Load development based on chronograph data

[iouri]

I was thinking since there's a lot of variables that might affect pistol group size (unless one uses ransom rest) wouldn't it be more beneficial just to find load with lowest SD. Shouldn't it be most accurate ?

[Sevens]

Hasn't proven to be so, and if the Chrony is what you are using and with no extra calculations, your SD is horribly limited to 10 shot strings. This is absolutely my biggest complaint with the Chrony, which is an otherwise decent tool.

[iouri]

No chrony was a "general" term - I've recently bit the bullet on LabRadar (after I've shot my Chrony ). Don't think it has that limitation. Tested it indoors and after I've figured out muzzle location it was registering every shot and it's less dependent on that "ideal" shot placement across sensors.

[JBnTx]

Buying a Labradar is the best money I've ever spent.

[iPac]

Originally Posted by JBnTx

Buying a Labradar is the best money I've ever spent.

I just looked one of these up on Midway, and let me just say WOW!!!

For both price, and what the machine actually does. Sounds like an incredible piece of equipment if you want the best.

Unfortunately, I'm with the ProChrono. Limited load development doesn't require much.

[rockquarry]

In theory, I suppose a low SD would be a contributing factor to best accuracy, but probably only noticeable at long range. In actual practice, at least at distances most shooters use a handgun, it makes little or no difference. An unusually accurate handgun, however, might skew my guess.

Unless I get very large variances on chronograph readings, I don't even consider SD when working up a load.

As for a Ransom rest, it works very well, but the biggest advantage this device has over good bench technique is that it does not suffer shooter fatigue which leads to wild shots. I owned a Ransom rest with windage base years ago and used it considerably. However, as some experts have remarked, one should be able to duplicate 50 yard Ransom rest results at 25 yards if skilled at shooting from a benchrest. I tend to agree. If a shooters requirements are more stringent than that, certainly a Ransom Rest would be necessary.

[Wise_A]

As soon as chronographs became widespread and affordable, precision pistol shooters immediately figured out that low standard deviations had absolutely no correlation with accuracy. In fact, they found out that the most accurate load was rarely-to-never the lowest SD load.

[JBnTx]

Originally Posted by Wise_A

As soon as chronographs became widespread and affordable, precision pistol shooters immediately figured out that low standard deviations had absolutely no correlation with accuracy. In fact, they found out that the most accurate load was rarely-to-never the lowest SD load.

I've noticed that too.

Some of my most accurate loads have had some scary SD numbers.

I understand what SD is, but I have never understood its worth as an indicator of a most accurate load.

All I've ever been told is that a low SD number is better.

[BB57]

Originally Posted by Sevens

Hasn't proven to be so, and if the Chrony is what you are using and with no extra calculations, your SD is horribly limited to 10 shot strings. This is absolutely my biggest complaint with the Chrony, which is an otherwise decent tool.

A long, long time ago I started copying (or printing) my numbers on to a sheet o paper and then dropping my data into an excel spread sheet later after I got home. a long, long time ago. It's a more efficient use of range time, you can shoot strings as long as you like and add the strings together if you prefer.

Originally Posted by Wise_A

As soon as chronographs became widespread and affordable, precision pistol shooters immediately figured out that low standard deviations had absolutely no correlation with accuracy. In fact, they found out that the most accurate load was rarely-to-never the lowest SD load.

That's a large and very sweeping statement to make given the variables involved.

SD isn't the only concern a shooter should have, and it's not the only thing driving accuracy, but it still tells you a lot about the load.

The velocity of the load has a significant effect on accuracy in a rifle due to barrel harmonics that are probably just not an issue in a handgun. However the effects of shot to shot differences of recoil on accuracy are more pronounced in a pistol.

A smaller SD also confirms that that things are happening consistently with a load - adequate but not excessive ignition, uniform and consistent powder burn, uniform and consistent case length and crimp, etc.

And of course a chronograph in general will tell you what you are actually getting velocity wise from a load - which varies a lot from revolver to revolver, given differences in chamber sizes, throat dimensions and cylinder gaps, even before the bullet reaches the barrel.

If you're shooting a self defense load, a low SD also gives a much higher degree of confidence that each round will produce velocities inside the bullet's optimum performance envelope.

That certainly doesn't mean I'll throw out an accurate load with a higher SD, but it does mean that all other things being equal, I'll take the load with the lower SD when accuracy is comparable.

[BB57]

Originally Posted by JBnTx

I've noticed that too.

Some of my most accurate loads have had some scary SD numbers.

I understand what SD is, but I have never understood its worth as an indicator of a most accurate load.

All I've ever been told is that a low SD number is better.

Pistols and revolvers are not the best examples to use as wider differences in velocity have less effect on consistent trajectory than in a rifle, given that most people seldom shoot a pistol or revolver at more than 50 yards, and in the vast majority of cases, the ranges are 15 yards or less. And of course most shooters don't shoot pistols and revolvers all that well anyway.

However, it's impossible to get consistency without a low SD, and consistency in your reloads is never a bad thing. If nothing else its a form of quality assurance.

[JBnTx]

Originally Posted by BB57

Pistols and revolvers are not the best examples to use as wider differences in velocity have less effect on consistent trajectory than in a rifle, given that most people seldom shoot a pistol or revolver at more than 50 yards, and in the vast majority of cases, the ranges are 15 yards or less. And of course most shooters don't shoot pistols and revolvers all that well anyway.

However, it's impossible to get consistency without a low SD, and consistency in your reloads is never a bad thing. If nothing else its a form of quality assurance.

Thank You.

How do I educate myself about ES and SD? Is there a book or a website?

Except for their definitions, I know very little about ES and SD. I had to look that up when I bought my Labradar.

I shoot handguns (popular common calibers) at 15yrds, from a rest, with 10 round strings of shots.

Most of my ES and SD numbers are low, but sometimes I'll get some really high three digit numbers. Some of those are my most accurate loads.

Where can I learn why this is happening?

[iouri]

Originally Posted by BB57

However, it's impossible to get consistency without a low SD, and consistency in your reloads is never a bad thing. If nothing else its a form of quality assurance.

That's was my reasoning as well: low SD would mean consistency and in theory consistency should lead to smaller groups.

[DWalt]

Originally Posted by Sevens

Hasn't proven to be so, and if the Chrony is what you are using and with no extra calculations, your SD is horribly limited to 10 shot strings. This is absolutely my biggest complaint with the Chrony, which is an otherwise decent tool.

The simplest Chrony has no statistical functions at all, just the MV readout. However, you really don't need any if you have a Microsoft Excel spreadsheet. Just write down each MV as you fire and enter them into a column in Excel later - however many you want. Then run the @AVERAGE() and @STDEV() functions on the column to determine the average and standard deviation. Excel has about every statistical function that you would want already built in. I have done that for many years.

"How do I educate myself about ES and SD? Is there a book or a website?"
Try Google. Or Wikipedia. Standard Deviation (SD) is principally a statistical measure of how closely data points are grouped around an average. The smaller the SD, the closer the data points are grouped together. In general, three SDs on either side of the average (or mean) will encompass effectively 100% (actually 99.7%) of the values. Let's say you fire 20 shots and determine that the average MV is 1000 ft/sec, and calculate an SD of, say, 20 ft/sec. Three SDs are 3 x 20 = 60 ft/sec. so the expected MV range of a very large number of shots would be 1000 +/- 60 ft/sec (940 to 1060 ft/sec). If the SD is 5, then the expected MV range of the population would be 985 to 1015 ft/sec, therefore more consistent than an SD of 20.

[Nevada Ed]

+1 with post #7......

I only worry about my test that go over 120fps +/- and I may
retest or drop that loading, however it very seldom happens.

My 38 J frame snub surprised me with the 135gr Gold Dot that
had a 9 ES with five shots !!

That is its high point.......
you don't want to hear about the rest.

[fredj338]

Originally Posted by iouri

That's was my reasoning as well: low SD would mean consistency and in theory consistency should lead to smaller groups.

Thats just it, its theory. Even in rifle, low sd doesnt automaticaly mean small groups, other variables are in Play. Some of my most accurate pistol loads will have SD in the high 20s (unique & lead bullets). Even in precisionrifle, small SD doesnt always give tightest groups. Es/Sd & a chrono are just tools. The proof is always measured on paper targets.

[BB57]

As noted above you can google standard deviation and get an idea how it works.

It's based on the idea of a normal distribution curve, where essentially all 100 rounds in a box of ammunition would have velocities within +/- 3 SD.

If the average velocity is 1,000 fps, and the SD is 15 (not impossible with a good handgun load), then the distribution should look like this for a 100 round box of ammunition:

---3rd SD (low)---

955-984 fps
2 rounds

---2nd SD (low)---

970-984 fps
14 rounds

---1st SD (low)---

985-1000 fps
34 rounds

Average velocity = 1000 fps

1000-1015 fps
34 rounds

---1st SD (high)---

1016-1030 fps
14 rounds

---2nd SD (high)---

1031-1045 fps
2 rounds

---3rd SD (high)---

The expected extreme spread is then 955 fps to 1045 fps, or 6 times the SD, or 90 fps, +/- 45 fps to each side of the average velocity.

More importantly however, you'll have 68 out of 100 rounds with velocities that are within 1 SD of the average, and 96 out of 100 that are within 2 SD of the average. That gives you a more practical spread of just 60 fps. (95% of the time), and the bulk of those rounds (68 of them) will fall within a spread of just 40 fps.

If the SD is 30, then you'll have the same normal distribution but the curve will be flatter as 6 times the SD will be 180 fps, which will be very close to the actual extreme spread you've observed.

-----


One other thing to consider is that sample size matters.

The statement posted above that "the most accurate load was rarely-to-never the lowest SD load" suggests that the shooters jumping to this conclusion were basing it on small sample sizes.

I've observed frugal shooters developing loads using a ladder approach with as few as 3 rounds per increment on the ladder. That's not enough to determine if the firearm is shooting better groups or not.

Normal distribution applies to group size as well as accuracy. A small group has lower odds of having a round that falls in the second or third standard deviation on the bad side of the distribution, and regardless of group size, the odds of shooting a single group that is above average are literally 50/50.

Small sample sizes also don't give you a solid SD number for the same reasons. If you get lucky with no 2nd or 3r SD rounds in your small sample, it'll make the SD look better than it really is. Conversely if you are unlucky and get some 2nd and 3rd SD rounds on each side of the average, the SD will look worse than it really is.

----

Personally, when I am developing a load using the ladder method, I load 15 rounds at each step, shot in three separate 5 shot groups in three separate "heats".

I then compare the velocity and accuracy data for all the rounds to see if there are any potential patterns related to velocity and accuracy in the data.

Those loads that look promising get another 15 rounds hand loaded for comparative testing purposes, and the load I then select for the final testing gets another 30 rounds produced.

By the time I am done, I've got a total of twelve 5 shot groups and 60 rounds of chronograph data with that load before it moves forward to production.

I also have six 5 shot groups and 30 rounds for all the second round loads - and 30 rounds is about the minimum number you need to have to draw any solid and meaningful conclusions on accuracy or on SD.

If you have a ladder with 6 loads, Round 1 will take 90 rounds. If you have 2 adjacent loads that look promising, I may add a load smack in the middle to see if there's a velocity node in between them where accuracy is even better, and consequently Round 2 will take another 45 rounds, and Round 3 to confirm the choice will take another 30 rounds.

165 rounds sounds like a lot, but when I'm talking about "production" of at least 2000 rounds, 8.25% expended on testing isn't bad - compared to the alternative of using a single round of testing with a single 5 shot group for six loads. That's just 30 rounds, but if you get it wrong, you're going to get less than optimum results with the next 1970 rounds.

Over the long term, that solid velocity data also helps when I switch to a different lot of powder. I can use the nominal load developed in the initial testing as a start point, and then adjust the load as needed to give me the same average velocity.

I'll also test loads under different temperature conditions to see how much velocity is affected by temperature change. If the change is significant enough to cause issues at extremely low temps, I'll develop a cold weather load with more powder in order to achieve the original warm weather velocity.

[BB57]

Something else I look for in load development are SDs that increase at the high and low ends of a load ladder.

Low load densities can cause pressure spikes and you start to see that potential develop in significantly increasing SDs before you get low enough to see a dangerous spike in pressure.

In some loads I also see the SD increase significantly near maximum pressures. That suggests very inconsistent burning of the powder.

[BB57]

Originally Posted by fredj338

Thats just it, its theory. Even in rifle, low sd doesnt automaticaly mean small groups, other variables are in Play. Some of my most accurate pistol loads will have SD in the high 20s (unique & lead bullets). Even in precisionrifle, small SD doesnt always give tightest groups. Es/Sd & a chrono are just tools. The proof is always measured on paper targets.

Let's shoot at 1000 yards sometime and see the difference between a load with an SD of 15fps and a load with an SD of 45 fps and lets shoot for total score with 60 rounds fired in three 20 round stages.

If our average velocity is 2800 fps, the load with an SD of 15 should have nearly all the rounds with velocities between 2770 fps and 2830 fps, with perhaps 2 or 3 rounds with velocities +/- 15 fps above or below those numbers (the 3 SD rounds we're expect in 60 rounds).

With the 45 fps load, most of the rounds would fall between 2710 fps and 2890 fps, with 2 or 3 rounds +/- 45 fps above or below those numbers.

In MOA, from a 100 yard zero, 2800 fps produces -35.5 MOA of drop (BC=.450) at 1,000 yards and 8.9 MOA of windage (10 mph full value crosswind).

2,770 fps produces 36.4 MOA of drop and 9 MOA of windage.

2,830 fps produces 34.6 MOA of drop and 8.8 MOA of windage.

Thats a total difference of 1.8 MOA in elevation and just .2 MOA in windage due to changes in velocity.

With the SD = 45 fps load, the velocity spread increases an so does the drop and windage:

2,710 fps produces 38.4 MOA of drop and 9.3 MOA of windage.

2,890 fps produces 32.9 MOA of drop and 8.5 MOA of windage.

That's a total spread of 5.5 MOA in elevation and .8 MOA in windage due to changes in velocity.

---

To put that in perspective, with the high SD load, I have 3.7 MOA greater variation in elevation at 1000 yards, and .6 MOA greater variation in windage. That works out to 38" more dispersion in elevation and 6.3" more dispersion in windage at 100 yards. I probably won't notice the windage differences, but I will definitely notice the elevation differences.

Precision shooters won't shoot with a sight or scope that has 3.7 MOA of slop in the adjustments, if they can avoid it - and the smart long range precision shooters won't use a high SD load with a comparable amount of "slop" in it.

A small SD, and an understanding of what might be due to the load itself, is also important when making adjustments at long range. For example, if I know I can expect +/- 9" at 1000 yards just due to velocity differences, then I'm not going to be too quick to make an adjustment on a shot that is 8" low. I'm going to take another shot and then decide if I need to make the adjustment. That's important if 2 sighters is all you get before you have to start shooting for score. If you've got a high SD load, with +/- 30" of potential dispersion due to the load, you're going to be wasting sighters and chasing sight adjustments the entire time.

----

In other words, yes, the proof is always measured on the target, but a 100 yard target tells you virtually nothing about how well the load will perform at long range as high and low SD groups all look good at 100-200 yards.

Thus, you've also got some solid chronograph data with reliable SDs to back it up if you want to select the load that will give you best results at long range.

-----

Now...we can argue about how relevant this is to pistol and revolver shooting, where the ranges are short.

It's a non issue in practical competition, but it may well make a difference in bullseye competition due to both trajectory and recoil induced angle of departure differences that result from greater differences in velocity.

The bottom line is that a load with a lower SD minimizes the magnitude of one of your more important variables, and that is never a bad thing.

[Sevens]

For the folks that suggest writing down each shot's velocity and then later, at home, entering each shot in to an Excel spreadsheet -- yes, this will absolutely work.

However, as BB57 points out above, getting a genuine and relevant look at a true SD simply requires a larger test pool. The Chrony Beta Master that I have limits to 10 shots.

If I had ten shots that returned a low SD, my next step would be 25 shots. If those 25 shots also returned a low SD, I'd wish to send 75 or 100 through.

I'm sorry, there's just no way I'm going to write down all these shots. No possible way. Are you suggesting that I click click click the Chrony and write after each 10 shot string? Or actually put the handgun down between shots and write?

And then, punch all this in at home? All because the little Chrony with absolutely plenty of computing power but NO WAY to access it thinks "meh, 10 is enough."

Sorry, the Chrony doesn't measure up. It has the power but it's programmed to be handicapped AND it has all this power and you're stuck trying to manipulate it with three push buttons.

[iouri]

Sevens, you may want to look on ProChrono Digital chronograph - they have a bluetooth module which connects to a phone or tablet and their application can export data compatible with Excel. I agree about Chrony's restricted potential. They do have the aux connector which probably can be used for similar bluetooth attachment. It looks like they just don't care.

[Rule3]

Originally Posted by iouri

Sevens, you may want to look on ProDigital chronograph - they have a bluetooth module which connects to a phone or tablet and their application can export data compatible with Excel. I agree about Chrony's restricted potential. They do have the aux connector which probably can be used for similar bluetooth attachment. It looks like they just don't care.

Absolutely, One can get a Pro Chrono Digital for less than a Shooting Chrony and the Pro actually works!

Yes, I bought and returned two Shooting Chronys.

The argument of well, if you shoot the Pro it is dead, well if you can not shoot through the uprights at 10 ft, You do not need a Chronograph. Even if you do they will repair or replace it for half price.

Many many threads on chronographs.

[Sevens]

Yes, many forum discussions on chronographs, not my first rodeo. And I appreciate the suggestion of the Pro Chrono Digital, I have been considering that one for some time.

I didn't post on this thread to cry about my Chrony Beta Master. (for a side laugh, the packing list shipped from Natchez literally called it the "Chrony Master Beta" and if you read that aloud, you'll see the humor and if you own one, you'll see how well that fits)

I started my first handloads in 1989. My first chronograph was this one, the Beta Master, in 2012. I call it "the fun sucker" because it's the handiest little device for funneling all the joy in a range day and swirling in to a toilet-like flush/loss. It wasn't necessary for my needs in 1989, it wasn't necessary in 2012 and it sure isn't necessary for my needs here in 2017. I have no problem admitting that this frustrating tool was worth the $138 I spent to get it in 2012. At it's core, this Chrony does it's basic job quite well -- it grabs shots almost always and doesn't miss them.

For certain, many folks can make far better use of a working chronograph than I care to try, and I'm definitely not suggesting that a chronograph isn't a solid tool for the kit. I do believe their useful value (especially to folks that are 98% handgunners such as myself) is limited. And they spawn all kinds of theoretical discussions on forums but I agree with most posting above that for the average shooter wielding a handgun you aren't likely to see the advantage of the "perfect" handload over a decent handload.

I posted in this thread because the OP called his tool a Chrony, he was using the term generally, and I don't think it's fair to discuss the value of SD if your strings are limited to ten shots. My care to discuss this device extends no further. If measuring SD is something that truly matters, using any device from the Chrony brand lineup is genuine mistake.

[Nevada Ed]

I don't know about revolvers and pistols but..................

I have bullets in my 22-250 and 30-06 that don't "Settle down,
until they get out to 200 yards...........
due to my barrel twist vs the bullet style and design.

There is also elevation and temperature that can change your data.
Shooting at an elevation of 5,000 feet is a lot different than sea level.

A chrony is fun to play with but can I do without one..... ??
I did for 38 years.

[5-Shot]

I once worried about low SD's too. Then I conducted the experiment shown below. I attempted to load each round 100 ft/sec faster than the previous one. They were fired in the desert through a 222, over a chronograph at a measured 100 yards. I had a duplicate target on my bench and a 30x spotting scope. That allowed me to fire a shot, measure its velocity and mark its location on the target. Each square on the target is precisely 1 inch. The results speak for themselves. (Yes, I admit the results would probably been very different at 300 yards.)

Ed

[BB57]

I used a PACT chronograph for years, until I literally wore the sensors out on it. It was a vey good unit.

I bought a Chrony Master Beta (someone wasn't thinking that name through very well) because it was cheap and available right now when my PACT died.

The main problem with the Chrony series is the incredibly poor user interface. Awful is too kind a word for it. I haven't messed with it's "features" in years, but instead just log the data. I've still not replaced it however because it works quite well at capturing and displaying accurate velocity information (provided you set it up properly).

Personally, I have no problem remembering 4 numbers, and the 5th shot will still be on the screen, so there's no problem shooting the 5 or 6 shot (revolver) string and then writing it down afterwards. If that's a stretch you can use the 10 shot string feature to record the 5, 6 or 10 string and then record it.

You can also get a printer for it that will connect to the unit and print all the shots (and you can write notes about the load on the print out. It's $110 as an add on. If you get the Chrony Master it's only about $75 more, and you get a 500 shot memory in the Master compared to 60 in the Beta.

You can also get a PC interface (although it's an RS-232 connection so you also need a RS-232 to USB adapter). That, in combination with a 60, 500 or expanded 1000 shot memory will allow you to export directly to an excel spreadsheet. But given the rather dated technology it's more of a pain than it's worth.

[DWalt]

"I've observed frugal shooters developing loads using a ladder approach with as few as 3 rounds per increment on the ladder. That's not enough to determine if the firearm is shooting better groups or not. "

If one is truly interested in determining comparitive grouping performance precisely with a high level of statistical confidence, he is wasting his time by firing any fewer than five 10-shot groups (all under identical conditions), and then calculating the average extreme spread (ES) of all groups fired as a basis for assessing grouping performance. Five shot groups are largely worthless unless you want to fire at least 25 of them to get an average ES for comparison. 10-shot groups are vastly more data rich than five shot groups. Consider that a single 10-shot group can be arranged into 252 different combinations of five-shot groups. To wit: 12345,12346,12347,...678910. Five-shot groups are fine for zeroing your sights, but that's about all they are good for.

There is one other item of interest, which is the circle of maximum dispersion (CMD), defined as the diameter of a circle which will cover an infinite number of shots at some distance. Statistical analysis has shown that the CMD is equal to about 1.25 x the average ES of at least five 10-shot groups. For example if the average ES of five 10-shot groups at 100 yards is determined to be 2", the CMD at the same distance will be approximately 2.5" (1.25x2"). Some call the CMD the "Covering Circle."

[BB57]

Originally Posted by Nevada Ed

I don't know about revolvers and pistols but..................

I have bullets in my 22-250 and 30-06 that don't "Settle down,
until they get out to 200 yards...........
due to my barrel twist vs the bullet style and design.

There is also elevation and temperature that can change your data.
Shooting at an elevation of 5,000 feet is a lot different than sea level.

A chrony is fun to play with but can I do without one..... ??
I did for 38 years.

I use a whiz wheel when shooting at long ranges where I am ranging the target to get the distance (using either a laser or a ranging reticle).

It accounts for pressure and temperature (density altitude), and you can "true" the output to account for velocity differences that occur due to differences between ambient temp now versus ambient temp when you zeroed.

However, unless you have accurate velocity information to send off to generate the specific wheel for your load, it's a case of "garbage in, garbage out".

[Wise_A]

Originally Posted by BB57

The statement posted above that "the most accurate load was rarely-to-never the lowest SD load" suggests that the shooters jumping to this conclusion were basing it on small sample sizes.

Originally Posted by BB57

Now...we can argue about how relevant this is to pistol and revolver shooting, where the ranges are short.

It's a non issue in practical competition, but it may well make a difference in bullseye competition due to both trajectory and recoil induced angle of departure differences that result from greater differences in velocity.

The bottom line is that a load with a lower SD minimizes the magnitude of one of your more important variables, and that is never a bad thing.

Ironically, it's Bullseye shooters that I'm referring to, that tested and confirmed all this.

I'm also a bit confused as to how them finding that the most accurate loads had miserable SDs implies they used small sample sizes. Given that these were the findings of many shooters, using many guns, chronographs, bullets, powders, cases, conditions, and so on, I'd say their sample sizes were more than sufficient.

But hey--pop over to their forum and ask 'em yourself if you don't believe me.