Thread: Testing a conventional approach to load development

Originally Posted by No.3

So, just to chuck a variable into the equation that hasn't been factored in - what affect if any would adjusting the seating depth of the projectile deeper in the two loads you've selected for the second round of the testing have?

Noting that you've settled on magazine length for the seating depth which is fair as most would either do that or set their length to the "book length" and forget it, yet seating deeper by a nominal amount is what some pundits recommend as a final step in tuning the load...

Would be a rather interesting final comparison just to check what difference if any this adjustment would have on your data.

If you take a look at Hornardy podcast #50 ( and subsequent investigations) you'll find that seating depth has no influence on accuracy when accuracy is tested to statistically valid samples.

Originally Posted by Tentman

If you take a look at Hornardy podcast #50 ( and subsequent investigations) you'll find that seating depth has no influence on accuracy when accuracy is tested to statistically valid samples.

I'm interested to test for myself but I anticipate no difference.

Thanks - great write up.

Originally Posted by No.3

So, just to chuck a variable into the equation that hasn't been factored in - what affect if any would adjusting the seating depth of the projectile deeper in the two loads you've selected for the second round of the testing have?

Noting that you've settled on magazine length for the seating depth which is fair as most would either do that or set their length to the "book length" and forget it, yet seating deeper by a nominal amount is what some pundits recommend as a final step in tuning the load...

Would be a rather interesting final comparison just to check what difference if any this adjustment would have on your data.

Does not matter, whatever he would change if he did the 20 shot group thing again it would likely print a similarly "enlarged" group, "hiding" any effect any change would have unless it makes things wayy worse (which it might) One could then claim that all those changes do not matter, which is not always true...

Problem when comparing 0.25 to 1.03 MOA groups is that things like shooter fatigue, posture, perfect (and consistent shot to shot) cheeck weld, often have a much larger effect than 2 very accurate loads themselves, especially over 20 shots...

Originally Posted by Tentman

Hmmm STC contributes to the thread but fails to advance the argument. Pity you don't appear to understand how the cancel button works . . . . may I suggest that you check out the thread on its use.

Try to read and understand what I wrote please, and maybe try to understand sarcasm as well.

I'm not a reloader, although I have whakamoled 10 rounds of 270 140g interlocks together and shot a deer with one.
What this thread explains is great advice to show the average hunter that as long as their velocity is fast enough and they get a group under 2 inches at 100m they can save a lot of time and money trying to get things better that are already good enough.
STC could keep telling everyone to try 15 rounds and pick a load from that but this thread shows full facts of speeds and groups that can help people from doubting themselves about whether their load development is good enough.

Originally Posted by Tentman

If you take a look at Hornardy podcast #50 ( and subsequent investigations) you'll find that seating depth has no influence on accuracy when accuracy is tested to statistically valid samples.

I may be wrong, however if my memory serves me correctly they said they didn't observe any noticeable difference with seating depth. however they also said they weren't really looking at seating depth and may well do some testing specifically on that subject.

I'm looking forward for someone to actually test specifically for seating depth in a similar manner to what Hornady (and Gimp) did.

Thanks @gimp for sharing your data & deductions, makes for thought provoking reading.
A couple of observations, #1 comparing the 2 20x shot pictures or patterns, the 43.6 group shows more horizontal dispersion & less vertical dispersion compared to the 44.5gn group & vice versa. At least that’s what pops out to my eyes. Both show some very good shooting skills btw. This brings me 2nd point, statistics & normal distribution….. this assumes a random error or wobble in data with a uniform set of underlying factors. I wonder if we truly have that as shooters, when the nut behind the bolt (the shooter) is as much part of the outcome as the load, pill, powder, barrel, barrel temp, fouling, etc, etc. I know my own shooting skills suffer from ups & downs. I think we’re not always dealing with a random normal distribution in load dev especially when running large numbers of shots.
Like you mention most of us started off learning load dev based on running 3 shot groups at increasing powder increments. My own learnings have lead me to run a ladder first with chosen pill & powder starting at a mid point on published data & proceeding up to book max and if no pressure signs detected then cautiously above book max until pressure signs observed, & record velocity & POI for all shots. Often the POI shift across much of the ladder is minimal, which helps establish likely sweet spot for velocity, safety & accuracy potential.

It would be interesting to test the sensitivity of a load to powder weight variation. In Gimps 3 shot groups ladder there are changes in POI but I wonder if these would pull in in a decent sized sample.

My experience with "accurate rifles" is that their ladder (say 10 shots at 0.3 gn increments) is often one hole at or close to MOA across an "enormous" powder variation. Ofcourse the change in velocity will tell a different tale as range increases beyond 100M etc but that's a different test.

Originally Posted by chainsaw

Thanks @gimp for sharing your data & deductions, makes for thought provoking reading.
A couple of observations, #1 comparing the 2 20x shot pictures or patterns, the 43.6 group shows more horizontal dispersion & less vertical dispersion compared to the 44.5gn group & vice versa. At least that’s what pops out to my eyes. Both show some very good shooting skills btw. This brings me 2nd point, statistics & normal distribution….. this assumes a random error or wobble in data with a uniform set of underlying factors. I wonder if we truly have that as shooters, when the nut behind the bolt (the shooter) is as much part of the outcome as the load, pill, powder, barrel, barrel temp, fouling, etc, etc. I know my own shooting skills suffer from ups & downs. I think we’re not always dealing with a random normal distribution in load dev especially when running large numbers of shots.
Like you mention most of us started off learning load dev based on running 3 shot groups at increasing powder increments. My own learnings have lead me to run a ladder first with chosen pill & powder starting at a mid point on published data & proceeding up to book max and if no pressure signs detected then cautiously above book max until pressure signs observed, & record velocity & POI for all shots. Often the POI shift across much of the ladder is minimal, which helps establish likely sweet spot for velocity, safety & accuracy potential.

Technically I believe the distribution of radii from shots forms a Rayleigh distribution rather than normal but central limit theorem still applies.

With the approach of shooting alternately between each load, I'm confident enough that shooter error is held consistent between my 2 20rd groups shown, while it may be additive to the dispersion it will be additive in equal measure, thus the results are comparable. The point however is valid which is why I refer to the "rifle system" - this means the rifle, shooter, sight and ammunition as a unit. Any change to a part of the system may change results.

"Does shooter fatigue contribute to larger groups" is a good question to go away and test empirically, if anyone is so inclined

I suspect the only way this could be improved is to remove variables, i.e. taking the action and barrel out of the human handle and bolting it to a big chunk of steel on a slide setup with a positive return to battery to reset after each shot and some form of mechanical device to operate the trigger to take the finger out of the equation.

That's going more to the proof barrel idea though, and I have heard of random examples where people have bolted rifles into sleds and got one hole groups and expected the same when the setup is hand held and been hugely disappointed that no human can shoot the thing near as accurately. Put down to the strapping into the sled doing something to the bedding or other variables that human hands and shoulders can't... It's getting outside of the original testing hypothesis though.

As far as does shooter fatigue contribute to larger groups - yes. It's been proven several times over and there are several sources of fatigue that do it - eye fatigue from not blinking enough or focusing too long or optics that aren't set quite right for your eye, muscle fatigue, cumulative effects of recoil and blast, excitement and/or pressure on weird bits of the body from lying over something awkwardly all play a part as does nutrition and hydration. I can't shoot for nothing when I need to take a piss for one!

Originally Posted by No.3

I suspect the only way this could be improved is to remove variables, i.e. taking the action and barrel out of the human handle and bolting it to a big chunk of steel on a slide setup with a positive return to battery to reset after each shot and some form of mechanical device to operate the trigger to take the finger out of the equation.

That's going more to the proof barrel idea though, and I have heard of random examples where people have bolted rifles into sleds and got one hole groups and expected the same when the setup is hand held and been hugely disappointed that no human can shoot the thing near as accurately. Put down to the strapping into the sled doing something to the bedding or other variables that human hands and shoulders can't... It's getting outside of the original testing hypothesis though.

As far as does shooter fatigue contribute to larger groups - yes. It's been proven several times over and there are several sources of fatigue that do it - eye fatigue from not blinking enough or focusing too long or optics that aren't set quite right for your eye, muscle fatigue, cumulative effects of recoil and blast, excitement and/or pressure on weird bits of the body from lying over something awkwardly all play a part as does nutrition and hydration. I can't shoot for nothing when I need to take a piss for one!

Other factors that have not been taken into account are warming up of the barrel, wind, temperature changes, parallax...

Shooting a 20 shot group to try and see a difference between 0.25 and 1 MOA essentially covers what you want to see in a shit ton of noise.

Right so I do believe there is one other test that hasn't been done that would answer more questions than most others and allow some folk to rest easy and be happy with what they are using. RUMPY has said it well above,others along same lines. The how far away was last deer poll was eye opening. Bottom line is this. Good enough is good enough. It's a personal thing how good you decide is actually enough. A deers vitals are generally considered a 8" vital area. If we halve that to try and remove some of the environmental and human factors we get 4" as ethical must be able to consistently get group size for given range....which by default tells us our own personal ethical range EG when you can't hold 4" or better you need to get closer. Your rifle n load might be good for 1" on a bench rest on bluebird day out to two hundred yards (oh the skills needed) but on steep slippery slope in howling norwester you may struggle to achieve that 4" freehand at 50 yards. So how about let's cut the crap and try one more test???? Same rifle,same shooter,same conditions and use every combination of run of the mill loads you have got access to. Cartridge doesn't really matter,the 223 is possibly cheapest beast to do this with. I may have been nice to old ladies in previous life but at hundred yards my centerfire rifles don't change enough between yards to matter at or under hundy.

Why bother??? To show that maybe all the fuss n bother isn't NEEDED

Greetings @gimp and all,
A nice piece of investigation with some solid conclusions. To me the early validation of load data using velocity is an important part of the process. Velocity is also a good indicator of pressure. The really telling part of your results was the minimal difference in mean radius between the two loads over 10 shots each.
Regards Grandpamac.

Originally Posted by STC

Other factors that have not been taken into account are warming up of the barrel, wind, temperature changes, parallax...

Shooting a 20 shot group to try and see a difference between 0.25 and 1 MOA essentially covers what you want to see in a shit ton of noise.

4x 5rd groups actually.


Barrel cooled between 5rd groups (even though it doesn't actually matter), fired each load alternately to minimise any differences in fouling etc. Fired 10rd of each then went and did other stuff (helping others sort guns) for a couple of hours then fired the last 10 of each. Wind very low and consistent, <1ms 12 o'clock. Temperature consistent at 12 degrees C through the day (kestrel). Parallax dialled out.

Shooting high numbers of shots through other rifles with different loads clearly demonstrates differences in precision, e.g. .223 with handloads 0.9moa group for 20, 1.5moa for 20 with factory "match" ammunition.


How do you propose to quantify a valid understanding of the real precision capability of a system?