounds of ammo and barrel. Shooting 10 shots and declaring that neck turning is significant is a waste of 10 shots and barrel (and you very well could be wrong, which is also a big waste in the future).

It's all about knowing what you can get away with, and relying on other shooters as much as you can. You don't need to try too many combos if you're shooting F TR with a .308 - putting a Berger 200.20X or Juggernaut on top of Varget is almost certainly going to work. So you're starting out WAY ahead. It's all a matter of what you're trying to do. If you want to try a different bullet, you're still in the right ballpark.

One thing I will say is that sometimes shooting a lot just to get data is interesting and illuminating. If you find that sort of thing fun, you can learn a thing or two. It's just not necessary to develop a good load, and shooters would be wise to not confuse the two.

I agree if your trying to prove something, like how a tuner acts, you need to shoot a lot and gather that data. But the guy that just wants to use a tuner does not. As an example. But thats not whats being put out by everyone.

 

[JFrank]

When I first joined AS I was finding a basic tune window in about 10 shots, then I learned how I was doing things wrong and that would never work very well, so now days I use more shots to get the same information.

 

[gunsandgunsmithing]

I agree if your trying to prove something, like how a tuner acts, you need to shoot a lot and gather that data. But the guy that just wants to use a tuner does not. As an example. But thats not whats being put out by everyone.

I'm not following you. My tuner test, that some refer to as a sine wave test, is about seeing sweet spots at predictable places along the waveform as well as predictable group shapes on either side of a sweet spot..at those proper poi's. IOW, using several three shot groups that are supported by predictable shapes as well as where they hit on the target. I'm not a statistician but the probability to have all those things come together and be predicted before the test target is even fired...well, has to be pretty phenomenal. I've reviewed well over a thousand of these test targets now and it virtually always happens. That test is just solid gold!

As I said though, I may not be understanding what you're saying and there are many other methods out there that definitely seem to be, lets say, not nearly as methodical and are highly likely to give less than predictable results. Just like loading ammo, I wouldn't move the tuner randomly and expect to get anything repeatable from it.

 

[chkunz]

Equating "statistics" with requiring enormous amounts of data, thereby creating an approach to tuning that is mutually exclusive with "traditional", well established approaches is a pet peeve of mine...

Statistics is the analysis of data for the purpose of making relevant inferences. The amount of data necessary to make a practically significant inference is based on:
1. The normal variation in the system (how precise/accurate your system is...)
2. The amount of "difference" you decide is practically significant... (ex. how much improvement in group size, POI, etc is "important")
3. How certain you want to be of the results (how much risk are you willing to accept).

For an experienced BR shooter who knows his/her equipment and can honestly call shots, very little data is required to find the diamond in the turds... for less precise systems, more data will absolutely be necessary, unless the size of the differences is great or one is willing to accept more risk... note the times the "golden load developed with 3 shot groups turned out to be "ordinary" or a turd in disguise...

I think this is good advice, if I want to know if load A is better than load B it takes less data to answer the question if the difference in the two is large. But the smaller the difference (variance) the more data it takes to answer the question with a reasonable confidence level. I understand statistics well enough to use it but not well enough to teach it.

 

[J.d. Popkes]

In addition, sites like Accurate Shooter have given me access to a ton of valuable information from other shooters. I don't buy into everything I read here but I have learned a lot from others that I can quickly validate and add to my trusted data base.

^^^^^^this

Being able to sort out the BS is a big part of any forum. People are interesting sometimes, and impractical. Others love to hear themselves talk, feel the need to debate and be a naysayer, basically post a resume, probably too intelligent for any practical purpose.

Separate the wheat from the chaff, I believe is the saying.

 

[Webster]

I want to stir the pot a little on this discussion. To ad another point of view to the discussion around statistics. Lately theres been a lot of talk about what is statistically relevant when it comes to testing things or tuning a rifle. Most of the stuff I have seen comes from people with a scientific background. I wanted to ad a point of view from a background of building rifles that have to win matches and shoot the smallest aggs possible.
In my opinion, we want to tune a barrel in the least statistically relevant way possible SO LONG as it produces results. If I could tune a barrel in 5 shots and go break a record and win matches consistently with that method it would be the best tuning method there is. But also the least likely to stand up to scientific scrutiny.
The reality is that none of my tuning methods, or any of the methods used by my customers to win or break records would be considered statistically relevant. It blows my mind when I read some of this stuff. We have to test powders, primers, neck tension, seating depth, tuner test, ext. There are ways to do this with only a small sampling. When I am starting with a new rifle and I want to try 3 or 4 different powders, it only takes about 10 shots to know if that barrel likes that powder or not. Same goes for primers. Group size is another thing. 3-5 shot groups are not relevant? I dont have enough fingers and toes to count the records that say they are. What are we trying to do? Tune a rifle or create data?
What I am trying to stop is the guy that calls and says he is struggling to tune his rifle, hes shot a ton of groups. Piled up plenty of statistically relevant data, but has only tried one powder, one bullet, and one primer, and now has his barrel 1/3 shot out because he would not make decisions based on a small sampling.

I look at a lot of videos by top shooters because it's interesting to see what the best in the world do. It's seems interesting that many of the top shooters find a good load with maybe under 50 shoots. Large numbers of shooters on this website are obsessed with 10 details and they are always looking for something they cannot find. They assume the barrel is good and if the try enough loads it will shoot small groups. Of course the best shooters have great equipment and shooting skills. They quickly decide if the barrel isn't good enough. Top shooters buy large numbers of barrels every year. If the barrel doesn't seem to perform, they sell it for a hunting rifle. Most of the people working up loads by group size have no idea what their skill, bench setup, or conditions contribute to the group size. I keep quoting Tony Boyer. Here goes Tony said he used to give bench rest lessons after a while he said he got tired of telling students that the rifle they just put a lot of money into wasn't up to par and would never be competitive. He quit giving lessons. Tony once stood in front of a good shooter and to the side. After each shot he told the shooter where the shot would fall on the target based on how he held the rifle. He always called it correctly (high, left ect.). I'm a varmint hunter with a nice rifle so I don't have to get OCD with details and statistics. I have done failure analysis my entire career and tend to say this is good enough for what I need.

 

[Alex Wheeler]

I'm not following you. My tuner test, that some refer to as a sine wave test, is about seeing sweet spots at predictable places along the waveform as well as predictable group shapes on either side of a sweet spot..at those proper poi's. IOW, using several three shot groups that are supported by predictable shapes as well as where they hit on the target. I'm not a statistician but the probability to have all those things come together and be predicted before the test target is even fired...well, has to be pretty phenomenal. I've reviewed well over a thousand of these test targets now and it virtually always happens. That test is just solid gold!

As I said though, I may not be understanding what you're saying and there are many other methods out there that definitely seem to be, lets say, not nearly as methodical and are highly likely to give less than predictable results. Just like loading ammo, I wouldn't move the tuner randomly and expect to get anything repeatable from it.

I mean if your going to try to figure out how or why something does what it does, that requires a lot of data to prove your theory. But if you just want to use it, you can shoot a relatively small amount to make it work.