Shot my first handloads today, ladder test question?

[Onewolf]

I shot my first handloaded rounds this morning at the gun range. These rounds are for my Savage Long Range Precision .260 rem. I have only been shooting a rifle for about 3 months so I am also learning how to shoot a rifle accurately. My first longer term goal is to qualify to shoot with the longe range group at the local gun club (field match shooting (prone/offhand), not bench-rest).

The first 20 rounds I made were 'test of concept/process' rounds. These were all Hornady A-MAX 140gr with 37.7 gr of Alliant RL-17 powder (92% of max load). At 100 yds my LRP 260 cold bore shot is always 1.25" left and 0.5" high but I forgot to rezero my scope from its last range trip so it was set to 300m range and therefore the cold bore shot (top left corner) was left 1.25" (expected) and about 2" 'high'.

My next group #1 was where I was relearning how to shoot the LRP 260 (after shooting AR, Mosin, and Marlin 336 last time at the range). Groups #2, #3 were not bad (less than 3/4") and my last group #4 was the best group I have ever shot (~ 1/3").



Next I did two powder load ladder tests at 300 meters using Hodgdon H-4350. One with the Hornady A-MAX 140 and one with the Barnes MatchBurner 140 bullets. The manufacturer spec BCs of these bullets are almost identical (0.585, 0.586) and being the same weight I used the same powder loads for both tests. I made two rounds for each load weight for each bullet type. The 7 powder loads I used were 40.5, 40.8, 41.1, 41.4, 41.7, 42.0, and 42.3. Therefore I had 2x2x7 = 28 rounds total. I also made 6 of the 40.5gr MatchBurner 140s for sighting/zeroing at 300m.



Because there were 3 other people shooting/bouncing the bench while I was trying to perform the ladder tests I knew that a couple of the shots were 'not good' and based upon the small overall group size I wasn't sure if there would be any meaningful data produced.

Last night I made two beautiful ladder test targets with black sharpies and rulers. Of course I left them at home this morning when I went to the gun range :banghead: At the range I used a ball point pen to scribble target diamonds on the back of two of my normal paper targets. Nice....

My ignorant newbie analysis of the ladder test targets is that there are no discernable 'nodes'. The only pattern that jumps out is that both of the 42.0gr loads were very consistent.





Do any of the more experienced guys here see any discernable pattern/nodes?

All in all, it was a beautiful 72 degree sunny day at the range and I got to shoot a Barrett 50BMG as a bonus. Not bad for a January 12 trip to the range.

Link to my Barrett 50BMG video (note there is _a_ bad word (not from me) http://www.youtube.com/watch?v=ePl52db5PLo

 

[Trent]

I'm not sure what you were trying to accomplish with the second test (20 rounds, 2 different bullet types, 10 different loads of each).

Different powder charges cause the bullet to leave the muzzle at different velocities, which will cause groups to be delivered to different points. Velocity is a contributing factor to bullet upset as it leaves the barrel.

E.g. If I load and shoot the same bullet, 5 shots each, with 2 different powder charges giving a velocity of 100fps difference, both the x AND y axis of the resulting groups will be affected. The faster rounds will tend to hit higher, but there is also a noticeable difference in horizontal impact that can ONLY be explained by two phenomena.

The first of these is spin drift. A bullet leaving at higher velocity will have more centripetal force will cause the bullet to drift naturally in the direction of the twist (and also has other interesting side effects when braced with a cross wind causing the bullet to climb in to a wind, or drop away from a wind, as appropriate for the conditions)

The second condition is gas pressure. Bullets leaving at different velocities remain in the "zone" in front of the muzzle where gas pressure is still acting on the bullet for different lengths of time. The bullets are still accelerating AFTER leaving the muzzle for a noticeable distance (8" or so), and can be be tipped by this force depending on the purity of the rifle crown, the purity of the base of the bullet AFTER the high pressure gasses have had their way with it, and whether the bullet left the muzzle tipped or straight. Bullets leaving at the same velocity will tend to have the same amount (and direction) of forces exerted on them, compared to bullets leaving at a different velocity.

Combining these factors the only way to establish any data of substance for statistical analysis on the inherent accuracy of a given load, you need to shoot relatively large group sizes. (I use 5 shots when working up loads, then subsequent 10 shot groups on loads which show "promise", to validate them.)

The ONLY time I shoot strings of independently loaded cartridges is when I'm searching for max pressure signs. Then I'll load up +1 grain increments up to about 2.5% of max load, and then 5 loads approaching max. (This reduces the amount of bullets you have to "pull" if you find that you're getting max pressure signs prior to hitting max).

The only time I push past maximum is if a group on the upper end of the spectrum shows accuracy promise AND I'm not showing overpressure signs. At that point I might continue pushing beyond max at .5% increments with a small string of individually loaded rounds, find a "true" safe maximum, and then shoot a 10 shot group at that point to compare.

After getting the "promising" groups nailed down to "confirmed" I'll experiment along the side of those in .1 grain increments with additional 10 shot groups to see if I can hone in on something slightly more perfect.

 

[NeuseRvrRat]

in my experience, you need more like 700m to get meaningful data from a true ladder test.

try this instead: http://optimalchargeweight.embarqspace.com/

 

[Onewolf]

The first "test" was simply to see if I had made (handloaded) rounds that were going to go bang and not "blow up" in my face since they were the first rounds I had ever handloaded. The fact that they shot decent 3 shot groups was just a bonus.

The 'second' test was my first attempt at a powder load ladder test. I realize that normally you would test one bullet type at a time but since I am a brand new handloader I was curious to also load/test with the second (nearly identical in external ballistics) bullet as an independent test.

If you read my text you will see that each independent ladder test was done with 14 rounds (7 different loads with 2 rounds of each load). I had also 6 additional identical rounds of the lowest powder load to zero at 300 meters.

I assume you are familiar with the ladder test methodology (http://www.6mmbr.com/laddertest.html) so I will not get into the details other than to say you look for clusters/nodes of hits at different elevations. I don't see any discernible clusters/nodes in the two ladder tests I performed. This may have been caused by poor shooting technique or invalid test process.

I am seeking assistance on analyzing these ladder tests and advice for the next ladder tests I perform.

Thanks are any/all comments and suggestions.

Doug

 

[Onewolf]

in my experience, you need more like 700m to get meaningful data from a true ladder test.

try this instead: http://optimalchargeweight.embarqspace.com/

Thanks, I have heard that the farther you shoot the ladder test the more valid the data, but I am currently limited to 300 meters.

I will definitely investigate the OCW method as well.

 

[Trent]

Onewolf,

I understand the premise of ladder testing, but I don't agree with it. This being said, there are many ways to find the nugget of truth in loading and whatever process is fun and fruitful for you, is fine by me.

Today I took my 300 Win Mag out. I was doing 1 grain increments of H4831SC with 210 grain Berger VLD bullets, starting at 65gr and working up to 69gr, using 10 shot groups. The conditions were very poor, bottomed at 34 degrees F, intermittent and annoyingly unpredictable crosswind of 15mph. I was underdressed for the rapid drop in temp (20 degrees in 2 hours), and it took it's toll on me (numb hands, slight tremors in the extremities indicating early signs of hypothermia).

Anyway, about as bad of conditions as one could go out an test loads in. But enough of that; Back to the nugget of truth.

I reset my scope based on some basic mathematics off my last shoot; was zeroed at 300 yards, for 220 grain matchking bullets travelling at 2600 fps (74gr H1000). Pleasantly, my first shot of the lighter 210 grain bullet was on the top edge of the 1" bullseye at 100 yards. My math worked.

Anyway the first two 10 shot groups were 1.690" and 1.558" extreme spread. Horizontal max was 1.690" and 1.204", vertical spread was 1.175" and 1.245" respectively. Not exactly earth shattering groups, but consistent.

The next group started out with the first three shots making a ragged hole. I remember thinking "man wouldn't this be great if I just stumbled on the magic load in the first round of testing!?".... the fourth round kept the one ragged hole going. But unfortunately things went down hill. That 3rd group at 67.0 gr ended up at 1.580" extreme spread, 1.594" horizontal max, 1.358" vertical max.

(At this point it's worth noting that my vertical spread is showing consistently smaller than horizontal; a lot of that was the 15mph off and on wind I was fighting - no flags out and a frozen ground with no plant life waving around, so I couldn't read it well; well, I couldn't read it at ALL, to be honest.)

That 3rd group, if I'd stopped at the first 3 rounds, would have been a whopping 0.110".

But 3 round groups don't tell the whole story. Take any 10 shot group you shoot and circle 3 of the closest holes - any three, just as long as they're close.

Now imagine those are the ONLY shots you fired.

Your impression of reality would not match the reality to be found.

The 4th group ended up being the exact opposite behavior of the 3rd group. I started out with three shots forming a triangle (actually set my extreme spread on the first TWO shots). It ended up at a whopping 1.898", with a horizontal deflection max of 1.670 and a vertical max of 1.562.

What's interesting about that 4th group though is the last SIX shots formed a tight little cluster that measures only .220" from center of impact to center of impact, at the extreme spread.

Had I only shot THOSE six shots, I would have thought I'd found the "magic load"... and probably chased my tail for several subsequent shoots wondering why I couldn't reproduce it! I would have been blaming my scope mounts, or my seating die, or inconsistent bullets.. ANYTHING except .. statistics.

By the time I got to the fifth and sixth groups I was becoming an ice cube. Those were 1.482" and 1.938" extreme spread, 1.248" and 1.910" horizontal spread, and 1.562 and .816 vertical spread, respectively. I actually shot my best horizontal and vertical spreads of the day, and the best group, while shivering in a wind that was growing from an occasional 15mph stiff breeze to a blasting 20mph sleet/snowstorm.

Now, a more useful foundation is to analyze the standard deviation of the bullets; when group size itself is inconclusive or I see a "tight core" with a couple of shots to the extreme that upset it, I plot each impact out in X,Y coordinates and determine the zone where the majority impact. (this is also useful for obtaining a true "zero".)

Once I find a highly consistent load (particularly in vertical deviation, which is the important one from a ballistic perception, as horizontal deflection tends to be missing wind cues), I will shoot a series of groups and plot them overlaid to determine where the true "center" is, and what percentage of rounds land within X" radius of the center. Such a composite group of say, 100 shots, might be 5" at 300 yards, but if 75% of the bullets are landing within 1" of the center, then I'm getting closer to "truth".

In the end, it's a statistics game. What are the chances of shooting a 10 shot 1" group at 300 yards? With my last barrel on this gun, it was about 80% - that is, 80% of the time, I could put down a sub-1" 10 shot group at 300 yards; my average group size over 10 strings of 10 was 0.92". So far, I haven't got that level of accuracy with this rifle on the new barrel. But it took 12 years to get to that point with the last barrel.

I don't believe in shortcuts. I don't believe 3 shot groups can prove anything whatsoever. I believe overlaying 33 3-shot groups together in one graph WILL prove something, as will overlaying 10, 10-shot groups together on one graph. It's a numbers game - you need sufficient data to form real conclusions.

Lacking sufficient data, you can form no real conclusions. You cannot form conclusions shooting single rounds or 3 round groups or 5 round groups or even 10 round groups. It's not until you overlay that data in to even larger groups together that you can begin to form - with any degree of certainty - any real conclusions about where the next round you shoot is going to land.

As you hone in on the best load for your rifle, there's no substitute for data.

Not trying to sell my way of doing things; this is my way of doing things, what I've found that works for me. But as a new reloader I'd strongly encourage you to not try to shortcut the process of collecting data, because the end result can be quite frustrating. You don't want to shoot a perfect group where 3 shots go through the same hole, believe you have found the ultimate load, only to find yourself blaming loose scope screws or gremlins for the next year as you are unable to reproduce it under any conditions, fair or foul.

Just my .02.

 

[Trent]

Composite of all 60 rounds. I plotted the "center" of each group, and then by simply aligning that center to the center of a blank target, I marked the center of each impact. (Each group was elevated from the prior one due to the increasing powder charge from the previous group, centering is required to get a composite view)



22 shots were within a .5" radius. (36.7%)
44 shots were within a .75" radius. (73.3%)
57 shots were within a 1" radius (95%)

I can compare this powder test with future tests. Different bullet type/weight, different powders, etc. From this test, despite the horrible conditions and shot off a bipod on a concrete bench, I can tell that the Berger 210 gr bullets fly pretty true.

The goal of all of this is to determine the consistency of the powder (how consistent the burn rate is for my barrel), the consistency of the pojectile, and the combination thereof. By comparing future tests with different powder against the same projectile I can exclude the projectile and have meaningful statistics for the powder. By comparing the same powder with a different projectile, I can exclude the powder and have meaningful comparators for the projectile.

Some deductive reasoning as the statistics are collected will allow me to determine powder preference, bullet weight preference, and so on.

Because, in the end, I want to find high probability hits. Maximum likelihood that the single bullet fired will fly true to the mark.

Group size, and statistics, are a mean to the end, but utterly useless unless records are kept for comparison. If you don't collect the data, organize and analyze the data, then some of those "truths" might remain hidden.

 

[kelbro]

You get better data performing ladder tests when the wind is not a factor.

 

[WNTFW]

"I have only been shooting a rifle for about 3 months so I am also learning how to shoot a rifle accurately."
That is why I asked about your shooting skills in the other thread. You need to pick a load and let your groups shrink from gaining skill. You also have to make the best decision you can from what you have.

42.0 & Barnes is what I would do for your next range trip. Do 10, 20 or 50 and shoot a good bit. Load up exactly the same & just shoot.

Also in the other threads a few guys mentioned Sierras. I have had a few guys that know their stuff tell me you can get a Sierra to shoot easier than just about anything. But the Barnes are proving themselves in the limited test you have so far.

You really need to understand the OCW method a little bit when trying it. Or at least put some faith in it & try it as per instruction. OCW descended from the ladder to be an improved ladder.

.22lr is also a great way to practice. Picnic tables aren't shooting benches either. I'd go prone before using the table in the .50 video.

 

[helotaxi]

The first of these is spin drift. A bullet leaving at higher velocity will have more centripetal force will cause the bullet to drift naturally in the direction of the twist (and also has other interesting side effects when braced with a cross wind causing the bullet to climb in to a wind, or drop away from a wind, as appropriate for the conditions)

Firstly, spin drift is negligible at the range that the OP is discussing. Second the difference in 100fps in a fast twist barrel which exhibits more spin drift than a slower twist barrel works out to less than 0.5" at 1000 yds. IOW spin drift has no place in this discussion because is isn't a contributor to shot-to-shot variance. The wind-spin relationship could be at close range if we were talking about small fractions of an inch that a benchrest shooter would care about and the wind was quite variable but outside 1-200 yds and benchrest sized groups, it's pretty much lost in the noise of normal variance.

Secondly, barrel harmonics aren't a simple vertical oscillation. There is often a horizontal component to it as well.

Finally, in all of your statistical analysis, I noticed something that was missing: chronograph data. No real analysis is complete without knowing what the real consistency is with velocity. It's really difficult to develop a onlg range round without knowing whether the vertical variance that you're seeing on paper is due to barrel harmonics, bullet inconsistencies or velocity variance. With a good barrel and a good bullet, it's not difficult to find a load that shoots really well at 100yds but is terrible at 600 because the velocity is inconsistent.

 

[Trent]

Spin drift is a factor at any range, the only difference is the amount. At 100 yards, depending on the rifle and twist rate, I'll see 10 shot group centers move between 0.050-0.125" (measured) with each incremental velocity increase of 100fps (I measured this same amount again yesterday on the groups I shot). This is remarkably consistent, and continues as the range (time of flight) increases. Point of impact on the 210 gr bullets I was shooting moved a lot more than the horizontal component; but both were measurable (and scaled with velocity).

But I agree that'll tend to get lost shooting small groups (or singular rounds) in the noise of normal shooting at a set distance of 300 yards.

Chronograph data isn't something I collect when it's 34 degrees and sleeting. Any more, with the amount of brass prep I go through, collecting chronograph data is not as important as it once was. It was useful when determining to what extent each individual stage of brass prep contributes, it was useful when determining how consistent powders burn, but after 15 years I've got a pretty solid understanding of what's going on with the components I use. The consistency of my prep work will allow me to observe any velocity issues first hand, when moving from 100, to 200, to 300, to 440 yards (the ranges I have available close to home).

Sure, if I was shooting at only one distance, they'd be hard to pick up on, but I don't just shoot a single distance.

I don't agree with the OCW method, mentioned specifically in the OP.

#1 The sample sizes are too small
#2 There's no control group for comparison
#3 It ignores the scientific method

Experimenting hasn't changed much. You get a control established. You change ONE thing, and only one thing, repeat the experiment, observe the differences, and record them.

The OP at least made a control; the two 4-shot groups of 40.5 gr.

But even assuming ideal conditions - an exceptionally fine shooting .5 MOA accuracy rifle, with each shot sent down range done so solid, by an experienced shooter with no aiming error, each round is going to impact somewhere, and quite randomly, within a roughly 1.5" circle.

Conducting 6 additional one shot experiments, with each round subsequently landing in it's own 3" circle of random distribution, with each load (in that relatively narrow power band he's testing) giving approximately another 10fps shifting the point of impact roughly .2" (according to calculations).

I would argue that the 0.2", compared to a best case margin of error of 1.5", is going to be lost in the noise even MORE than the spin drift you were so quick to dismiss.

 

[helotaxi]

I don't agree with the OCW method, mentioned specifically in the OP.

#1 The sample sizes are too small
#2 There's no control group for comparison
#3 It ignores the scientific method

Experimenting hasn't changed much. You get a control established. You change ONE thing, and only one thing, repeat the experiment, observe the differences, and record them.

The OP at least made a control; the two 4-shot groups of 40.5 gr.

But even assuming ideal conditions - an exceptionally fine shooting .5 MOA accuracy rifle, with each shot sent down range done so solid, by an experienced shooter with no aiming error, each round is going to impact somewhere, and quite randomly, within a roughly 1.5" circle.

Conducting 6 additional one shot experiments, with each round subsequently landing in it's own 3" circle of random distribution, with each load (in that relatively narrow power band he's testing) giving approximately another 10fps shifting the point of impact roughly .2" (according to calculations).

I would argue that the 0.2", compared to a best case margin of error of 1.5", is going to be lost in the noise even MORE than the spin drift you were so quick to dismiss.

If you actually read up on the OCW method, it is very scientific. The sample size is as big or small as you need it to be to get satisfactory results. The size of the groups shot when doing initial OCW testing is irrelevant. All you care about initially is the center of the group.

Spin drift is perfectly deterministic and can be calculated. It is also exactly the same, shot to shot. Therefore from a grouping standpoint it can be completely ignored. With a 1:8 twist and a good .264cal bullet fired at 2700fps spin drift is 0.06" at 100yds. That is the definition of insignificant. It stays below 1" out to past 400yds and is less than 7.5" at 1K. But again it is the same from shot to shot and thus will not affect group sizes.

 

[Trent]

If using 3 shot groups, sure OCW is scientific (to a point - 3 shot groups themselves don't tell the whole story).

If using 2 shot sample sizes, it tells you very, very little.

If using 1 shot sample sizes, it tells you absolutely nothing. This is even more true if your inherent margin of error is much larger than the increments you are testing.

The OP established a baseline in his first target showing his rifle (and him) shoots about 1 MOA on the two three shot groups at 100 yards. His subsequent 2 shot groups at 300 yards continue to bear out this trend, from what I can tell.

With a margin of error of 3", shooting small sample sizes (1, 2, or 3) can reveal any NUMBER of falsehoods.

For instance, if I were to use a sample size of 2 as the OP did, with the stats I collected this weekend on the last test, using this powder and bullet combination, 18% of two-shot groups I fire will show sub .5 MOA results (approx 1 out of 5 groups. This would be an inaccuracy introduced in the test process, because the sample size is too small.

That's scientific - but not good science.

The next time he repeats the test and obtains a different result, what's he to blame when it doesn't work out the same? Mirage? Spec of dust in his eye? A slight gust of wind that didn't register on the flags?

Now if he repeats the same exact test with 3 more shots, he's approaching a usable sample size. If he does so twice or three times, now he's getting somewhere, so long as the results are combined to form a larger sample. (You don't need to shoot 10 shot groups all on the same setting. You can go out twice with identical loads and shoot 5 shot groups, overlaying the results of day#1 and Day #2 together on a clean target to form a 10 shot group).

Why 10 shots?

Statistical smoothing. Nothing more. Each result contributes 10% to the overall view. An error of 10% induced on a particular sample contributes only .1% to the overall result. Whereas with 2 shot groups each result contributes 50% to the overall view. An error of 10% magnitude on a single sample contributes an error of 5% to the overall sample. (And with inherent inaccuracies of rifles, even a good shooting rifle / rifleman of 0.3 MOA, that's still a big deal at any real distance.)

18% of the time even with a BAD load of 1.2 MOA, as I showed above, I will see two bullet holes touch each other at 100 yards, with the distribution groups I posted above. This would give me the false impression at 100 yards that I was shooting a maximum of .3" groups, one out of five times!

The OP did 14 such samples. How many of those are "true"?

In my case, I can tell you that a full 1/5 of those would be invalidated right from the get go, because the sample sizes aren't enough to iron out the inherent randomness.

Now, the question is, WHICH 1/5 of his results are invalid? He has no idea, you have no idea, I have no idea.

So the entire experiment is invalid.

 

[helotaxi]

You're looking at group sizes, which with OCW load development is incorrect. You do not care about it when finding the stable range of charge weights. Since you're focusing on the wrong thing, your understanding of what is statistically valid is skewed. All you're doing in the initial incremental charge testing of OCW is finding a series of charge weights that shoot to the same POI based on the group center. It doesn't matter if the groups are bigger that you want, you fix that later by experimenting with seating depth. All you're doing shooting 10shot groups during load development is burning up components.

 

[Trent]

Assume that you are shooting under perfect conditions with a barrel capable of 0.5 MOA in a machine rest, no human error introduced, with an ammunition capable of a 95% standard deviation of velocity at +/- 5fps.

Sighted for 100 yards, the .264 bullet has a predicted point of impact at 300 as follows:

2655 fps: -18.20"
2600 fps: -18.13"
2605 fps: -18.05"

On TOP of the 0.75" inherent in the 0.5 MOA barrel. Which brings us to an extreme spread of -0.82" to + .83" high (1.65" vertical spread).


Now, you shoot X strings, increasing powder charge slightly with each load, giving, say 20fps increases in velocity with each load.

2620 fps @300 = -17.84. Point of impact (average) will move only .29" up on the target. Assume brass prep is stable and we retain a +/- 5 fps 95% standard deviation. Our new +/- is:

2615 fps @300 -17.91 (-.07)
2625 fps @300 -17.76 (+.08)

Which gives us (again) a spread of 1.65".

Assuming even shot distribution in each oval of error, since the ovals overlap due to the small component increases and relatively minimal velocity adjustments, rounds will land in a "convergence zone" about 70% of the time. (In fact, they will "cross each other" VERY often).

Here's a convergence zone chart plotted for a VERY accurate 0.3 MOA rifle shot under perfect conditions with a load increment increasing velocity a measured 0.200" from one group to another.



Note that even the SLIGHTEST increase in vertical skew due to velocity differences on the loaded ammo (even as little as +/- 5 FPS) will result in elongating the ovals which converge, INCREASING the convergence zone and the chance that rounds will form "Nodes", as you call them.

The ONLY way that the OCW method would work is if you had a perfect barrel and perfect ammunition which could predictably put the bullet through the same hole every time. (Such a barrel does not exist.)

Otherwise you have a flawed system which will report "false nodes" a substantial percentage of the time, a flaw which can ONLY be remedied by increasing the amount of sample sizes.

Which, you claim, simply wastes components.

 

[NeuseRvrRat]

i think some folks are mixing up Newberry's OCW method with a "ladder" test. the OP performed a ladder test. he did not follow Newberry's OCW method.

 

[Trent]

My comments were about the "ladder testing" the OP was using.

But OCW also typically works with insufficient group sizes. You need MORE than 3 shots, though, to iron out statistical aberrations. The 3 shot groups called out for in OCW are not sufficient to do this; with 3 shot groups, 25% of the time you'll get a 3 round group in which all three shots land within the upper or lower sector of the group, moving the group center artificially closer to the group above or below it than the "true" group center, which would be discovered on a larger sampling. (The chance of all 5 shots landing together, ABOVE or BELOW the group center is 6.25%; while the chance of all 10 shots landing above or below is only 0.195%.)

The more shots you shoot, the more your "group center" reveals itself. If you plug in the coordinates of the x/y for each shot you can obtain a true mean, which, after 10 shots, which will be within a .02% margin of error to the actual group center if you were to shoot an unlimited number of shots.

In other words, 10 shots give 99.815% certainty that you have revealed a group center accurate to within .02% margin of error of overall group size.

3 shot groups give a remarkably small confidence that you are finding the "real" group center.

If you read what I wrote in post 2, I use 5 shot groups on load development; which is sufficient to reveal a good amount of information about the actual group (with a confidence of 93.75%).

I shoot 10 shot groups to confirm, which give confidence of 99.815%.

Although this topic is not specifically about OCW, it demonstrates that OCW has the same shortcomings as the ladder testing the OP used.

 

[helotaxi]

I don't agree with the OCW method, mentioned specifically in the OP.

#1 The sample sizes are too small
#2 There's no control group for comparison
#3 It ignores the scientific method

And my response was to this post of yours, as quoted in my original post. All you're looking for with the initial OCW round is the likely area(s) to focus your attention. Close enough is close enough for this purpose. Once you have it narrowed down you expand the sample size. It isn't that difficult to comprehend and it doesn't take nearly as much in the way of components or time to arrive at a good charge weight, or better yet range of charge weights, from which you can work with seating depth to tighten the group size up.

 

[oldfortyfiveauto]

My 2 cents....I typically load 8 rounds divided through the min/max range. Then at 200yds (max at my range) I shoot one at each each target point. In this case shots 5, 6 & 7 show little vertical change. I then load 5 or ten at the middle range of the three for a group or two. If it still looks good I'll load and shoot another 4 or so groups to verify. This has worked well at weeding out bullet/powder combinations.