ES is it really that important?

[jungle_re]

For year on hand load when using the chrono id be looking for the best speed possible with the smallest ES. Fairly straight forward logic, speed equals less drop and a more consistent speed equals less vertical range so better accuracy.

 

On this credence all my load development was based.

till now! Some strange results! Worked up a simple load in the 300 win mag, pushing a 208gr amax at 2904 avg but with a crazy ES of 115! I'd expect it to not shoot well at distance as it was only managing around the inch at 100yrds. Only had chance to take it back to 500yrds (working out drops for a few other rifles also) but I was surprised that once I had it dialled in every shot I could ring the 4x3" plate. It's confused me slightly and tiff and I had a chat in which he also said his 338 lap mag also had a triple figure ES yet held consitantly 1 Moa out to a mile. Looking forward to seeing how it behaves in the next 1000m additional range.

 

any idea why this would be? If this is the case then is there any point at all in chasing a tight ES?

thougts and comments sought

[gbal]

Slightly hard sums...optional,for both knowledgeable,and challenged ! Just skip to the last para ****for the key points relevant to the ES issue raised in this post....(rest is a rainly day attempt to explain for those not hot on descriptive stats..tempted to just delete....but it's like "Average family in China has .9 kids"-well,none at all have .9 kids (impossible) but the vast majority are limited to one,and a few don't have even one...hence a 'correct' but totally "unrepresentative" average of .9...due to change,of course...to somewhere betweeen 1 and 2 .

 

 

(((Sampling -better yet "representativeness'-is always an issue,even in simple 3 or 5 shot 'groups'-much mischief can arise when numbers are not 'representative' enough ( some variation is a given).

ES per se isn't very informative- it is the extreme spread (honest!) ie the difference between the highest and lowest MVs in the string/sample.....it could be very large,but as it's only from two shots,it is possible-if a little unlikely-that all the other shots are the same/very close......then it's no surprise that almost all those POIs are consistent.....but the high/low are way off.

 

The SD is more informative (Standard Deviation-the average amount by which a shot deviates from the mean...though again,that can be skewed by a few extreme (actually even one) measures...." a flier"..

 

The eyeball test is useful-is there consistency in the string of shots-with only a very few wildly different?

 

3100,3000,3000,3000,2900 ES 200 and SD 80 but might well give 3 very close shots,and two different 'fliers"-one high,one low ....

 

Key is consistency,and neither ES nor SD can always capture that in detail-and just like an average,can work out the same for very different distributions of measures (see below).

 

Just for some scale comparison, a 6mm 80g @ 3000 fps will give a vertical change of only .16 inch at 300 yards for a 10 fps MV change-though calibre etc matter,and at distance vertical does become a real issue if MV fluctuates. Fclass shooters will know the larger effects of bigger calibre sat distance.

 

Maybe your shots are quite consistent,with only a few 'fliers', despite the fliers that would come from a longer string.

 

It's a probability thing-you need a big sample-or inspection of internal string consistency beyond the 'summary' average type SD and ES,useful though they can be-when low (then mvs have to be similar).BUT if they are high-then the above comes in,maybe....and a closer look is needed.

 

3002,3000,3001,2999,3001. Average 3000-which represents each mv very well (SD is 1,ES is 3).POI consistent.

2700 3200 2800 3100 3200 .. Average 3000 which does not represent any mv at all well (SD is 200,ES is 500).POI all over the place.

 

3000,2900,3010,2990,2995...Average 2979 (SD 18,ES 20)....POI pretty good,but one will be well out-the 2900 shot).

 

You really need quite large samples though for effects to be reliably clear in short test strings....and human's tend not to think that way....heads/tails/heads/tails/heads/tails....well,not really ....runs of two,three,four of either are not unusual,even though on a large number of tosses it evens out to 50/50 heads/tails....

4 heads in a row is not good evidence of bias, nor heads,tails,heads,tails good evidence of no bias...yet.

 

three swallows don't make a summer...or a drunk...(sample too small.....))) :-)

 

****SO- you need to check on the real spread/consistency-or not-of readings...a high ES may coincide with much variation in the loads,but might only indicate two extremes in otherwise consistent mvs.( Compare two strings of shots : 2800,3000,3000,3000,3200- versus 2800,2900 3000, 3100,3200..both strings of shots have the same latge ES of 400,but the second has much worse inconsistency)

A low ES is always indicative of consistent mvs.(3000,3005,3005,3010,30015 MS is only 15). The same is true for SD-low is good-consistency,very probably...and high indicates -even more probably - inconsistency. In the limit, High ES and SD might only diagnose one or two way out mvs, in otherwise consistent loads(and the more so the smaller the sample),so need to look closer. Low ES and SD are much more representative of all the MVs-they are likely consistent.So,which is it here,given high ES-lots of big variation,or just a few/couple that are way out?

 

 

 

gbal

[Chris-NZ]

"Jungle"-

 

If you rifle exhibits what is known as "positive compensation", you may not get the expected elevation diffs that the chrono suggest. My deceased gunsmith (who knew a LOT) talked about it occasionally. Basically, the theory goes that some rifle systems deliver the lower velocity shots with a variable upward whip of the muzzle as they exit. I haven't seen a detailed scientific analysis of this but it has appeal in explaining why slow shots don't end up going low.

 

It's a characteristic attributed to No.4 .303s at long distance. Personally I never had issues with a .308 at 900yds+. Possibly just nostalgic thinking of old fullbore shooters

[ds1]

I guess you would need a considerable sample to establish any reasonable level of significance. Also worth trying 2 matched chronographs, one at the muzzle and one at distance (500m)if you can without shooting it. The theory is high muzzle ES is not seen at distance as velocity is lost quicker the faster a bullet starts....hence smaller ES and relatively smaller groups at longer range than at 100m. Ballistic calculator suggest about a 25% - 30% reduction in ES from 500 to 100m with 155g scenars comparing mv of 809 m/s v 839 m/s.

 

Other than that chrony error can add up to a lot - try putting 2 chronographs back to back.

[Murph]

"Jungle"-

 

If you rifle exhibits what is known as "positive compensation", you may not get the expected elevation diffs that the chrono suggest. My deceased gunsmith (who knew a LOT) talked about it occasionally. Basically, the theory goes that some rifle systems deliver the lower velocity shots with a variable upward whip of the muzzle as they exit. I haven't seen a detailed scientific analysis of this but it has appeal in explaining why slow shots don't end up going low.

 

It's a characteristic attributed to No.4 .303s at long distance. Personally I never had issues with a .308 at 900yds+. Possibly just nostalgic thinking of old fullbore shooters [/quote

 

Plus 1 on this. I remember reading an article on this somewhere- cannot remember where!. I had the same thing happen to me with a load I had for 53gr vmax in 223. It had a es of 70 fps or thereabouts but shot really well at 100 yards. Best group of 0.180 inches at 100 and under half moa at 600 yards. On paper it should not have shot like that, but infield conditions it told a very different story. Its was a ice surprise to say the least. But then I had to change powders but luckily I get similar results with a lower es with my new powder. Strange to say the least.

[Andrew]

I have always contended that small ES and resulting Standard Deviation were more an indication of the quality of the ammo, not a guarantee of accuracy. ~Andrew

[brown dog]

Worked up a simple load in the 300 win mag, pushing a 208gr amax at 2904 avg but with a crazy ES of 115!

 

Have you worked out the theoretical vertical dispersion that could cause at 500?

 

ie have you actually quantified what the vertical dispersion could look like with that mv spread?

..... in your ballistic prog compare the impact heights by simply changing the mv run it at 2847, 2904 and 2961 - how do the impact height numbers look in inches at the target?

 

(noting also gbals valid points on SD and 'quality' as well as the, quite possible, 'jump' ideas behind the barrel whip theorists, but I'm always "hoofbeats generally mean horses, not zebras" ... I'd want to quantify how big this mv vertical may be just at the theoretical number crunching level before looking for the zebras )

[gbal]

Just reading Matt's points,I was pondering similarly,though not wishing to be too provocative on a (damp) Sabbath morning.

Much mischief can follow from chrono errors,and imprecise 'fall of shot' data,all fed into a ballistic calculator,often-evidenced on here-with dubious accuracy on all parameters.

 

So,assuming that there are good samples-not just 3 shots-and fall of shot is reliably measured (not "three shots at 600yards,one off the bottom the card,so averaged the other two holes" ),check out Matt's suggestion-see what the 'theoreticals' are..." go compare,...."

 

I was wondering about such relatively high ES figures- a quick look suggests something like 3 to 4 grains of powder would be needed for a difference of 115 fps in 300wm....ballpark,depending on bullet,that's for 180 ish-and the 223 with 70 ES might differ by .6g of powder. Those are improbably high 'reloading errors'..

I'd hear the horse's hooves here as chronic chronos ( at least enough duff readings-two is minimum- to give a non representative high ES ).Check is look at the consistency-or not- of the actual fps readings,all of them (as earlier discussed,statistically). ES etc is not a lie,but can be very misleading.High SD even more so.But we don't have SD data here.....That said,zebras do exist...

 

(Just a secondary point-- if the rifle was shooting an inch at 100,hitting a 4 inch plate at 500 isn't too surprising,especially on a casual measure,no indication of 100y being vertcal or not, and not too many 500y shots?

 

Tiff's 3 figure ES yet 1moa to a mile is more perplexing,admitedly,but that ES may be misleading....or just possibly fps vertical is attenuated in very heavy bullets-though the f classers really care about consistent fps,as the major reducible error component in cartridges.

 

"Positive compensation" is worth a look.....and Geoff Kolbe,onetime chief honcho at Border Barrels,has done just that-well a good hard stare,really,as his excellent resume of work to date shows there is very little indeed real hard evidence from actual test firing,though some theory and speculation, since at least the early 1900s.

The central notion seems to accept that barrels vibrate in harmonic motion when fired,and so the muzzle moves up and down relative to central rest position,assumed by the sights. As I understand it (see in a moment) this could mean that a fast bullet exits as the barrel is rising,but a somewhat slower bullet might exit a tad later,when the muzzle is higher, so the slower bullet shoots higher, and this compensates for its increased drop compared to the faster bullet. (I'm left wondering if it might all just as easily reverse on the down twang ,and the advantages be reversed...indeed it seems rather variable,all round,but the principle has some merit ).

Kolby does note that 'positive compensation' is sometimes,but by no means always shown.....

 

But he then carries out an elegantly designed-and carefully measured- experiment with a 22rf barrel-which initially shows considerable vertical with Tenex at 50 yards. Adding a weight (200g,calculated) to the end of the barrel damps down the vertical to a round hole,comensurate with the predicted dispersion-random-of the Tenex ammo.....

 

It seems that removing-or better yet-compensating the barrel vibrations do indeed reduce almost all of the extra vertical dispersion.

 

Snag is of course,weights are not added in the field observations we have to hand,though the use of tuners (weights adjustable near end of barrel) is widespread in 22rf Bench Rest 50 yard shooting-helping validate the idea. Browning did try the "Boss" tuner on cf,but what happened to that....??

The Kolbe report is quite excellent though (OK,I am disposed to like clear, scientific reporting,when I need data to check out what mates say over a pint) :-)

 

The specific explanations for the original post observations remain uncertain,and I am prohibited by my religion from serious thinking on the Sabbath. :-) :-)

 

Kolbe,Geoff "The Vibrations of a Barrel Tuned for Positive Compensation" 2011

 

 

gbal

[ds1]

At 500m for 338 Lm 250g Scenar bc. 675 @ 880 m/s drop is 134.4 cm, @ 910 m/s drop is 123.6 cm.....or about 3/4 moa in old money.

 

The high bc is efficient so relatively little drag, but the 100fps difference at the muzzle is about 85 fps difference at 500m.

 

But 3/4 moa is still to me significant.

[gbal]

Here is another way into this question (does ES/SD matter).

Perfect accuracy (putting shots on target) is rare in real shooting;most ballistics programs are about giving optimal ballistic solutions,even so,even with the best,you will miss sometimes,because of bullet dispersal. Bryan Litz' WEZ analysis shows the effects on one disperser-variable MV,expressed here as SD of MVs.

 

Specific results will vary with the ballistics,and shooting parametes calibre.MV,bc and wind/range error -but shooting error-wobble etc-is excluded).

 

In this example,we have a 30 cal 175g SMK @2600fps;

and assume very good shooting parameters (wind speed judged +/- 1mph,range +/- 1 yard,.5moa rig

 

Table entries are % hits .

 

Range y SD 10 fps 15 fps 20 fps

 

Target disc 5" / 10" 5" / 10" 5" / 10"

 

500 93 / 100 91 / 100 83 / 100

600 73 / 99 67 / 98 55 / 96

 

700 50 / 94 43 / 88 34 / 79

800 34 / 78 26 / 69 20 / 54

 

900 21 / 57 15 / 48 11 / 37

1000 13 / 39 9 / 32 7 / 24

 

 

So hit % with SD 0f 10 fps on the 5" disc goes from 93 at 500y,dropping to 13 at 1000,and for the 10" disc it drops from 100 to 39% hits; and so on -clearly getting considerably worse as SD increases to 15 and 20 fps. (sorry for condensed table)

It matters-but there are some issues-larger ES/SD for MV wll cause vertical dispersion,whether it 'shows' depends on target size and hit % probability....on fig11s eg there is a lot more vertical hit area to play with...and a hit is a hit..down. As target size diminishes,the target tolerance for dispersed shots reduces,quite rapidly.....as above,even under very good overall conditions,and modest SD changes.

 

A related issue is that for large targets (fig 11s) any increase in hit % is greater for relatively slower bullets,but for small targets the hit improvement is greater for faster bullets....because for the large/fast combo,the hit rate is already so high,there is much less room for improvement by by better MV consistency. But for the smaller targets,much of the low hit rate is caused by inconsistent slower bullets,with more vertical,so improving SD has more effect on hit %.

But let's not get too complex.

 

gbal

 

I am arguing that the above is not 'thinking' on the Sabath,merely passing on Bryan Litz analysis.

 

I have said 20 "Hail Bryans".

 

I have checked the ballistic sabbath guardians,and had the same somewhat limited reassurance as the honeymoon couple who asked the Hebridean minister if it would be OK to have conjugal relations on the Sabbath,and got a somewhat grudging:

"Well,in your circumstances,I suppose so,provided you don't enjoy it."

 

:-(

[brown dog]

I think the simple bottom line here will be found to be a consequence of using very high BC bullets (eg Will's 208gr .30 AMax) which means that ES matters less as distance increases because 'the graph lines converge'.

 

Eg David, how would your numbers look for a 300gr Scenar - even less vertical spread with the same ES?

 

At 500m for 338 Lm 250g Scenar bc. 675 @ 880 m/s drop is 134.4 cm, @ 910 m/s drop is 123.6 cm.....or about 3/4 moa in old money.

The high bc is efficient so relatively little drag, but the 100fps difference at the muzzle is about 85 fps difference at 500m.

But 3/4 moa is still to me significant.

[bradders]

I think the simple bottom line here will be found to be a consequence of using very high BC bullets (eg Will's 208gr .30 AMax) which means that ES matters less as distance increases because 'the graph lines converge'.

 

Eg David, how would your numbers look for a 300gr Scenar - even less vertical spread with the same ES?

 

Not an expert by any stretch of the imagination, but looking at Davids figures you have an average of 895 m/s and therefore a SD of only 3/8 MOA.

And to me that ain't bad

[ds1]

BD, yes 300g scenars would look better for ES but as always there are other compromises, mainly my 26" barrel and wanting a flat trajectory for ukd targets out to 800m for tactical matches. Also I don't feel inspired to test Hornady 285g melting tips when Scenars shoot well

If I get my ass kicked at 1km next month I might not be so lazy.

[jungle_re]

10 shots were put over the magnaspeed. I don't have any doubt of the unit and as it was being used with two other rifles which worked so I'm ruling that out. As for the re loading each charge was weighed on a rcbs 10/10 beam scales with targetmaster power trickler so each powder charge should be as close as cock on as possible. 15 shots once drops confirmed with 1 miss on the gong.

 

Pushed the figures though on the applied ballistic kestral with the lower and higher figures of the ES. This works out as a 4.18" vertical difference. The gong was 3" wide and 4" high surely more misses should have been experienced if that Es in real world equated to such a wide vertical displacement.

[brown dog]

Pushed the figures though on the applied ballistic kestral with the lower and higher figures of the ES. This works out as a 4.18" vertical difference. The gong was 3" wide and 4" high surely more misses should have been experienced if that Es in real world equated to such a wide vertical displacement.

 

68% of your shots would have fallen within 1 SD of the mean mv's impact height - what vertical dispersion does your 1 SD mv variance give from that centre point? Around +/- 0.5"?

94% would have fallen within 2 SD - how high is that? Around +/- 1" ?

 

If each shot is a dice role, and you only fire a few, which ones are you most likely to see? 94% of your shots will be around +/- 1" with the data you've given.

[jungle_re]

One sigma is 42.53 equalling 1.61" leaving 0.39" in the error budget to still hit.

I'm an average rifle shot so to me the ammunition must be producing less vertical shift than the figure would suggest.

 

It's interesting to discuss but I guess it will be proven one way or another once that distance is doubled again

[brown dog]

One sigma is 42.53 equalling 1.61" leaving 0.39" in the error budget to still hit.

 

 

? With an ES of 115, I'h have fagpacketed your SD as 29ish and an inch ....not enough data here really... will be interesting to hear the outcome........but I think it's inevitable it will get -relatively speaking- better with distance as the lines continue to converge. I'm still not seeing zebras!

[gbal]

Matt's (6.23 post) points are cogent-68% of all data is captured within 1SD,and 94 % within 2 SDs.

((And 7.57 post-added;key issue-just not enough data reported,understandably,as the 'statistical' explanation might not have been predicted-when will we all accept shooting is probabilistic !)

Without the actual measured MVs for all the magnetospeed measured shots -I accept they are probably pretty accurate(and sufficient powder differences seem unlikely-my earlier point)-we just cannot work out the actual SD. Ten shots may be a marginal sample too,if there is a ES as reported...

 

But the approach is to llok at dispersal from the data ,compared to target size...fifteen target shoots is decent data,probably. The WEZ hit % figures are illustrative of the effects of tightening the SD of the MV.

 

But without the actual calculated SD for all the MVs,not a lot can be said-if indeed it is somewhere in the 40s,that means a lot of variation between shots-probably,but the only figures given-the two ES velocirties can't progress us much.

 

Probably,your comments are fair-but it's taken a long time to get near the SD for the MVs-the SD value is one critical bit of evidence- (probably,because if it's low.....the world is bent in some unknown way. At least we have some values for the drop with the two extreme ES values,which are tight but not impossibly so...under very good conditions.

Whatever,maybe it all helps make the point that you need much more data than we modally have in these cases ,to do any 'significant' analysis....one reason WEZ etc are based on minimum 100 simulations.

Human's are vulnerable to assumed effects in very small samples-not a criticism,maybe the way the brain is semi-hard wired-yoy don't get a hundred escape attempts from serious predators...

 

Goodness knows where the large ES /SD MV dispersion comes from,if it's not the powder!

 

But good shooting,anyhow!

 

gbal

[ds1]

I still think this is best guess as sample sizes are small and belief in accuracy of commercial chronographs without testing / comparison is also for me not best practice. An eye opener for me was at a comp there was an opportunity for people to test their loads over a doppler radar and get actual traces from it. To get the information you want, you need that sort of technology and a fair bit of testing. Only problem is a doppler radar cost cca 200,000 quid.

[jungle_re]

I provided the one sigma value above George it's in my post as 42.53. Not a guess calculated from the 10 shots taken.

Fair call on the chrono but you have to base it on something and as two other rifles were chronic at the same time did not exhibit the same level of spread I think you can rule that as as much as possible. It couldn't be a systematic error as this would have propagated in the other strings. For example the ar over 10 shots with factory FN ammo showed an ES of 23. Shot in the same hour, in same condition as the win mag.

 

The values don't seem to show any huge blunders that are I ediatley well outside all other measurement nor do they show an increasing or decreasing trend. I guess I could reject all reading that fall outside of one sigma and use those as the caculation though 10 is a small sample unit to do that too. With old score surveying this was carried out for each observation but a minimum of 16 rounds for a tertiary survey and multiples of 16 increasing up to first/special order. Can't remember the exact maths behind the reasoning but it's to do with 16 and multiples yielding better results in later least squares processing. I would want to dismiss any of those data points out of hand as none look like they are outrageous misrepresentations. The chrono is what it is and as one of the best hobbiest units cross used with other loads and weapons I can't see what I could do to improve its results?

 

To be continued......

I'll update for interest once I have some more shot data, though that may be a while

[gbal]

Will, Fair comment-let's accept the chrono data-magneto s usually quite reliable.And you have reasonable co-roboration from other rifles.I never intended this point to be central-just neded confirmation-ditto the powder-as I said,it would have to be out big time,and that is implausible....at least for most loads.

 

The basic point about not giving the actual MVs though was serious-without those the SD can't be checked,nor any vagaries in the actual velocities noted...there n=may be none,of course...and you suggest you have looked. So we progress,though mainly by eliminating some possible explanations !!

 

The 'stats' issue -maybe now to be rejected (sample size ,though) runs like this (in case anyone is trying to follow !)-no one statistic (mean,mode,median,ES SD etc etc can hope to capture all the information) and it matters,thus:

 

Suppose we load two batches of 100 cases each identically and chrono them :ES is 100 fps in each:

 

Batch A chrono thus : one shot at 3050,one shot at 2950,and 98 at 3000

 

Batch B chrono thus ; ten shots at 3050,ten shots at 2950 and 80 shots at 3000

 

We go shoot at 500 y gong,4 inches tall (and we know,from analysis like Davids,that this load shoots 4 inches different vertical for 100fps difference in this approx MV range (more like 308 than 300wm,but live with it meanwhile,just to make the sampling/consistency -ES can mislead-point.)We might expect:

 

Batch A to shoot two overlapping 'groups' of 50 shots each,about 4 inches vertically apart,centre to centre,so with a 4 inch vertical dispersion,and maybe one low,and one high flier.We might expect 50% hits on the 4 inch gong,but NOT 14/15

 

Batch B to shoot a tight group of 80 shots,with some high and some low a couple of inches out,and a very high hit rate,probably 80% plus,as some fliers will be hits-just,and 14/15 is very likely.

 

Just cross refering to (my) post*10,at 500 yards for a good 308 rig/shooter,the hit rate on a 5 inch disc drops from 93% to 83% as SD goes from 10 to 20 fps,and at 600 it goes from 73% to 55%

 

Clearly SD matters-it measures consistency. ES may or may not be much of a guide to consistency-low ES good,high ES tells us little,as in Batch A v Batch B.

The 300 wm will do better at longer ranges,but the same general principles apply-and SDs in the 40s just accelerate the decline in hit %,over a reasonable string.We still are short of an explanation for those kinds of SDs,of course.

Note it does take quite a bit of dispersion to get misses from such as the 300 wm at moderate distances,like 500 yards-the ballistics tolerate abuses quite well,by some analyses,the 300wm with 230 BergerHybrid is pretty well up with the 338L in hit%..

 

....so get some,get 15/15 hits and we can consider the stats as a deviation for wet dark days! :-)

 

g

[brown dog]

I provided the one sigma value above George it's in my post as 42.53. Not a guess calculated from the 10 shots taken.

 

 

My point was, SD from only 10 shots is a bit dodgy; the SWAG rule of thumb to estimate what an SD should look like is to divide the ES by 4.