Fantastic information - It has changed my whole approach to reloading and will make it more meaningful and less tedious! I watched episode 50 and 52 on the Hornady KZfaq channel - it explains everything well. My take on these videos on this subject are as follows: 1. You need to shoot 50 rounds to get a 97% valid result for your group size and SD and ES. 2. A 20 round sample will get you within 25% of the valid SD and ES for your group size. 3. Sometimes you are lucky and get a very small group in only 5 shots. However the more you shoot that load the larger your group size will become. 4. Powder charge does not have a significant effect on group dispersion until you get to the max load +p level. Even then it is barely discernible after 50 shots. 5. Powder choice and bullet choice are the two most important variables in group dispersion for a given barrel. Chose a powder appropriate for the caliber you will be shooting. Choose a well made bullet appropriate for your caliber and rifle. Powder charges lower than max in any powder will tend to be more accurate. Velocity has no effect on bullet dispersion other than normal bullet drop, so any charge weight will be accurate with the same powder. 6. Seating depth does not have a significant effect on dispersion for modern match rifles. Maybe for older military surplus or sporter rifles designed for older calibers with less precise chambers. You would need a lot of rounds to get a significant test result in this area. 7. Barrel heating does not have a significant effect on group dispersion in Hornady's testing, although they did not test with thin barrels. They think that cooling off after every 10 rounds should work for any rifle that may tend to heat up more quickly. 8. Hornady recommends using the distance of the bullet impact point from the aiming point to calculate the mean radial deviation from the aim point instead of using the group size. You can record this for each shot and calculate SD and ES for it. It is more predictive of the expected bullet dispersion from the aim point of your next shot. Group size only measures the distance between your two worst shots and has no predictive value. 9. If your groups ( of 20 or more shots) are not circular, you may have a loose piece of equipment causing the issue; e.g. loose scope mount, stock forearm, etc. 10. Primers had not effect on dispersion. Shooting should be fun again - pick a powder and a bullet and any charge weight between starting and max and shoot twenty round and see how it goes. if it doesn't look good, choose a different powder and try again! Anyway it's going to save me time and frustration loading up 5 rounds each for 5 different charges weights and thinking I am going to get results that indicate the most accurate load! Now I will just pick a non max load and load up 20 rounds and shoot them as one string! If I like what I see I can load up 30 more and get the actual dispersion values. Thank you Hornady for all your work on this!!

At the risk of some hyperbole, episodes 50 & 52 are the most important videos in years for anyone interested in precision shooting. I’ve read a lot of the available military reports on small arms precision and conducted dozens of my own computer simulations to evaluate the usual suspects, the invariant measures (mean radius, CEP, standard deviation) and size dependent measures (ES, figure of merit, covering circle) to demonstrate to myself the costs of small group sizes. What I lacked was real data. Like most of us, I’m constrained by time and money. Your real world test data and podcasts are tremendously valuable as guides for us all. Thanks. Keep up the good work. Nag, nag … you should really write this up and publish.

Episode 50 and 52 were great topics with Miles and Jayden and it did change the way I view my process, it very much helped. These podcasts are a win so keep up the great work Hornady team.

Episodes 50 and 52 are my favorite by far. I just got into reloading and precision shooting very recently. I use statistics professionally and most of the load development tribal knowledge on the internet seemed more like magic or voodoo to me. i.e. shoot 1 shot ladders varying by .3 grains, stand up, turn around, bow to the east and shake a stick, etc. The info in these episodes is something I can really sink my teeth into. I’m glad that you folks do this kind of statistical testing and that you share this info with the public. I’d much rather shoot a .75 inch 20 shot group than a .25 inch 3 shot group. I’m the type of guy who is happy when I’m proven wrong about something because it increases my knowledge and gets me closer to the truth. Awesome work guys.

What I learned in '80s: Use one target for your first group (3, 5 or 10 shots, your choice), leave it attached and put target for second group over it. Keep shooting your "one group" targets with the first one underneath. After 10 groups shot, you will end up with 9 "one group" targets and one that shows what you, your rifle and your load can reliably do. You might be surprised how the "flyers" even out nicely into the pattern. No excel or other apps necessary!

More great info. I used to check my zero with a minimum amount of shots. If I hit the X first shot, that was good enough. Listening to what you said in episode 50 about a large zero sample was like a light being switched on! Common sense when you think about it, but I never thought about it!!

I want to compliment you on your emphasis on sample size when evaluating data. I have degrees in math and physics so I understand the importance from a statistics standpoint of larger sample sizes. I'm now 76 years old and have been shooting all my life. I recommend to shooters that they learn what standard deviation is and how to apply it to their testing results. It can be a great help. For example, if you measure the muzzle velocity of 20 shots and get an average of 1100 fps and an SD of 10 the two SD spread (95% = 20) you can expect, on average, that 19 of the MV measurements will be between 1080 and 1120. But that also means 1 of the 20 will be outside that range, either high or low. Understanding how this works helps a shooter decide if that occasional flier is significant. I never do any testing with less that 20 shots, usually 25. I'm also a strong believer in evaluating group size with Group Mean Radius. Here's an example. Lets say you shoot a five shot group with four shots hitting the same hole and one shot being off by 1 inch. If you measure the spread of this group it is a 1 inch group. If you use Group Mean Radius it is .320. Keep up the good work,

Another great episode! I'm actually making a playlist of your podcasts as great information I could look back to in my KZfaq library! 👍😀 After listening and learning about everything long range shooting for load development, I figured the best thing to do now is just to pick charge based on the velocity you want to run with. Very interesting though because when you listen to benchrest shooters, they are never afraid to go add more powder based on their observations on how their bullet holes look at the target!

Both of your episodes were extremely dynamic and helpful. I’ve been reloading for over 50 years and I’m still learning. Consistency, is the thing that always drove me crazy. I like how you mentioned that changing the bullets or the powder, normally is the main ingredient for the basic research for the best loan.

I caught this just in time. Makes perfect sense and I often catch myself missing the simplest explanation. Thanks for catching me this time. I just was falling into the vortex having done ten 3 shot groups of different charges, followed by three 10 shot groups of the charges with lowest ES/SD and really ended up being kinda stuck scratching my head on where to go, planning on seating test depth next for some reason even though all my loads so far are 20 thousands off the lands. Thanks again.

Very good video. Thanks Love it! Enjoyable too

This is third time I have watched these and I pick more data and information each time. This time, I picked up on the statement that you will not advise doing a 10 or 20 round group at one sitting with a button rifled hunting rifle. And that depends. With some of the budget rifles, the button rifling is not removing metal, it is displacing it. And it could move or deform in the heat of the barrel. Therefore, the better 5r button barrels are de-stressed by a heating cycle in an oven. This will keep the lands from mutating. But it could cost more. It depends on the job of the rifle. If you have a Mossberg or Thompson Center Compass hunting rifle, both being 5R button rifled, you don't need to do more than a 3 shot group. I am totally on board with the limit of 3 shots on a game animal, maybe a fourth as a "hail Mary" shot. So, consider how your barrel was made. That being said, you cannot find information on how Mossberg barrels are made but there is a company that does pre-fit precision barrels for them and you can choose options and the barrel will cost more than the original rifle did. Choose the job of the rifle and build accordingly. Deal with real MV. I have seen and heard recently that the Precision Hunter 7 mm PRC 175 gr ELD-X is not achieving 3,000 fps regardless of barrel length. On average for a 24 inch barrel, 2,800 fps. Which is fine. 7 PRC is still a great bullet and more available and practical than the 6.8 Western and easier to shoot than the 28 Nosler. In fact, thanks to another series by another guy, I may switch to the Hornady CX Outfitter 160 gr. And even if it is not 3k fps in real light, I am not going to shoot past 600 yards on an elk, which would still be doable with a PH but the CX is more likely to create an exit wound that aids in exsaguination and blood trail. Group the rifle for the job and be honest with yourself about it. Close in the range that you need for that bullet to function properly. I am a hunter and I like the safety net of 2000 fps impact velocity. So, I am going to close distance to ensure I have that velocity at imact.

I'm new to reloading (~1 year) and been having trouble coming up with a load development method in which I have a high degree of confidence that gives me reliable data regarding what my rifle is capable of for a given set of components. I'm also an engineer and familiar with statistics and have been wondering how some of these load development methods using minimal rounds for a given set of components could possibly be predictively accurate. Thanks very much for presenting some pretty compelling hard data and conclusions that make a lot of sense.

Brilliant! Please keep up producing these technical podcasts. No one else is putting out this information.

I always love these podcasts. It always ties something that I am learning in my engineering studies! Just enjoy it!

I have shot 3 shot groups as tight as .360 up to 1.297. same rifle same ammo same point of aim. I’d say your data is spot on.

Great podcasts, lots of very useful info. Well done and thanks a bunch, On testing that older caliber / chamber geometry, I'd like to throw the .30-06 into the ring, it's old but still in use almost everywhere.

Thank you for sharing this information! After reloading, hunting, shooting for 55 years you have proven what I have thought is reality. I have chased the 1/2" groups and about gone nuts with the unexplained point of impact change the next time at the range. I will use your information to set more realistic expectations for myself in the future. Also, I will not obsess over trying a bunch of different powder charges.

Marvelous insights guys. Very professional analysis and presentation. Thanks!

Another hour well spent--thanks! Part of the uproar is from the shooters who used 3-shot OCW testing "and my rifle shoots great!" Indeed, it does. Quality PRS rifles are remarkably accurate and consistent. But what they miss is that they could repeat that 3-shot OCW test five times and find five different "nodes". And the identified load would shoot well. But that 20-round group would show no significant accuracy or velocity SD differences between those loads. Litz did some interesting testing on barrel tuners, and I'd love to see you pursue similar testing.

don't stop, loving it.

Brilliant. This makes me wonder if what I think is a 1 MOA gun is really more likely to be 1.5 or more. All youtube rifle reviewers should watch this. Instead of using small groups to evaluate a rivle, they should be shooting 20 shot groups over multiple days to get a real idea of what a gun is capable of. This also shows the gun manufacturers "MOA guarantees" are completely worthless. All they have to do is come up with one 3 shot group sub MOA. Sounds like with normal distribution, even a 2 MOA gun can get lucky and string 3 together within 1 inch. The industry needs a more standardized test such as what has been explained here. I am constantly amazed by the great info in these podcasts. These feel like college level courses. Very helpful.

Another outstanding podcast that offers us shooters incredible insights that we could not gather by ourselves. Thank you all!

Fantastic work gents!! Thanks for the ground breaking research you are doing , and for educating all of us.

#1 good barre #2 good stock #3 good action #3 good trigger #5 good scope mount and level #6 good match ammo Sum of all is very very accurate system Spend the time shooting and not reloading and you will get good results.

An episode on cleaning and when it actually impacts accuracy would be awesome. Just as many divergent opinions/myths on cleaning. "I clean every (10,100,1000) rounds" information out there. Hard data on large sample sizes would be really interesting.

Great follow up to the last video on the subject of group size. I appreciate that you went further on the subject, it is great.

I remember White Feather saying that with a thin barreled rifle, after 3 to 5 shots, you’re going to be all over the place past 500 yards.

Great conversation, a great nugget of wisdom about every 30 seconds!

Yes!!! Get busy testing tuners!!! I'll be looking for that video. 👍

Very useful content. Thank you all very much.

Great follow-up, thanks guys!

Another awesome podcast. You're podcasts are super helpful in allowing me to realistically analyze my groups and load development.

Glad i found your channel. Excellent quality information with good explanations.

Another thing to consider is loading multiple small samples of say 5 per load with several different components then ruling out the bad ones immediately. Then load the goods loads in larger sample sizes. I think that's what most people were doing anyways.

Outstanding presentation, I am learning a tremendous amount from this project. It is changing my mindset both at the range and at the reloading bench. Again thank you for your time and effort posting this information.

I’m relatively new to reloading so this information is golden to me.

Excellent job guys. Enjoyed the follow up.

So, now that you are re-writing the book on all of our load development processes, I'd like to see a follow-up test, where you use conventional load development methods (OCW and velocity) with 3-5 shot groups to identify a node and a non-node. Then load up enough cases of both to be statistically sound. Then go shoot those two groups. My hypothesis is that if the load development doesn't matter, because it is the group size that matters, then the statistical groups should shoot about the same distribution. However, if the standard load development methods do have merit, then the group from the node would theoretically be smaller than the group from the non-node. I look forward to your findings.

You guys are on to something concerning this topic. Keep it up

On a similar note, at least from a statistics standpoint, I did a GPS accuracy experiment about 20 years ago. I was marking my property lines and I wanted to accurately note the corners so I could determine acreage so I decided to use the position averaging feature on my handheld GPS unit. So for a period of months, I went to each steel marker and took another waypoint on the each marker and let the GPS average the positions. I waited a few days between each position and waited until I had good GPS strength and a good accuracy circle. Every few weeks, I'd use the GPS to find the property markers. Now, I knew where they were. I could physically see them, but I used the GPS as if I was in the pitch black and stopped only when my Garmin unit told me I had arrived. The funny thing is, I never actually got to the posts. I'd get close, but I never stopped and looked down to find a piece of rebar... never. In a real world situation, this makes no difference because I save the location of my truck or campsite in the woods and as I get close to it I recognize where I am and shut the GPS off so if the GPS is within 100 ft, it's close enough. So I decided to test the position averaging theory with a spreadsheet. I created a "location" at position 0,0 and gave my spreadsheet GPS an assumed realistic accuracy of 45 feet, meaning I might be anywhere in the 45 ft diameter circle when I marked the location at position 0,0. So, for example I'd have the computer make a random mark at some position on the X axis that was on a radius of anywhere between 0 and 22.5 ft and at an angle of anywhere between 0 degrees and 359.9999 degrees. I then generated 10,000 random point locations and averaged the x and y offsets and compared them to the actual position of 0,0. The theory was that if I took enough samples, the average would eventually get me very close to zero or the actual position that I am "standing at" but what I discovered was that after 25 or 30 positions averaged together, the average never got closer to the real location of 0,0. In fact, it never got closer than about 10 feet of the actual location. The reality of it is that even if my GPS marked my location exactly at the coordinate where I was standing. Like if the earth was physically marked like a map with UTM/UPS meter markings and I wrote down the exact position where my rebar marker was on the earth into my GPS. Even if that was the case, the positioning error the next week when I went out will not likely be the same and the GPS will walk me to a point that is at least several feet, more likely 15 to 20 feet away from the pin. It will never get better, other than by chance and it will only be as close as the positioning error on that particular day. This discussion reminded me of that experience and as such made perfect sense. Thanks for the hard work.

I'd suggest making a cliff notes podcast on load development. Halfway through episode 50 I was about to say screw it and just sell my reloading stuff. Then you spent way too much time on methods that don't work and two seconds on the method you proved to be the best. And honestly I didn't even get all of that. Basically just pick a powder charge for a velocity you want, load up 10-30 rounds and see how it groups. If that doesn't work just buy a new barrel or try a different powder? Or a different charge weight that's completely different. Seems you're going to burn your barrel up just trying to find a load.

Horn. Please don’t waste time defending the excellent choices you have made in the last 15 years at least. Keep on it. Thank you. No better turn keys available.

Just installed a new barrel on my Ruger Precision in 6.5 Creedmoor. After the break-in, I began working up load data, and several trips to the range. Arrived at a powder and charge that was giving me good velocity and reasonably small groups - 1/4 " 3 shot groups :) Being honest with myself, I could see that the worst group was about 7/8". I have been trying Neolube for the case necks to provide even bullet release or what I thought would. Found that the non-Neolube necks produced better SDs. Trying different seating depths and primers was not producing any significant change. And now I watch both your Podcast on YOUR GROUPS ARE TOO SMALL:)

What a great set of podcasts! As a hunter I use smaller shot count tests than a bench rest guy should but invariably I ended up with no less than 10 superimposed 3 shot targets or 10 superimposed 5 shot targets and adjusted my zero based on the composite center of all the targets. I've done this for decades not knowing for a fact that it was the right waupy, it just seemed logical. The zero adjustment many times was 1/2" or more in either direction. The end result were rifles that were effective in the field. If bench rest shooters would double my test shot counts, (to 100 shots) it should improve their development. I know they have a high accuracy demand but they should try using a near stock 700 BDL Remington with a custom barrel on crows at 400 yards.

You make a good point with regard to determining small differences in hand loads by averages or standard deviation. However, it may be helpful to measure differences relative to the Standard Error of the Standard Deviation. What we are looking for in various loads is precision, i.e., tight groups. Precision is best measured by Variance (V) and to compare V across shot groups we can use Sample Standard Deviation (S). However, S only applies to the sample from which it was drawn. Since we are trying to avoid using extremely large samples, it is best to compare Sample Standard Deviations in light of their Standard Errors, meaning we need to calculate the Standard Error of Standard Deviation (SES). Unlike S, SES are estimates of error in the S based on a statistically infinite number of samples, in this case, an infinite number of rounds. For example, suppose I shoot a 5 shot group of load A. I first calculate the Center of Impact (COI) for that group. I then calculate the average distance to the COI of all 5 shots. Lets say that all 5 shots are exactly 3" from the COI. That gives us an average (A) of 3.0", and a S of 3.35". Then I shoot Load B. I shoot a 5 shot group with rounds that measure 1, 2, 3, 4, and 5 inches from the COI for an average of 2.8" and a S of 3.57". Though the average of load B is lower than A, , if we consider the S of each group we can see that load A has a lower S, which means a tighter pattern by 0.52". We might be tempted to say that load A is better than load B by 0.52", but there is a problem. The principle problem at this point is that the error in load A's S, as indicated by the SES, is +/- 1.18", which is greater than the difference we are trying to measure. All we can say so at this point is that load A is better than load B by 0.52" but +/- 1.18". That's not saying much. To make a statistically significant conclusion we must increase the precision our SES by increasing the number of shots. But how many shots do we have to take? Lets take a quick look at the formulae I am using for establishing SES. FORMULAE Variance (V): V = Σ e^2 ÷ ( n - 1 ), where e = individual errors of each shot as measured from the point of impact and n = number of shots Standard Deviation (S): S = √V, where n= number of rounds Standard Error for Standard Deviation (SE): SE = S ÷ √(2 * (n - 1)) (www.tqmp.org/RegularArticles/vol10-2/p107/p107.pdf, (19)) Crunching the numbers for a 5 shot group for each load we get the following as our baseline: Load A 5 rounds average errors in inches: 3, 3, 3, 3, 3 A = 3.0 V = 11.25 S = 3.35 SE = 1.18 Load B 5 rounds average errors in inches: 2, 4, 1, 5, 3 A = 2.8 V = 13.75 S = 3.71 SE = 1.31 We know we need more than 5 rounds. So, to find the exact number of shots needed for a given level of precision, we need to use the formula for SES and solve for n, then we can plug in the SES we need as our reference and calculate how many rounds we have to fire to obtain the level of precision we desire. SE = S ÷ √(2 * (n - 1)) ⇒ √( 2 * (n-1) ) = S ÷ RSE ⇒ 2 * (n-1) = (S ÷ RSE)^2 ⇒ n - 1 = (S ÷ RSE)^2 ÷ 2 ⇒ n = (S ÷ RSE)^2 ÷ 2 + 1 where S = sample standard deviation, RSES = the Reference Standard Error needed to measure the difference between load A's and load B's Standard Errors beyond the margin of error, and n = number of rounds that must be fired. We can now solve the question: How many rounds (n) would it take to get draw statistically significant conclusions about which round was better? Let's assume we want our RSES precision level to be 0.354". This means the error of our S will be +/- the diameter of a 9mm round. Then lets assume a confidence level of 68%, then plug in our numbers. n = (S ÷ RSES)^2 ÷ 2 + 1⇒ n = (3.35 ÷ 0.354)^2 ÷ 2 + 1 n = 45.77 n = 46 rounds If we assume that the delta in the SEs remains 0.52" after 46 rounds (a big "if"), we can be 68% confident that load A was patterning tighter than load B by 0.52", +/- 0.354". In real life, all the statistics change slightly as we add to the number of shots. But using our baseline numbers we can calculate a minimum number of shots needed to draw statistically significant conclusions without tedious iterations. The actual, accurate data will be calculated after the final number of rounds, and when those number roll in we would be 68% confident in our results +/- 0.354" without having to burn thousands of rounds. But suppose I wanted to increase both the level of precision and the confidence to, say, SE = 0.25" and a confidence of 95%. For this level of confidence I would multiply the results by the appropriate Z score (below) and calculate thus: CONFIDENCE LEVEL Z SCORE 0.90 1.645 0.95 1.96 0.99 2.58 n = 1.96 * (3.02 ÷ (0.25 ΔSE))^2 ÷2 + 1 n = 1.96 * 71.96 n = 141.0478 rounds n = 142 rounds For this level of increased precision and confidence, I would have to fire 142 rounds of each load, but afterward I could be 95% confident in the results +/- 0.25". Whatever level of precision and confidence we need, using Standard Deviation in light of the Standard Error of Standard Deviation instead of a simple Standard Deviation, we can obtain statistically significant results using a fraction of the rounds at a fraction of the cost.

*Please* tell me how this relates to barrel harmonics and barrel timing. I'm going to use the wrong words here, but you seem to be saying that both barrel 'whip' (up and down) and bore expansion with shock-wave (think OBT) aren't real effects on dispersion. Is this true? Any chance you could address these effects/theories in a part 3 please?

Great episodes! These are clarifying a lot of questions I've had. I completely agree with the comment made by Random Idiot who suggests using the 7x57 for testing throat diameter and principal axis tilt. I hunt with, and load for, the 7x57 and have heard for years that the reason it's so difficult to find accurate loads is variations in throat length. Now I am wondering if it really isn't an issue of throat diameter. The SAAMI chamber parameters for throat diameter are far more generous than the .0005 or .0006 inches over bullet diameter of the newer PRC-class cartridges. A comparison between two 7x57 chambers (one SAAMI and the other with a PRC-type throat) would be a good test to see which is really the culprit: throat length or throat diameter. Keep up the good work!!

For years I've been evaluating ten-shot groups by the mean radius and standard deviation. I really can't afford larger sample sizes. What I haven't been able to pin down is: within the constraints of a ten-shot group, how much difference do I need to see for it to be a real difference.

Outstanding and informative.

I am hearing from this episode and the 50 is that there are no nodes. That as you increase charge you increase dispersion. But this confuses me: the 20 round feeler groups. This suggests that one charge is better than another discounting the less charge = more accurate. I totally get the proper sample size aspect. I have beat my head against walls too often trying to verify results. It also seems that, generally, the heavier the cartridge, so to speak, the more forgiving and easier to reach a “good” charge than a lighter cartridge such as 30-06 vs .224.

These "your groups are too small" podcasts are some of the most amazing sources of data that I can't stop feeling giddy at getting to partake in all the data and conclusions that your hard work produced. Hearing that lower charge weights almost always result in lower dispersion, does that make it more likely for a 6.5 creedmore to have lower dispersion than a 6.5 prc given the difference in the charges of those cartridges? Seems like it would.

Even though I'm not handloading for a precision rifle, info is still good, started doing 10 round groups to get my carbines sighted in. Before I would chase my tail trying to zero with 5 round groups. Listening here I'll probably start going for 2x 10 shot groups to set my zero.

"The Law of Large Numbers" always tells the truth.

I’m shooting a new caliber. I shot about 300 rounds of load development to see what different loads did a bunch of times. People would say I didn’t need to. But for me, to have that confidence, I needed to know what my rifle did with all of those loads. Now I know.

Another breath of fresh, BS-free air. But I knew you guys were just pushing shooting more to sell more components and ammo, Miles confirmed it! You should next make a video on how to make foil hats...mine's kind of poking me. . FWIW, to underscore what you're saying about the non-universality of load info working well: Miles' two powders he likes in the 6 ARC, Varget and Leverevolution have both disappointed me in cartridges, with Leverevolution doing it in the 6 ARC...with 4 different bullets. What works well in one rifle will not necessarily work well in another which I feel you've touched on a little in this and Episode 50. There isn't *one* universally awesome component combination. Ya gots to find yer own path (load). Thanks for this great information.

Thank you guys for good talk, both 50 and this episode. I learned a lot. One question I have is about shooter skill. I'm a half decent shot, nothing amazing. I get that this kind of sample size (20-30) will handle a less skilled shot even better, because it gives a fairer picture of what I and the gun together can shoot. What I wonder is if you would say there's a skill floor to being able to say anything about a guns capacity? Like "You have to be x talk to ride" or "shoot to level of Y to use the data"?

The most interesting and informative podcast so far from a hands on standpoint but it predicates a question for the trio to answer please for me: I determine a load for my 6.5PRC using a certain powder and Berger 130VLD Fed 210 primer N165 59.2grains and I eventually obtain a three shot group giving 0.30 I then load another ten I clean the barrel thoroughly and religiously after every 10shots. Using this data obtained I then proceed to shoot a fouler after letting the rifle and ammo to adjust to sit outside to adjust to the particular ambient temperature invariably 3-7C at this time of year most mornings. I then fire off the fouler to warm and foul the barrel then proceed to shoot my next three /three shot groups resulting in a 0 40/0.50/0.30 SD across the nine shots is 8 and the ES was 12 This predicates the question Are these results a fluke? Can I assume that combination on that particular day weather conditions prevailing is capable of a consistent at least 0.5" group of three? The next day little or no wind again I shoot a 0.4 then 0.3 and think I'm good to go and load out 40 I then shoot several deer out to 325yds ok deer size target What's your take on this please. It took 21 rounds excluding foulers for which I use some very little used old design to get to where I got the 0.30" first group changing powder weight slightly and seating depth backed off from jam

The only question that I have is barrel node? Powder charge matters so much is possibly because of that node. If shooting in a anti node where the barrel is moving at the exit, changes that equal shots in the node when the muzzle isnt moving as the bullet exits. Next question, especially when shooting various distances, would be combustion process being stable to produce best or close to best extream spread, as well std deviations?

What Miles says about "bullets not liking certain barrels (or bullets/powder not liking certain barrels)" doesn't make any sense in the context you guys are speaking. Why does a bullet not like a certain barrel? There has to be a reason. How do you explain this? If there isn't a way to tell, then all the reloading efforts are pretty much moot points. But the extreme precision and accuracy demonstrated by some shooters indicates that they have an idea of what works. Also, there are competition shooters that consistently shoot (recorded) small groups. Much smaller than .5 moa, often close to .25 moa - yet what you guys are saying here is that this is pretty much impossible. How do you explain this? Thanks for the hard work and the time to present these technical subjects. This is very useful information, and you have me convinced to use 4DOF

A bunch of us would love an episode on your experiences with 6mm Creedmoor.

I really enjoyed this podcast. It really got me thinking about how many decisions we make in life based on small sample size. I do have a question. I have heard in this podcast as well as a previous one that you just pick a powder an velocity you like and go with it. My question is what is your criteria for not picking somewhere close to max charge? I know traditional thinking means you need to work up to make sure pressures remain safe in each individual rifle but other than pressure signs, for hunting and or long range precision, when would lower velocity be a preferred choice unless you are needing subsonic or you expected target range and impact velocity do not correlate to optimal bullet performance.

When I listened to the first podcast about sample size it was like a light switch turning on in my brain. I now know why one day I could shoot a tiny 3 shot group then a week later the group is much larger. One thing a 3 shot group will tell me is that it's not going to get any smaller if I keep shooting. Thanks for the information you guys are sharing.

With stringing, would that also apply to gas guns, as the piston would impart some vertical forces while firing?

Super great info. Thank you so much for the truth syrum

Would be awesome to have a part 3 that went deeper into the weeds on the math and statistics involved

I’m curious about your statement seating depth doesn’t really matter. I have to wonder if you’re not seeing effects from seating depth due to the fact that you’re using a truck axle for a barrel. Does seating depth have a bigger effect on a pencil barrel?

Great information again!

This is also a good follow up. Thank you guys for sharing all your hard work! It’s going to save me time and money! Please can you test whether barrel tuners work? After watching your videos and the people selling tuners videos it seems there is no proof of concept as the “tuning” involves a series of 3 shot groups

I've always worked up my gopher loads as being able to hold minute of gopher head at 200yds, will the edge of that group stay safely inside a gopher head, that is shooting 8-10 5 shots or 3-4 10shot grps. I like the hit percentage system. If I read the wind right and shoot the gun properly, I will most probably smoke that sucker, and I'd have to do something fairly drastically wrong to miss, and that will happen sometimes. I've seen the dispersion over a collection of targets over time, with all the same load, different conditions. It is reality. Where did the worst shot go, how far from where I aimed was it actually, would it kill what I was shooting at, or at least put it down, off cold bore shots? I like the way you folks explain your tests.

I revisited this podcast. May I suggest title be changed to be more descriptive on a search to "Your Groups are still too Small - Implications for Load Development"

Really excellent followup. I'm still unclear as to whether a slight change in charge weight (.3gr) will have an effect on poi. I guess my question is whether in your research and testing have you found there is such a thing as a pressure tolerant load moreso than other charge weights, all else being equal.

Why do you make your cases so short; ie less than trim length?

Wow! This is the first episode that I’ve watched. That was a lot of info delivered in a “digestible” manner. You changed my mind about load development and zeroing. Could you edit the video to better show the monitor that JD was using?

Question: How do you find the charge to start with between Max & Min? I use quikload and GRT to calculate a Node on the rifle and do load development with that. Is a barrels Node important then? Thanx for an awesome podcast!

Apologies if I missed it, but for the hunting load workup, do I understand that you pick a single charge weight per powder? Or do you shoot various charge weights before moving on to the next powder? If it's the former, how do you decide on the charge weight? I'm new to reloading, but I'm also a professional researcher who spends his days figuring out what we can learn from sample sizes that are almost always smaller than we'd like, whether or not we have enough samples to observe the effects we expect, etc etc. I have to say these two episodes are hands down the most refreshing reloading content I've found. In a world of VERY bad data logic, it's nice to know that there are at least some who understand these concepts. THANK YOU!

Amazing podcast! information based on significant data is what i have been looking for, without me knowing it! i have allways felt that most data is to random and therefore inconclusive could you assaign a number for the steps in reloading based on importance?. for example if powder and bullet are nr 1 and reducing powder charge is nr 2 How significant is selection of resizing dies, if at all? fulllenght sizing vs neck sizing, does it matter? How mutch does consentrisety matter, if at all?

More great information on sample size for a given load to see if it is legit. I would now like to see an hour and a half pod cast on your traditional hunting bullets that are not in your 4DOF system. The Interlock and SST bullets have worked great for years. How do they actually compare against the ELD -X at traditional hunting ranges, 25-350 yards? Coyote, pigs, and deer species. Some gel test at 2400-2700 muzzle velocity would be nice to see at 200 yards. What the hunting bullets actually do at range in gel. Not 100 yard high velocity test. I have never had a rifle that shoots a max load in your hand loading manual well year round during practice and hunting season. I have always had my best groups in my rifles just a couple grains past half to three quarters the way up the ladder from minimum to max load. Hopefully I said that right.

The cip reamers with zero parallel freebore (7x57 Mauser) would be about the worst you could use. Even the saami .308 is much better designed.

In general with load development, looking for the highest useable speed, would you start with the fastest burning powder first or the slowest? Would that change between spherical or extruded?

Thanks!!

Any thought on modeling your data to develop a teaching tool that would allow individual system inputs to suggest a dispersion outcome?

What increments did you use for testing seating depth? What if muzzle exit timing is very important in reducing dispersion but your test methods skipped right over areas where meaningful changes would be seen? If you test in 0.015” increments, say, but actually need to test in 0.003” increments in order to find those differences, how can you do confidently say it doesn’t really make any meaningful differences in dispersion? You say it doesn’t as far as you saw. This may indeed be true. But is it possible your methodology caused you to miss something? Are barrels infinitely stiff? Or do they move? If they move, can you do anything about it? If they move, can you work with it and improve dispersion? If they move, can you work against it and worsen dispersion? What I heard being said over episode #50 and in this follow-up #52 is that statistics are a thing. Normal distribution is a thing. And that you believe these things very deeply. I don’t doubt this. Nor do I doubt those things. And so I don’t have any issue with you believing in them. But what I seem to have noticed is you guys don’t think there’s anything benchrest shooters know that you don’t. You don’t think F-class shooters know anything you don’t. And that you think that the things they believe they know are based solely on small sample sizes and are therefore meaningless. Believing something based on small sample sizes is probably foolish. We are in agreement there, one hundred percent. But if these guys have found things in their decades of shooting in competition that make their results worse if they omit those steps, it seems to me those things are worth investigating properly. And there in lies the rub. I don’t think you’ve properly tested them out of sheer ignorance. “People say seating depth matters. Let’s test it.” “Sounds good. What increments should we use?” “I don’t know. How about 0.015” increments?” “Sounds good. Let’s do it.” And that’s what you do, rather than investigating to find out what increments these highly successful shooters test with and find very good results with, you test in some other random fashion and come to the conclusion that the act is random and produces random results and so does not matter. You have a wealth of experience doing the things you do. But this does not mean you already have all the answers. My suggestion would be to take advantage of it when any of these highly successful shooters are willing to share their methods and findings with you so that you can then design tests that are similar, if not exactly the same, so that you can then try to determine whether or not it is a meaningful thing to worry about. You say with a huge amount of confidence that seating depth doesn’t matter. But it seems to me you haven’t even tested it properly yet. Barrels vibrate. Which point during that vibration the bullet exits the muzzle makes a big difference. This seems to be the largest factor you are missing. You have practically unlimited resources to test with and you’re half-assing it. It is aggravating to witness. Go to winning shooters and ask them how they think you should test, even if you think what they’re saying doesn’t hold water, test it their way anyway. Approach it as though you still have things to learn and that others may still have things to teach you. You don’t know everything. Neither do they. But they seem to be doing a better job than you are, so maybe you should listen to them for a little while. Maybe you’ll learn something new. It is possible.

Is there a software that I can use that uses the standard deviation of the mean radius to help me select a load and use it to best predict future performance?

Would you consider the 7mm rem mag to have a "sloppy" chamber design compared to creedmoor, PRC chambers?

Hey guys I love your pod casts and the information you present. I've been interested in muzzle brakes have you done anything along that line in a podcast? More specifically radial brakes v ported break's, one better / worse, reduction in recoil which is better / worse?

I'm sure I'm not alone here but I wish I had this information 15 years ago.

What is an e.g. "1 moa rifle"? The bell curve is infinite so all rifles are infinite moa rifles. Might I say instead that a rifle has a 95% chance of shooting a 1 moa 5-shot group? Would that be an adequately constrained definition? 5-shot group size is such a good communicator that perhaps there should be a statistically valid standard so that when I say 1-moa group people know precisely what that means.

It's funny that Miles says pick one load between minimum and maximum, and it will shoot as good as any other then says on his own rifle that he will try two or three loads with a 1/2 grain or one grain difference to see how it shoots. Does he shoot 500 or 600 rounds to see if it's any better? He's contradicting himself and what he has told us his findings in a "lab" were.

You should have eric cortena on the podcast about this or go on his KZfaq interview

Talking about group shape. Agreed in a test barrel it would be around but most of us are shooting barrels that have harmonics so wouldn't that account for vertical dispersion more than loose fixtures.

A question for Miles. Do you shoot a group of 20 at sub MOA? That seems to be what I gathered from your load choice methodology. Do you put the rifle in a fixture in order to take the human out of the process to make this evaluation? Can you please elaborate on your process? I have enjoyed both of these podcasts and greatly appreciate the data driven approach.

Great content as usual, thank you! Just a thought, but could your rifle mounting system/accuracy fixture be causing some inaccuracies in your findings? If a straight taper barrel is clamped in a V shaped channel this could change the harmonics when fired and the way it behaves when it heats up. Would it not be a better option to mount it on the same way you would in a stock or chassis using the bedding area and action screws? Sam

26:02 I think that pretty much disproves the theory people have about 50 year old powders 'not being the same that they once were'.

Thanks for sharing. Making population inferences using small sample sizes full of bias….from hero to zero….unless you are lucky.

Agree with most of what was said here. However, loading the way Miles does you are almost never going to find the optimal accuracy of the system. It's entirely possible that a very accurate barrel could hate the bullet you started with but still give you that 3/4 MOA group. However, if you tested different powders/bullets it is quite likely that you could get that group size lower. The more accurate your system is the more margin for error you have when you pull the trigger. At the end of the day it's all about your own expectations but mine are that I at least try to find the best combination I can for my barrel.

Does the sample size variation apply to pressure as well as group size? Is a ladder test looking for pressure signs with one shot at each load invalid, or is the SD on the pressure generated much smaller? Currently working up a load for my dad's 7 mm rem mag and this has me thinking.

Real question.... Where you only using the factory loaded ammo and if so, how do you know that you weren't simply seeing the manufacturing variations in that ammo? Would a more precisely loaded ammo take smaller groups to get to that level of distribution? Where it took 2 boxes of factory ammo for the distribution to stabilize, would a more precisely loaded ammo's distribution stabilize at say 20 rounds?

What is the relevance of your testing to the reloader? Or is it?

lots of amazing info, wish SST bullets would come with better grain weights. the deviations in a box are way to big, almost a full grain differences!