What should a tuner be expected to do?

[Biologist]

Hey JEDelta: I just had a brainwave on how to re-analyze your data from post #155: Now that you are using OnTarget to measure radius from a central point of aim, you don't need to "group" the groups, and you don't need mean radius per group, and can save time and money in your tuner testing. You can use the single radius per round as the sample.

Groups: The traditional 5-shot group is the single sampling unit (1 group = 1 sample) per tuner setting (treatment), and it provides one number: the group size. But it requires 5 rounds. 8 groups requires 40 rounds to produce just 8 samples.

With OnTarget measures for the radius of each shot, you don't need to analyze by group for understanding dispersion patterns affected by a tuner.

Radius: Each round shot generates a radius from the central point of aim. The radius from each shot is the sample for each tuner setting. In your data in post #155, using OnTarget, you generated 40 radii samples per tuner setting, not 8. There is no need to group the radii into 5-shot mean radius stats. You have a sample set of 5 tuner settings x 40 rounds for a total of 200 samples! That is a robust data set that exceeds the magic number of 30 samples per treatment that the textbooks say is needed for a solid sample size.

If you have the time, it would be neat to see the 5 x 40 table of radii from post #155, and the ANOVA for it, and a pair-wise comparisons for the two best looking tuner settings.

 

[grauhanen]

With OnTarget measures for the radius of each shot, you don't need to analyze by group for understanding dispersion patterns affected by a tuner.

A serious question for the analysts about trying to "understand dispersion patterns affected by a tuner". Since tuners can't correct POI irregularities of individual rounds, how can this be accounted for when analyzing tuner results? In other words, when bullets don't go where they should, how can analysis of tuner influence on POI be reliably achieved?

This question is asked because a .22LR bullet's point of impact may not always be what MV or ballistics calculators may predict. The reasons for the mismatch are known and can be seen regularly with ammo that performs inconsistently. Indeed it's more of a problem with ammo that is not consistently consistent, such as .22LR entry level match ammos such as SK varieties, the less expensive Eley ammos, and similar RWS offerings. The problem can also occur with more expensive grades, but perhaps with less frequency.

 

[JEDelta]

I think measuring from POA wouldn't necessarily provide good data. There are many factors that could affect the POI: ammo, shooter skill, prevailing weather conditions during the time the group was shot, etc. It is important to note that because of the order the groups were shot, there was over an hour of time between the first and last groups of each tuner setting.

My original interpretation of Biologist's post was to break down the mean radius of each group into five separate radii, from the group center, for each shot, as more data points should give better data. This is just a different way to look at the mean radius.

When the data is broken down this way, it does provide stronger results, but perhaps still not as strong as the results for group size.



The only pairing that is as strong is the best and the worst (4 and 6)

6 vs 10 p = 0.191764
6 vs 4 p = 0.018511 (p = 0.058263 when just using the 8 data points from mean radius)
4 vs 2 p = 0.088946

Pretty much every way to analyze the data does show a statistically significant difference between tuner settings 4 and 6.
I feel confident using these two tuner settings to compare to a baseline without the tuner. Hopefully, I can get some good data on Saturday.

 

[Biologist]

A serious question for the analysts about trying to "understand dispersion patterns affected by a tuner". Since tuners can't correct POI irregularities of individual rounds, how can this be accounted for when analyzing tuner results? In other words, when bullets don't go where they should, how can analysis of tuner influence on POI be reliably achieved?

If I can borrow and paraphrase from Dr. McCoy in the original Star Trek TV series: "Damn it Jim, I'm a Biologist, not a Ballistician!"

Short answer: I don't know.

Slightly longer answer: Its all about sample size and replicates. With large enough sample size and enough replicates, statistical methods exist to provide clear answers.

Longer more boring version below: I will attempt a guess since I don't really know:

I think you could get the definitive answer from the folks who run the tunnel testing labs. I think they are testing for your exact question.

In their lot testing for customers, either with or without a tuner attached to their rifles, they are generating cloud patterns of hits around a fixed aiming point. The percentages of hits within concentric rings (rings being the same as radii) generates density stats per ring (radius). And they will have worked out what a sufficient sample size is.

Since world class smallbore competition uses decimal scoring, then I think the best score indicates the tightest pattern (smallest group) around the central point of aim. The best lots for the rifle are going to score the highest decimal score. And it indicates the lesser percentage of bad round fliers per lot.

Data generated in the tunnel testing method shooting in electronic targets can be portrayed as a competition score: 20 (or 25) rounds would be the rounds per replicate tuner setting since it emulates what the shooter is going to do in matches.

The fliers (caused by those bad rounds) are outliers in the cloud pattern of hits. With or without a tuner attached, there is no need to define exactly what a flier or bad round is. The best score speaks for itself and wins.

Tuner setting replicate scores are statistically compared to each other, and to scores with no tuner attached. This should determine if the tuner is a net benefit for that rifle-ammo combo, and if so, what setting is optimum.....I think.

Outside in the wind shooting in normal match conditions, obviously the degree of uncertainty increases. I think a larger sample size with more replicates would be needed.

 

[grauhanen]

If I can borrow and paraphrase from Dr. McCoy in the original Star Trek TV series: "Damn it Jim, I'm a Biologist, not a Ballistician!"

Short answer: I don't know.

Slightly longer answer: Its all about sample size and replicates. With large enough sample size and enough replicates, statistical methods exist to provide clear answers.

Longer more boring version below: I will attempt a guess since I don't really know:

I think you could get the definitive answer from the folks who run the tunnel testing labs. I think they are testing for your exact question.

The fliers (caused by those bad rounds) are outliers in the cloud pattern of hits. With or without a tuner attached, there is no need to define exactly what a flier or bad round is. The best score speaks for itself and wins.

Tuner setting replicate scores are statistically compared to each other, and to scores with no tuner attached. This should determine if the tuner is a net benefit for that rifle-ammo combo, and if so, what setting is optimum.....I think.

While Bones had an excellent point, Spock may have pointed to a logical question.

Without knowing which rounds are outliers and which aren't -- that is to say, which rounds that don't go where their MV predicts and which ones do -- is it possible for statistical methods to to provide clear answers? The answer must be that it's not -- unless accurate chronograph information is included, which, of course, likely risks adding bewildering complexity to the analysis.

Why? When someone analyzing target results knows the muzzle velocity of the rounds, he can determine what, if any, influence a tuner may have on them. This is because, when rounds have a POI directly related to MV, an effectively adjusted tuner will alter the POI of rounds based on their MV. Groups will get a bit smaller. With an effective tuner setting, individual POI will appear to be better than MV alone would predict as slower and faster rounds strike the target closer together than they would otherwise..

When he doesn't know the MV of the rounds in question, their POI can't be analyzed with respect to the effect of a tuner because he can't be certain whether or not they go where their MV predicts.

Good, consistent ammo will have a high frequency of rounds going where their MV predicts. Inconsistent ammo will have more incidences of rounds not going where the MV predicts because it has more rounds that are even less perfect than better ammo. Entry level .22LR match ammo is typically insufficiently consistent for effective tuning. This is why tuning with inconsistent ammo is generally unproductive.

Bryan Litz, who is certainly no stranger to statistical analysis, found that when he used non-lot tested entry level match ammo in his .22LR tuner testing he couldn't find evidence that tuners had made a genuine difference in his results.

 

[JEDelta]

Entry level .22LR match ammo is typically insufficiently consistent for effective tuning. This is why tuning with inconsistent ammo is generally unproductive.

Bryan Litz, who is certainly no stranger to statistical analysis, found that when he used non-lot tested entry level match ammo in his .22LR tuner testing he couldn't find evidence that tuners had made a genuine difference in his results.

After a third batch of testing, the findings suggest it would be hard to argue otherwise. I'm actually using the same ammo as he did in his test, SK LRM.

During the week, I did have to remove the action from the chassis as the pivot pin for the bolt release worked its way loose, but it was re-torqued to the same specs as before. I am not sure of the exact effect that may have had on testing.

This time I shot 12 groups of five shots, alternating one group per setting until I had twelve groups each. Then I took the tuner off and shot twelve groups, ensuring I broke my cheek weld between each group to have similar conditions to the tuner groups.

I had a complete reversal from last week. Setting 6, which was the worst setting from last weekend ended up beating setting 4 (previously the best setting) across the board.



Between any of the pairings, the lowest p-value was when comparing the mean radius between tuner settings where p = 0.121287
The second lowest pairing was p = 0.218818 for group size between the tuner settings.

I still think it shows that the tuner is definitely having an effect on the groups. However, since the groups without the tuner averaged right in between the two tuner settings, neither tuner setting is statistically significantly different from the groups without the tuner.

When you are looking at potential group-size improvements measured in hundredths of inches. It certainly doesn't help when you are working with ammo that results in SD's that are many times larger than that.

Anyway, it was a fun experiment, and I appreciate the feedback that everyone provided.

EDIT: I figured I would give a reference for how much of an improvement can occur without it being statistically significant. I got a new IBI barrel which I tested after this test. Visually and by every single measurement, the groups from the IBI barrel were better than the MTR barrel. And yet the p values were 0.133702 and 0.148542 for mean radius and group size when compared to the MTR barrel without the tuner. Even though the results aren't statistically significant, it is pretty clear the IBI barrel shoots better.

Based on the data from the last two tests, the improvement of group sizes would have to be over 0.1" before it becomes statistically significant. We are talking about improvements of over 20%. Based on the previous discussion, those kinds of gains are not expected, so perhaps statistical significance should also not be expected.

 

[emerson]

After a third batch of testing, the findings suggest it would be hard to argue otherwise. I'm actually using the same ammo as he did in his test, SK LRM.

During the week, I did have to remove the action from the chassis as the pivot pin for the bolt release worked its way loose, but it was re-torqued to the same specs as before. I am not sure of the exact effect that may have had on testing.

This time I shot 12 groups of five shots, alternating one group per setting until I had twelve groups each. Then I took the tuner off and shot twelve groups, ensuring I broke my cheek weld between each group to have similar conditions to the tuner groups.

I had a complete reversal from last week. Setting 6, which was the worst setting from last weekend ended up beating setting 4 (previously the best setting) across the board.

View attachment 700372

Between any of the pairings, the lowest p-value was when comparing the mean radius between tuner settings where p = 0.121287
The second lowest pairing was p = 0.218818 for group size between the tuner settings.

I still think it shows that the tuner is definitely having an effect on the groups. However, since the groups without the tuner averaged right in between the two tuner settings, neither tuner setting is statistically significantly different from the groups without the tuner.

When you are looking at potential group-size improvements measured in hundredths of inches. It certainly doesn't help when you are working with ammo that results in SD's that are many times larger than that.

Anyway, it was a fun experiment, and I appreciate the feedback that everyone provided.

EDIT: I figured I would give a reference for how much of an improvement can occur without it being statistically significant. I got a new IBI barrel which I tested after this test. Visually and by every single measurement, the groups from the IBI barrel were better than the MTR barrel. And yet the p values were 0.133702 and 0.148542 for mean radius and group size when compared to the MTR barrel without the tuner. Even though the results aren't statistically significant, it is pretty clear the IBI barrel shoots better.

Based on the data from the last two tests, the improvement of group sizes would have to be over 0.1" before it becomes statistically significant. We are talking about improvements of over 20%. Based on the previous discussion, those kinds of gains are not expected, so perhaps statistical significance should also not be expected.

View attachment 700376

My experience with SKRM leads me to believe it does not have the consistency required for precision in general, nor for tuner testing in particular. Perhaps a different case would give different results. I will not buy more.

 

[grauhanen]

My experience with SKRM leads me to believe it does not have the consistency required for precision in general, nor for tuner testing in particular. Perhaps a different case would give different results. I will not buy more.

Your experience is one that confirms that effective tuning with entry level .22LR match ammo is generally unproductive. A different lot of SK, if there was one that shot exceptionally well, might make successful tuning possible. The difficulty lays in finding such a lot. While there are many posts on forums claiming success with entry level ammos, they may be the product of insufficient testing.

Experienced shooters with considerable successful tuner experience invariably assert that the ammo must be consistent. The question of exactly how consistent the ammo needs to be is one that seems addressed too rarely.

For example, is ammo that can satisfy this forum's sub-half-inch five-shot groups challenge consistent enough? That's likely too low a bar. Many successful challenges have been done with entry level ammos such as SK.

How about ammo that can shoot ten-shot groups at 50 that are sub-half-inch? That is a somewhat higher bar.

Below is an example. Five ten-shot groups at 57 yards, all under half-an-inch. Will this ammo be good enough for tuning?



Is the bar high enough? Is this ammo good enough to use for successful tuning?

To help answer the question, consider first what a tuner does when it's doing what it should. In short, according to positive compensation, it helps alter the exit timing of slower and faster rounds so that they arrive on target with MV variaration between rounds mitigated. That is to say, when the tuner is effective, rounds with different MV's should have a similar POI.

Of course a tuner can't work miracles and alter MV's that are far apart. In other words, when the MV variation between rounds is too great a tuner can't overcome the difference. For a tuner to do its job, it's best if the ammo used has a small extreme spread.

The ammo used in the example shown above does indeed have a small ES. The box of 50 had an ES of only 22 fps. That's quite good.



Will that do the trick? Will this ammo be "good enough" for tuning?

This question will be left for readers to suggest answers. The information on the picture should be a clue.

 

[Biologist]

Hey JEDelta: Nice work again! I really appreciate the tuner testing effort, and the data you are posting.

In Post #166's data, I looked at the bare barrel (no tuner baseline) groups data MTR vs IBI. I think there is a significant difference between those group sets. My stats knowledge is not enough to know why the ANOVA showed no significant difference. I will guess its because the sample size is too small for the ANOVA test to be sensitive enough to the differences, and maybe the variance is too large? Twelve 5-shot groups is alot of samples, but stats-wise, maybe not enough for the width of the range of the distribution and variance?

But there are other ways to visualize this data that perhaps are better. I tried two methods below:
- Pairwise bar graph, which is straight forward in Excel to do.
- Box and Whisker graph, which Excel can easily do, but a little trickier for the interpretation, but I will elaborate below to simplify this. The Box & Whisker analysis is very useful for all kinds of comparisons between two data sets, although it does not calculate significant difference. However for shooting a rimfire match for score and what you want to chose for barrels, ammo or tuner setting, I think the B&W analysis is rock solid. It identifies outliers. Once you learn the tricks in Excel, its a breeze to do.

Pairwise bar graph comparison of the barrel's groups, no tuner: I ordered the groups from smallest to largest. In every case the IBI barrel shot smaller groups than the MTR:



Box and Whisker Graph: For a how-to Tutorial, this YouTube video is helpful: https://www.youtube.com/watch?v=39lsUsJsc2c
However the instructor does alot of things manually in this video to show you the meaning for every graph element and what the software is doing. These are useful to understand the graph, but once understood, you can skip the manual steps because Excel will make the graphs automatically from your data table with the Box and Whisker choice in the "insert graph" list.

I will briefly explain key B&W concepts below this graph:



Box and Whisker graphs show the quartile ranges and the middle "interquartile range": 0-25%, 25%-75%, 75%-100%, plus outliers if they are present.
I lightened the colours so as to be able to see the numbers and lines.

The box in the middle 25% to 75% is the "interquartile range", which is the middle 50% of the distribution.

The median is the 50% value within the box, marked with a solid line. Note how the median is not always in the middle of the box.

The X is the mean. Excel will show this number's value in the box, but I deleted the number to reduce overlap of the numbers for visual clarity, and besides, its already in the original data table. The median is sometimes a better statistic that the average (mean) for understanding data dispersion.

The whiskers show the statistical max and min, IF there are no "outliers".

The dots above and/or below the whiskers are the "outliers". In Box & Whisker graphs, outliers are defined as more than 1.5x the interquartile range spread (upper and lower bounds of the box), minus or plus the bounds of the box. The tutorial explains the outlier calculation.

Box & Whisker Graph Interpretation:
A wide box relative to narrow whiskers is good. That means less variation in group sizes, and less crazy flier groups on the large end of the distribution.

A narrow box relative to wide whiskers is bad. That means the groups were widely variable in size.

Outliers on the upper end (top above the top whisker) are bad. These are the very large groups well beyond the size of most of the other groups. If you did not pull the shot, or if there was not a bad gust of wind, these could indicate a bad round in the group that caused one of those crazy fliers that is unexplainable.

Outliers on the bottom end (bottom below the bottom whisker) are really good. These are exceptionally small groups for what the distribution was.

The MTR barrel showed outliers above and below, a small box, and wide variation. The IBI barrel showed a wide box relative to group spread, no upper end outliers, and no lower outliers.

Note that the boxes (middle 50%, or interquartile range) overlap in distribution. Both of these barrels can shoot small groups, and there is alot of similarity.

In my opinion, these two graphs above show clearly with this data set, a significant difference in consistency of group size distribution between the two barrels. The IBI is shooting a tighter consistency with mean and median smaller groups. We already knew that from the SD's, but this shows it in an interesting visual.

 

[Biologist]

Grauhanen: That looks like a beauty lot of ammo! That is an amazingly small and consistent ES for the MV. And very nice shooting with those 10-shot groups! Care to tell us what ammo brand that is?

RE question: I would guess "yes" that ammo would be good enough for tuning (assuming the rifle was responsive to the tuner).

 

[JEDelta]

That is definitely another interesting way to view the data. You can view the data and see that it is clearly better, but the box and whiskers chart seems to give a bit of "mean radius" of the data so to speak, while also identifying statistical outliers.

The IBI vs MTR barrel is a bit off-topic, but I think it gives a good example of perhaps the difficulty in proving statistical significance when it comes to tuners. Something can be clearly better but for one reason or another you just can't get p < 0.05.

I applied the same chart to the rest of the data. Tuner setting 6 obviously isn't as good as the IBI barrel, but it certainly seems to be better than no tuner.

 

[Biologist]

That is definitely another interesting way to view the data. You can view the data and see that it is clearly better, but the box and whiskers chart seems to give a bit of "mean radius" of the data so to speak, while also identifying statistical outliers.

The IBI vs MTR barrel is a bit off-topic, but I think it gives a good example of perhaps the difficulty in proving statistical significance when it comes to tuners. Something can be clearly better but for one reason or another you just can't get p < 0.05.

I applied the same chart to the rest of the data. Tuner setting 6 obviously isn't as good as the IBI barrel, but it certainly seems to be better than no tuner.

View attachment 700584

Yes, nice analysis!

I might challenge (with trepidation), with the thought that tuner setting 6 is not as good as the IBI barrel with no tuner. This is because of sample size and the craziness of what bad rounds may lurk in the box waiting to be crazy fliers.

When I think about competitive shooting for score on bullseyes, this is where these comparison statistics can melt my brain! In score there are no groups. At my club's 50m matches we shoot single rounds into 25 mini bulls (a common target format). We score inside edge, and its all about cutting that 10 line and X ring line, and one bad flier in the 9 or 8 ring can lose me the match. Or if I score a perfect 250 points (our 10 ring is generous and 249's and 250's are not uncommon if its not windy), not enough X's can lose me the match (happens all the time). The losses often correspond with the usual frequency of one or two bad fliers per box of my mid grade of ammo. We have half a dozen shooters who all can win on any given day with a perfect 250 and high X count, so it always raises questions about the ammo and bad rounds in the box. Therefore the frequency and spread of the upper 50% and the outliers in the B&W graph are informative and really matter. The upper 50% in the graph are hitting the 9 and 8 rings causing me to lose the match.

In the above Box and Whisker example, note that the box, i.e. middle 50% of the results for the IBI bare overlaps fully with tuner setting 4 and 6. That means on any given day these all have the potential to shoot the same mid-size groups, and beat the others in a score match.

The median line marks the boundary between the upper 50% and lower 50% of group size. Using the median line, note that both tuner settings 4 and 6 shoot smaller groups 50% of the time better than IBI bare did. This example demonstrates the usefulness of the median statistic. Thinking about a score match, 50% of the time tuner setting 4 and 6 could score slightly better. (if this pattern theoretically were to hold over many replicates).

Note that tuner setting 6 has no upper outlier, whereas IBI bare has an upper outlier that is higher. That single outlier for IBI bare could have been a bad round in the box - it likely was because its way out of proportion with the top of the box (75th percentile). So if we ignore IBI's upper outlier, the below the top of the box proportion (75% of the time), IBI bare could score slightly better than tuner 6.

Note that 6 has a wider box with smaller groups sizes than IBI bare, and its median is lower. The median indicates that for 50% of the time (lower half), tuner setting 6 shoots smaller groups than IBI bare. Again applying this concept to a score match, tuner 6 might be a better choice than IBI bare.....but its debatable too.

Tuner setting 4 has a wide ES. That wide upper 25% and upper 50% above the median, can and will lose you a score match, hitting 9,8 and 7 rings. But interestingly its median in the 50% box is lower than both 6 and IBI bare. That means 50% of the time tuner setting 4 can shoot a better score in a match. But Tuner 4 is showing high risk of missing the X and 10 ring, and the risk factor of tuner 4 would cause me to chose tuner 6 or IBI bare for the match. However, with the sample size and ammo, we cannot know if tuner 4 experienced a fluke for having a few more of those bad rounds in the box that the tuner cannot control for POI.

It just goes to show in rimfire how difficult statistics can be for interpretation!

 

[grauhanen]

Grauhanen: That looks like a beauty lot of ammo! That is an amazingly small and consistent ES for the MV. And very nice shooting with those 10-shot groups! Care to tell us what ammo brand that is?

RE question: I would guess "yes" that ammo would be good enough for tuning (assuming the rifle was responsive to the tuner).

The ammo used was RWS Special Match, from the Premium Line of RWS smallbore ammo that includes the top tier RWS R50. RWS factory rates Special Match at 330 meters/second (or 1083 fps), which is the same nominal MV as R50. It's usually priced similarly to Lapua Center X and Eley Team, although recently a Canadian dealer had it available for a slightly better price. Although lot to lot variation of any .22LR match ammo makes comparisons unfair, RWS Special Match may be considered comparable to Center X and Team, both of which are the "low" grade of their respective ammo maker's standard .22LR match ammo.

The chronograph data certainly suggests that the lot is indeed quite good. But no chronograph has ever shown how the .22LR ammo will actually perform. They give an indication how it might be expected to shoot, but they don't give a performance guarantee. As a result, the question remains whether it is in fact a good shooting lot and sufficiently consistent for successful tuning.

Three boxes of this ammo had MV's in the 1092 - 1119 fps range, with a fourth box having a single outlier of 1136 fps. The ES figures for the three boxes without the outlier ranged from 22 -25 fps (SD's of 5-6), while the box with the outlier had an ES of 44 fps and SD of 7.

 

[grauhanen]

The chrony data suggests that the ammo should perfom very well. Unfortunately, MV is only a single factor in determining the ultimate POI of a .22LR bullet and too often it seems like it's not even the most important.

A ballistics calculator indicates that two rounds of ammo with a 10 fps difference in MV, one with an MV of 1100 and the other 1110, will have about 0.08" of vertical dispersion at 55 yards (which is very close to the 57 yards at which I shot this ammo). Consider the vertical dispersion actually seen on the target shown below, which has the vertical dispersion size as well as the ES of the rounds in each group.



Next, compare the vertical dispersion on the target with the expected vertical dispersion predicted by the ballistics calculator. Calculations are rounded for simplicity.

 

[grauhanen]

Clearly the vertical dispersion predicted by ES alone is much less than actually occurs. Why is the actual dispersion much greater than the MV differences would predict?

Excluding the influence of air movement or wind, with perfect ammo and as nearly a perfect a bore as possible, five or ten rounds shot at the same POA should produce a vertical group with a dispersion pattern determined by the MV differences between the individual rounds. This rarely or almost never happens.

There are several possible reasons at play.

One is the rifle/barrel. A poor barrel will disperse shots over a wider area than a good one. Even if it was possible to have rounds identical in every way, it's likely that a poor barrel couldn't shoot them into the same place on demand. With consistent ammo a better barrel will give more consistent results. The barrel is question is a relatively new Anschutz barrel and there's nothing to this point to indicate that it's less than average for an Anschutz-made barrel for a Model 1907. No doubt there are better barrels, but in this case its quality may not be an overriding issue.

Of course, no bore is perfect. When excluding air movement between shooter and target, even with very good barrels and very good ammo, bore imperfection and slightly imperfect chamber/leade/bore alignment help explain the lack of a vertical-only dispersion.

A second and more important factor is the ammo itself, or more specifically, the bullets themselves. Imperfectly formed heels on .22LR ammo will cause inacccuracy and bullets going where they aren't expected. Furthermore, unjacketed, soft lead projectiles such as those on .22LR are usually characterized by imperfect centers of gravity. It's not easy to make them perfect. Center of gravity offsets cause rounds to have unexpected trajectories which produces unexpected POI. When several rounds with imperfect centers of gravity are shot, the dispersion pattern becomes less predictable the further they travel. The more center of gravity is a problem, the wider the dispersion. (This explains why as distance increases some lots of the same kind of ammo do better or worse than others).

When bore imperfection, chamber/leade/bore alignment combine with bullet center of gravity issues, the result is that group dispersion patterns vary between rifles and ammo. Bullets will not go where they should as often as shooters would prefer.

The bottom line of this post, literally and figuratively, is that, not counting wind or air movement, the better the barrel/bore, and the better the ammo, the more likely that the .22LR bullets will go where they should.

 

[grauhanen]

To answer the question about whether the RWS SM ammo refered to above is consistent enough for effective tuning, the answer is no.

Sure, five ten-shot groups under half-an-inch sounds good. The problem is that too many of the bullets in those groups don't go where they should.

Instead of the small vertical dispersion predicted by MV differences, the actual vertical was considerably greater. ES predicted vertical never more than 0.2" and sometimes half that, yet vertical actually was two to four times greater, it's hard to avoid the conclusion that the ammo is imperfect. Bullets don't go where they should.



It's worth keeping in mind that, according to positive compensation, tuners affect dispersion caused by MV variation. They don't correct or mitigate trajectory irregularity caused by factors such as imperfect bullet heels or center of gravity issues. They don't cure flyers.

It would be convenient to have been able to track all rounds shot in the target shown. But it's not easy when many shots go into a relatively small area at a shorter distance. (Even at 100 yards, tracking individual rounds can be challenging.) Nevertheless, the tracking shown below gives an indication of recurring MV/POI mismatch.



When an ammo's MV and POI don't match, is it possible to get an effective tuner setting? Perhaps some rounds out of many may benefit. Is that good enough? It may be for some shooters. But for consistent results, it's necessary to use ammo that's much more consistent.

It was suggested in this thread (possibly in another recent one) that serious shooters in 3P or Olympic-style shooting (and elsewhere) assess ammo that produces consistent ten-shot groups at 50 yards/meters that are 15mm outside-to-outside as "meh" or just acceptable. For competition they look for ammo that's consistently under 14mm oto.

SK ammos don't give that level of performance, no matter the lot, no matter the rifle. Neither does the lot of RWS Special Match shown above -- at least in my Anschutz 1907.

 

[Williwaw]

Another way of looking at this is to not expect MV variation to account for the whole of vertical dispersion. One approximation of the MV attributable vertical component is to subtract the horizontal dispersion and what remains might better approximate the portion due to MV. That said, if the naked barrel or tuner setting has the rifle in anything but a near stopped(max up or down)position then either negative or positive compensation will mess up the math anyway.

This whole question of whether a system is tunable can be looked at another way if you subscribe to two things ... the barrel vibrations are repeating predictably and a tuner can alter them. With that said the system is in some state of tune or out of tune regardless. The challenge is whether we can figure it out and move it to a better place.

Great discussion.

 

[_Shorty]

Barrels move. Tuners make barrels move slower. That's all they do. It is possible to find a tuner that slows a barrel down jjuuusssttt enough to reduce vertical at a given distance. But that doesn't mean any random tuner will do that job well on any random barrel.

Example:

Eley Tenex
50-yard target

1050 fps round needs a launch angle of 10.827 MOA.
1085 fps round needs a launch angle of 10.368 MOA.

An ideal tuner doing its job in an ideal manner will slow the barrel down just enough so that it has moved/aimed 0.459 MOA further upwards when the 1050 fps round exits versus when the 1085 fps round exits. Then, both bullets will go through the same hole, ignoring wind and other unavoidable randomness for the sake of discussion.

The problem is finding a tuner that's ideal for a given barrel length and contour. Will any random off the shelf tuner be serviceable for a given barrel length and contour? Maybe. Maybe not. If the vast majority of shooters using that tuner also just happen to have pretty similar barrels, the odds are much better that things will work out. But vastly differing barrels will require vastly different tuners. I have an older Harrell here that I don't think was ever bored out, which weighs about 252 g, and a newer one that has been bored out for one of my barrels, and it weighs about 239 g. They're both roughly 2.5x too heavy for one of my barrels, as near as I can tell. I've been playing with ones that are in the 100-104 g region lately, and think I may finally be in the right playing area for that barrel. Lots of trial and error, and some thinking about the goal and how to measure whether that goal is even being approached. But, again, that doesn't mean you'd have any luck with any of those tuners if you put any of them on your barrel. You'd kind of need my barrel to use my tuner. They're a pair. You need a barrel that starts out moving too quickly when bare, and pair it with a tuner that slows it down just enough.

http://www.geoffrey-kolbe.com/articles/rimfire_accuracy/tuning_a_barrel.htm

 

[Leuchtkafer]

Hey Shorty, out of curiosity, what is the length and diameter of the barrel where you're using the 100g tuner?

 

[_Shorty]

Hey Shorty, out of curiosity, what is the length and diameter of the barrel where you're using the 100g tuner?

Well, that's not as easy to answer as one would normally rightly think, and that's probably another discussion all on its own. Hehe.