Bullets matter!

[JEBruns]

I know everyone knows that, but here is a good example. All 3 rows are using the same powder (H335) charges, 5 of each charge for each bullet. First row is 52gr SMK's. 2nd row is RMR bulk Hornady 55gr FMJ-BT with cannelure. 3rd row is RMR bulk 55gr FMF-BT. All shot out of a new Tikka T3X Super Varmint in a Whiskey-3 chassis. Groups shot right to left with the highest charges on the right.

Some decent groups in the first 2 rows I'll focus in on more. But I found it interesting how much the 3rd row opened up. I did let the gun cool down between ladders. And I shot a control group of a known very accurate load at the end. The control group was very tight at just under .300.

EDIT: 100 yards, .223

 

[badkarmamib]

I run 150 SSTs out of my .30/06 Remington 700 ADL. I load 150 SP's for my son's Howa 1500, since I had bought a bunch for my Garand. I thought I could save money by switching to the SP's in my Remington 700. Nope. All else being equal, group size doubled at 100.

 

[red rick]

Some good groups in the first 2 rows . I would clean the barrel and shoot that 3rd row again .

 

[Blue68f100]

You really need to work up a load for each bullet to do a true comparison. I've never found swapping bullets gave me the best groups.

 

[JEBruns]

You really need to work up a load for each bullet to do a true comparison. I've never found swapping bullets gave me the best groups.

That's what I was trying to do. Each of the 5 targets on a given row were a different powder and bullet charge. What I'll do with the results is fine tune the better loads and dial the SMK's and RMR Horn 55's in more. I won't put any more effort into the bottom row's results with H335. Or CFE223, from prior testing. I just can't get those to shoot well out of any of my .223's. I bought 2000 of them, so I'd just like to find a good recipe for them too.

 

[eddiememphis]

What distance?

 

[Varminterror]

A node takes 4 charge weights, at minimum, to define. Two inside, and one on each side which are outside of the bounds. 5 groups of 5 shots isn't a very conclusive test.

To look at each row, I'm not certain I see anything which suggests any group stands out as better or worse than the rest in each series. I'm not certain I see anything in the test methodology which should produce differentiated results. But 75 rounds of burnt powder...

What's the goal for this load with these bullets? What are you wanting to do with the rifle? What does "success" look like in this particular pursuit?

 

[ackley man]

In BR a 3+1 or 4+1 group indicates your load is a little too hot. Almost all of your targets in row 1 show this.

 

[JEBruns]

What distance?

Sorry, thought I'd mentioned that. 100 yards.

 

[JEBruns]

A node takes 4 charge weights, at minimum, to define. Two inside, and one on each side which are outside of the bounds. 5 groups of 5 shots isn't a very conclusive test.

Didn't mean for them to be conclusive, just a starting point. As I said above, now I'll do some fine tuning on the better groups.

It was conclusive with the 3rd row though.

To look at each row, I'm not certain I see anything which suggests any group stands out as better or worse than the rest in each series. I'm not certain I see anything in the test methodology which should produce differentiated results. But 75 rounds of burnt powder...

Interesting. I certainly see, and calipers back it up, some groups that are much better than others.

What's the goal for this load with these bullets? What are you wanting to do with the rifle? What does "success" look like in this particular pursuit?

Goal is to see if I can find what I consider good groups using H335 in my new Tikka. Have had no success with that powder in my AR's.

Success is repeatable, sub-.300 5 shot groups at 100yds.

 

[Charlie98]

good groups using H335

H335 is a great blasting powder, but I've never found it to create particularly great groups.

some groups that are much better than others.

Indeed! I would say that not only do bullets matter, but powder charge as well. Just looking at them, I would say your Tikka likes the lower charges better.

 

[eddiememphis]

I think the clear take away from this is how far guns and ammunition have come.

Even as recently as 20 years ago, the shot groups as tight as the ones on the bottom row would've been cause for celebration

 

[maxxhavoc]

@Varminterror
A lot of people have been working up loads in this method for a very long time, and it does work. Maybe not as efficiently as other methods, but it does work. There is a rather obvious "analog" trend of these groups to group better at lower powder charges. It would be worthwhile to look for "the load" on this in the leftmost targets on this, and maybe just a bit further left on the last one (if you even want to use them at all.)

I am genuinely interested in YOUR methodology. How do you generally develop a load? Does it require more or less rounds fired? Why is it better? If it requires to have groups of n>30 to establish a proper statistical distribution, that may just be too time consuming (and primer-consuming!)

Not trying to bicker, just curious to hear another method than this. You do not sound like you lean toward a ladder test either.

 

[Varminterror]

A lot of people have been working up loads in this method for a very long time, and it does work

The reality is this:

1) They likely could have chosen any of the charges on the page, randomly, and had the process "work" just as well.

2) They likely missed out on truly improved performance by chasing ONE thing they think "worked better" but didn't actually understand that it wasn't truly better in any way.

In other words, folks pretend this method works because they kinda ended up with a small shooting group (often ignoring how often they say "I don't know what happened, that load was awesome, but it all fell apart," or "it wants to shoot small, but I throw flyers out of the group sometimes," or "all day long if I do my part,"....). What they do NOT do, however, is do further experimentation which proves those OTHER parameter combinations wouldn't work just as well, or even better.

I am genuinely interested in YOUR methodology. How do you generally develop a load? Does it require more or less rounds fired? Why is it better? If it requires to have groups of n>30 to establish a proper statistical distribution, that may just be too time consuming (and primer-consuming!)

My methodology isn't mine - it's a regurgitation of methods I was taught by other shooters and which are commonly known the world over. My personal adoption of these methods does use a VERY low round count, because I don't let myself chase smoke. There's 75 shots on that page - and the OP claims it is a starting point... Were it MY rifle, I'd be done already by 75 rounds - and I wouldn't be looking at these 75 shots and thinking I needed to explore anything else. None achieve the goal stated, and none appear to have a true advantage - a true differentiation - which suggests any of them are better than the other. So I'd ditch that load, either ditch the bullet, ditch the powder, or ditch the barrel - or all of the above.

For my PRS rifles, typically my load development is ~40 rounds for a new bullet and cartridge. 10 to foul, 3 shot groups for 10 weights, and I find regions which are NOT differentiated, and that's my load. They shoot small, they shoot stable, and the load is forgiving. I'm not wasting time (or primers) looking for needles in haystacks, I just pick my piece from the side of the cake which doesn't have a booger on it...

There's nearly-exhaustive science out there which effectively disproves, time and time again, the methods we all used - MYSELF INCLUDED - because we all chased smoke at some point. "A lot of people have been working up loads [with] methods for a very long time" which don't actually work. Folks willingly assume that there is an immediate, irrefutable response for all groups fired to the input variable changes, rather than acknowledging that in reality, they are simply seeing statistically certain probabilities. Shoot 5 groups, ONE will be smaller than the rest, ONE will be larger - do we assume all of our ammo is different if we do that with the same lot of ammo? Or do we accept that it's simple statistical distribution? Accept that rifles just don't shoot the same group size for every group they ever fire? Yes, we realize that group size isn't a fixed result - there is a distribution with standardized error margin. So why then are folks so damned certain that EVERY hole on a page during load development is meaningful - that EVERY change we make in one variable or another is ALWAYS driving a difference on the page?

You do not sound like you lean toward a ladder test either.

One reason "similar POI" focused methods like the Audette Ladder, the Newberry OCW, or Satterlee's rendition of the Audette Velocity curve have been proven to WORK is because they do NOT try to prove groups are "different," but rather these methods look for indication that groups are NOT different. There's SOME invalidity in the analysis as to the ANOVA between what we assign as "anti-nodes" versus what we believe is in the node (the non-differentiated portion of the ladder), but overall, ladder tests aren't looking for what is different - not looking for the small group among larger groups - it's looking for NON-differentiated results. Proving a null hypothesis is much, much easier than proving a differentiation. So in these methods, we are NOT looking for the ONE small group, and we're NOT chasing our tail by jumping up and down because one random group was coincidentally smaller than the others - but rather, we're looking for a cluster of results with relative insensitivity - where the change didn't appear to make a difference.

Which is really my point in responding to the OP in this thread - he can chase his tail and convince himself that the ONE small group on the page is meaningful, OR he can take heart that no matter what powder charge in that range, that powder and bullet wants to do the same thing, over and over, so he can trust those results.

 

[Varminterror]

@JEBruns - what's the diameter of the outer circle on these? If these are the USBR 50yrd targets, I think they would be 2.05", is that correct?

I'd never say anyone SHOULD do ANOVA or T tests on their load development targets, but confirming that information, I can offer some insight as to why I encourage folks to question whether distinctions on target arrays are actually differences or not. I see a LOT of folks finding confidence in coincidence - and as someone who also fell for that trap in the past, I hope to help show folks what I was shown, and help avoid seeing someone else chasing smoke.

 

[Varminterror]

There is a rather obvious "analog" trend of these groups to group better at lower powder charges.

I'm not intending to come across as contentious or snarky in any way, but I look at data for a living, and I absolutely LOVE helping other shooters improve their load development methods - and I absolutely HATE seeing other shooters spinning their wheels and chasing smoke, just because something we all did the wrong way for a long time is what we all did the wrong way for a long time...

The human eye WANTS to find patterns, and we find ourselves justifying them however we can - we ignore flyers even when we don't truly call them during a string, because we WANT our groups to be smaller...

For example, we're seeing that the rightmost load, claimed to be the heaviest charge, shoots smaller than the 3rd or 4th charge weights, but your eye WANTS a trend where they shoot smaller for lesser charges... And as I've described above, knowing what we know about statistical distribution of group sizes, it's highly likely that we don't actually have any statistically substantiated difference between the 5 charge weights for the top 2 strings at all.

We can also make a general observation that the higher charge weights have a more consistent vertical POI than the lower charge weights - for example, the top row left and 2nd from left shift by what appears to be approximately a half inch, which suggests instability by known load development methods...

Of course, we also don't know whether these steps are 0.1grn, 0.015grn, or 1.0grn, but we're WANTING to make observation and interpretations, and WANTING to let our minds fill in gaps with assumptions and wishes.

If we're wanting to save time and primers, I'd want more than coincidence to convince me to chase MORE rounds down any rabbit holes. And with a goal of 0.3" 5 shot groups, the data here suggests we scrap these combinations entirely, because none of them are 0.3" anyway...

 

[FROGO207]

The two words the OP used say a LOT. Every firearm will have one or several bullets/weights that will work best. No two firearms will have identical results as a rule. Some bullet shapes and compositions are inherently better than all others in a certain caliber. But if you bank on it being so in your gun you might be dissapointed. Of course there are other factors that affect accuracy but it comes back to a good bullet selection can make or break a good load. Years of testing have proven this out to me.

 

[maxxhavoc]

One reason "similar POI" focused methods like the Audette Ladder, the Newberry OCW, or Satterlee's rendition of the Audette Velocity curve have been proven to WORK is because they do NOT try to prove groups are "different," but rather these methods look for indication that groups are NOT different. There's SOME invalidity in the analysis as to the ANOVA between what we assign as "anti-nodes" versus what we believe is in the node (the non-differentiated portion of the ladder), but overall, ladder tests aren't looking for what is different - not looking for the small group among larger groups - it's looking for NON-differentiated results. Proving a null hypothesis is much, much easier than proving a differentiation. So in these methods, we are NOT looking for the ONE small group, and we're NOT chasing our tail by jumping up and down because one random group was coincidentally smaller than the others - but rather, we're looking for a cluster of results with relative insensitivity - where the change didn't appear to make a difference.

So basically, you would:
-Shoot 10 Foulers
-Shoot 10 Groups of 3 shots each (at 200+?)- I assume:
-wait x minutes/seconds between shots
-alternate charge weights beween shots
-reset rifle position between shots
-Plot the POI for all shots
-Look for a mean (median?) POI and narrow down to the groups that are closest to the calculated POI
-(would you use actual distance from POI for this in both x and y, or plot the mean for x and y separately and use each to create a "composite" mean?)
-use the load that most consistently groups near the poi.

Something like this?

Usually I combine the OP's method and a velocity-based ladders. Usually my best "eyeball" groups hit within velocity plateaus. If they do, I use the average load of the plateau. I find that method helps minimize variances such as temperature/headstamps/I'm getting old/etc...

I'm limited to 100 yards, so velocity seems to me to be a better method for me. Most of the POI stuff I have seen say there is not enough variance at 100. I'm interested in more advanced methods of load development, but I do not look for perfection, I am not in competition, am on a 100 yard range, often on a somewhat rickety bench, with an entire rifle setup that probably cost less than your scope.

 

[thump_rrr]

I'm not intending to come across as contentious or snarky in any way, but I look at data for a living, and I absolutely LOVE helping other shooters improve their load development methods - and I absolutely HATE seeing other shooters spinning their wheels and chasing smoke, just because something we all did the wrong way for a long time is what we all did the wrong way for a long time...

The human eye WANTS to find patterns, and we find ourselves justifying them however we can - we ignore flyers even when we don't truly call them during a string, because we WANT our groups to be smaller...

For example, we're seeing that the rightmost load, claimed to be the heaviest charge, shoots smaller than the 3rd or 4th charge weights, but your eye WANTS a trend where they shoot smaller for lesser charges... And as I've described above, knowing what we know about statistical distribution of group sizes, it's highly likely that we don't actually have any statistically substantiated difference between the 5 charge weights for the top 2 strings at all.

We can also make a general observation that the higher charge weights have a more consistent vertical POI than the lower charge weights - for example, the top row left and 2nd from left shift by what appears to be approximately a half inch, which suggests instability by known load development methods...

Of course, we also don't know whether these steps are 0.1grn, 0.015grn, or 1.0grn, but we're WANTING to make observation and interpretations, and WANTING to let our minds fill in gaps with assumptions and wishes.

If we're wanting to save time and primers, I'd want more than coincidence to convince me to chase MORE rounds down any rabbit holes. And with a goal of 0.3" 5 shot groups, the data here suggests we scrap these combinations entirely, because none of them are 0.3" anyway...

Why don't we flip the script here and explain how you would you do it if you had 100 rounds to get to a load from start to finish.
I have a rifle that's giving me fits.

 

[AJC1]

Why don't we flip the script here and explain how you would you do it if you had 100 rounds to get to a load from start to finish.
I have a rifle that's giving me fits.

For the sake of not repeating his effort, a thread with his methodology would be good for easy reference later.... besides then I can book mark it to try myself.... I'm absolutely humble enough...

 

[jmorris]

I agree that bullets often matter more than what gun, shooter, press, powder, dies, measure, runout, ES/SD, primer, case, moonphase but sometimes not.

 

[Varminterror]

how you would you do it if you had 100 rounds to get to a load from start to finish.

As I described in this thread (and have described many times over on this site):

For my PRS rifles, typically my load development is ~40 rounds for a new bullet and cartridge. 10 to foul, 3 shot groups for 10 weights, and I find regions which are NOT differentiated, and that's my load. They shoot small, they shoot stable, and the load is forgiving.

Those 10 weights are 0.2-0.3grn steps, fired round robin.

Jason Baney describes his version of the Creighton Audette Ladder test on this webpage, which is effectively what I do, with the following modifications described below:

6mmBR.com LR Load Development

How I differ from this article is functional facilitation: I often am not so diligent as to paint my bullets, so I'll shoot separate POA's instead of a single POA, in the same way does the Dan Newberry OCW test. Audette's original articles also included chronograph velocity curves as well to corroborate the results on target, so I do include the chronograph result as well, which is largely attributed to Scott Satterlee these days - but was published by Audette before Scott was even born. I prefer to shoot these tests at 600yrds, with a minimum comfortable distance of 300yrds - but when I find myself stuck on a 100yrd range, I have found that the velocity curves give better indication of what WOULD happen at 600yrds than do the POI shifts on the 100yrds target. Sometimes the velocity curve agrees with the 100yrd target, sometimes not, but the velocity curve has ALWAYS agreed with my 600yrd targets, so I know it's the 100yrd targets telling lies, not the chronograph. (Equally, for rifles which don't shoot as small, or for shooters which don't shoot small, evaluating huge groups at 600yrds isn't fruitful, but weak precision doesn't trick a chronograph, so I've helped a lot of new shooters find their loads by corroborating both halves of the Audette method - POI on target and speeds on the chrony). Again, make no mistake, I'm not looking for groups which are smaller than others or just looking for low ES/SD for a given charge weight, and I'm not looking to avoid groups which are larger than others, I'm ONLY looking for the forgiving windows where potential influences - expected influences - are proving to NOT be influential. I'm ONLY looking for areas where the loads are NOT responding to changes in charge weight.

I used to do relatively extensive seating depth testing in the past, but honestly, I haven't found a statistically differentiated influence for most of the bullets I choose to shoot (and I choose to NOT shoot bullets which are jump sensitive) which is significant for the sports I shoot.

I do not do any of this, and didn’t make any insinuation that I do:

-Plot the POI for all shots
-Look for a mean (median?) POI and narrow down to the groups that are closest to the calculated POI
-(would you use actual distance from POI for this in both x and y, or plot the mean for x and y separately and use each to create a "composite" mean?)
-use the load that most consistently groups near the poi.

As I stated in this thread:

I'd never say anyone SHOULD do ANOVA or T tests on their load development targets

So I shoot my ladders and by the time I'm driving home, I know which loads are in the node, and I know where I need to be loading. I'm not doing any complex mathematical analysis, and in fact, I'm actively avoiding the need to do so - I'm looking for very, very simple patterns where the loads nearest one another are NOT different, such analysis of variance or any other statistical analysis simply isn't necessary... As has been my point in entirety throughout this thread - if we have to try REALLY hard to prove that one group is .42" because and only because it is a different charge weight than the .56" group beside it, and it can NOT be smaller because of random distribution of statistically certain outcomes, then we're chasing smoke.

I do NOT bring up statistical analysis in these threads to imply it is necessary to prove one load better than another, but rather, I only bring up the statistical analysis in threads like this just to encourage folks to question their own observations when they are quick to believe one small group is meaningful - trying to encourage folks to question themselves as to whether they actually can trust that the difference they think they see is really a difference, or if it's just coincidence. If we have to do in-depth statistical analysis to even determine if a group size is meaningful or not (which we do for a test like the OP's), then the null set is likely the true result - that there really is no difference. Occam's Razor is often misquoted - we simplify it as "the simplest answer is most often true," but the saying was actually "entities should not be multiplied more than necessary," as a manner of saying - if we have to dig and scratch to find ways to validate a belief (a hypothesis), then it's probably not actually valid.

@Nature Boy and @JFrank have great threads going on this site which describe their iterations of relatively the same method for load development. I've shared my own targets and velocity curves to describe this process here numerous times as well. Working from a basis of "try a ton of combinations and pray something kinda shoots smaller than the rest," isn't a productive methodology. Grab a proven bullet, a proven powder, good/consistent brass, and a good barrel - make it shoot small by running through these proven methods, and it will. If it doesn't shoot small enough for your needs, then get a better bullet or better barrel or better smith. Once you reach the mountain top, shooting among the smallest group-shooters in the world, then you're just praying for luck on day of matches so the wind blows your bullets into the group instead of out, and so your finger jerks the bullets into the group instead of out so you come out on top.

 

[JEBruns]

@JEBruns - what's the diameter of the outer circle on these? If these are the USBR 50yrd targets, I think they would be 2.05", is that correct?

Correct.

I'd never say anyone SHOULD do ANOVA or T tests on their load development targets, but confirming that information, I can offer some insight as to why I encourage folks to question whether distinctions on target arrays are actually differences or not. I see a LOT of folks finding confidence in coincidence - and as someone who also fell for that trap in the past, I hope to help show folks what I was shown, and help avoid seeing someone else chasing smoke.

Hey, I'm always willing to learn. But I really don't know what you are saying. I guess I need to research some of the methods you mention at some point. Ladder development might burn more bullets, but has brought me some great (or at least what *I* consider great), consistent, small grouping loads. Plus I bought the gun to shoot. If I burn up a barrel, I get to start over. A good thing!

 

[Charlie98]

Grab a proven bullet, a proven powder, good/consistent brass, and a good barrel - make it shoot small by running through these proven methods, and it will.

In all honesty, I have to agree with this.

My recent tests with my .308 Savage 10TAC, trying to find a load that beats 168grn FGMM... using 3 different bullets, 3 different powders... and too many components. At the end of the day, and something like 150 rounds, none of my loads would beat FGMM. So... if you can't beat them, join them. I used the 168grn SMK (used in FGMM,) and IMR4064 (reportedly used in FGMM currently,) and commercial brass, seated to 2.800" What do you know... I finally beat FGMM's groups.

Thinking back, that's also how I wound up with my original recipe for 6.5CM... I looked at what everyone was shooting (on a different forum) and just copied those component choices.

 

[AJC1]

In all honesty, I have to agree with this.

My recent tests with my .308 Savage 10TAC, trying to find a load that beats 168grn FGMM... using 3 different bullets, 3 different powders... and too many components. At the end of the day, and something like 150 rounds, none of my loads would beat FGMM. So... if you can't beat them, join them. I used the 168grn SMK (used in FGMM,) and IMR4064 (reportedly used in FGMM currently,) and commercial brass, seated to 2.800" What do you know... I finally beat FGMM's groups.

Thinking back, that's also how I wound up with my original recipe for 6.5CM... I looked at what everyone was shooting (on a different forum) and just copied those component choices.

One could say it's the only cost effective way to start competing due to cost unless one has substantial resources and time.... get competitive and then experiment for better results.