by B.B. Pelletier
Before we begin today, I’d like to ask you readers if there are any products you have evaluated that you think might be worth your review. There are already over 14,000 customer reviews on the Pyramyd AIR website and more are added every day. Customer reviews are a great way for shooters to read unbiased reports from those who have tried the products before them.
Gamo’s Lady Recon has open sights, so I tested them first as part of the accuracy test. I’ll test the rifle with a scope, too, and I think it’ll be interesting to see how it performs between open sights and the scope.
I tested velocity of the rifle with Gamo Match, Crosman Premier 7.9-grain pellets and RWS Diabolo Basic pellets, but I don’t think they will necessarily be the best as far as accuracy goes. I’ll test it with the Gamo Match, the Crosman Premier 7.9-grain hollowpoint pellets and the RWS Meisterkugeln 7-grain pellets. That should give us a pretty fair idea of the potential accuracy of the rifle.
Focus on some design features
Before I begin shooting, though, I’d like to point out a couple design features that are worth considering. First, let’s look at the open sights. Gamo put a nice set of adjustable sights on the Lady Recon, and they aren’t fiberoptic. That’s a plus for those shooters who use open sights and want precision. Fiberoptics are great for rapid target acquisition, but they lack the precision needed to shoot tight groups, because you’re aligning such large sight elements–the green and red dots. Without them, you’re aligning the edges of the front post with the rear notch–both formed correctly on the sights of the Lady Recon.
Another curious feature is an articulated cocking link–the link that connects the barrel to the piston and pushes it back when the barrel is broken down. Back in the 1970s, before we knew much about the design of spring-piston powerplants, the articulated cocking link allowed for a shorter cocking slot in the forearm, which helped reduce the buzz of vibration when the gun fired. The Lady Recon forearm has a short slot, and no doubt it does help control vibration in this rifle.
The barrel is sheathed with plastic that’s formed into a fluted design–reminiscent of the barrels on fine vintage and modern target rifles. It looks great and helps your hand hold the barrel during cocking, but some may object to plastic on the outside of an airgun. Gamo uses the same sheathing on many of their more upscale rifles, such as the Big Cat and the Whisper, so Lady Recon owners shouldn’t be put out by it.
Accuracy testing
Because I’m not running at 100 percent, I cut the number of shots per group back to 10. It still seems to give a good indication of accuracy, and I think I’ll continue to use it in future tests, except for special tests I want to run.
The shooting was done at 25 yards with the open sights that come on the gun. I did not wear prescription glasses. The rifle was bag-rested with the artillery hold. I used the backs of my fingers because the Lady Recon is so light it doesn’t hurt to hold it that way. My fingers were touching the triggerguard.
Gamo Match
This time I used the new heavier 7.71-grain Gamo Match wadcutter pellet. No special reason for that. I just wanted to try it. The group was not centered on the bull, but as you can see, the rifle gave up a 10-shot group measuring 1.15″, center-to-center.
Crosman Premier hollowpoints
Things started out well, so I put up a new target and switched to Crosman Premier Hollowpoints. This 7.9-grain pellet is available over-the-counter at discount stores, yet offers almost identical accuracy to the boxed Premier 7.9 grain pellet. However, in the Lady Recon, it turned in the worst result–a 10-shot group measuring 1.856″ c-t-c. I had adjusted the rear sight, but as luck would have it, I adjusted it the wrong way, so the group is even farther out to the right. The Gamo rear sight has very tiny icons that are supposed to show you how to adjust the sight, but they’re too small for me to see clearly without a magnifying hood, plus I was in a room too dark to see them anyway.
RWS Meisterkugeln
Next, I tried 10 of the lightweight 7-grain RWS Meisterkugeln pellets. They surprised me with a tight 1.018″ 10-shot group, which was the best of the test. That’s probably as good as I can shoot with this rifle with open sights at 25 yards, so I’m impressed. This time, I adjusted the sights the correct way, but I went too far and wound up on the left side of the bull.
RWS R10
The last pellet I tried was the RWS R10 target wadcutter. In a target rifle these will usually surge ahead of everything else, or sometimes equal what H&N Finale Match pellets can do. This time, though, they were slightly behind the Meisterkugeln, so I would save some money and go with the cheaper Meisters.
The result of this test is my observation that the Lady Recon is a very accurate youth spring rifle. Next, we’ll see what it can do with a scope.
B.B.
Sounds like a nice little sparrow popper for around the barnyard. Could maybe use a little more zap for starlings. Another 100 fps or so.
Reviews???
Too many look like they were written by 10 yr olds who would be impressed to find out they can shoot spitwads with McDonalds soda straws.
twotalon
twotalon,
Then someone needs to show them how it's done! 😉
B.B.
Not bad, and it's good to see that the cheap Gamo Match did well at that distance with a group less than 1/8" larger than the best.
I'll be really curious to see if you get better results with a scope… that should give a good indication of the consistency of the lockup mechanism.
B.B.
That sounds like a hint.
Been thinking about getting a Diana 34 or 48 in .22. Been looking at them for a long time, and think I may owe myself one after all the junk guns I bought at discount stores over the years.
Was looking hard at the 48 for quite a while, but PA just had to keep showing the AF guns on the home page.
Might just get a Diana and start breaking it in while I'm waiting to get my eyes fixed.(cataracts)
twotalon
twotalon,
If you can afford it, get the 48. Besides being more powerful, it's also less hold-sensitive. It's perfect for the .22 caliber.
B.B.
First, thanks for blog and please keep up the great work.
Reviews I would like to see:
Cometa! The Fenix and now the Fusion are starting to trickle back into the USA (Pomona, Air-Venture). I would love to know more about the new Fusion, and the Fenix is available in .25! I have a pair of RWS94s and love them. Are these news ones as good? Better? What is up with the unique muzzle ports on the Fusion? We must know!
Also, the Cometa Indian pistol is a unique but quality springer handgun that I would love to see the bb take on!
B.B.
Pretty sure it will be the 48. Plan on shooting cp and exacts. Possibly polymags.
This will be my open sights heavyweight for 25yds or less. Scopes turn me off at close range.
In fact, one of the biggest "no sale" points for me is rifles that do not come with open sights. AF rifles with high mounted scopes really drive the point home on that one….hard to use up close.
twotalon
Twotalon, BB's right. The '48 in .22 is a nice rifle, and the heavyish weight is (realistically speaking) somewhat offset by the balance and short length of the gun.
Is the price a bit steep for ya?
Vince…
No, price is not too steep. Possibly a bit of paranoia about where my fingers will be while loading it.
I have one of those chinese $20 Cummins tool sale clunks that scares me to load even while holding onto the underlever. Has a lot to do with the construction quality.
twotalon
Daveshoot,
I tested the Cometa Indian many years ago, when it was the RWS 9N. I liked it, if it was a bit quirky to cock. A good spring pistol.
B.B.
twotalon,
The anti-beartrap mechanism in a Diana sidelever is pretty foolproof. When you see what blocks the sliding compression chamber you will understand.
B.B.
Twotalon, I know what you mean about the finger-smasher, but the 48 has a direct, mechanical anti-beartrap mechanism that you can both see and hear engaging. It's pretty secure.
twotalon,
I know what you mean cause my SS would drive me nuts with close shots until I put one of the Daisey Lasers on her, PA DY-7951 @ $24.95, and sighted it in at 8 yards for rat shooting in a friend's back yard.
Works so well that I've been looking at their green laser, but it looks like it has a built in Weaver mount. I want to keep the laser close to barrel for obvious reasons and don't know about using an adaptor and raising it up too high.
I agree with you about rifles coming without open sights– they just don't look right to me.
Good luck with the cataracts!
Mr B.
I haven't been able to keep up with the blog here, but in case some folk where unaware, W. H. B. Smith's "Gas, Air, and Spring Guns of the World" is available as a hardback, released this past July. I picked up a copy from Amazon.
But it's pink…………….
Now this is funny. As you guys know I sell photographic equipment…I deal strictly with schools, gov't and miltitary.
I supply quite a few cameras for Canadian trooopers headed for Afghanistan. A couple of months ago a lot of the orders came in (mostly small Canon point and shoots) specifying pink cameras.
When I asked the purchaser he said "pink is the one color that the soldiers don't permanently borrow…if you know what I mean"
T thought that was hilarious!!
CowBoyStar Dad
Yeah, I owe PA a few reviews. I keep buying so much stuff from them I can't keep up with writing them.
I think the Recon (I have the black one and it looks cooler than the pink but then I don't have any grand daughter shooters) is really fun to shoot. It's small and easy to cock. I haven't let my grandsons shoot it yet. But at 11 & 12 yrs old I think they could handle it fine. When I get home I'll look to see what shoots best in my Recon. I think it was the JSB Exacts.
I got a shocker yesterday. My brother here in Arizona subscribes to "The Rifleman" and the August issue has an article review about the Tech Force Contender Model 89. It's a break barrel made in China. $270 with scope. The review of the rifle itself is not what bothers me, it's the general statements they made about its place with airguns.
What appalled me about the review was that they said it was "…an excellent example of this breed of modern air rifles…" and "…accuracy was good…"
After having said the above they showed the test results. 10 rounds at 10 meters, group sizes in inches:
Gamo Hunter 8.3gr Avg group size 2.93
Gamo PBA 5.4gr Avg group size 3.26
Tech Force 7.3gr Avg group size 2.30
I don't know about you but statements made like the ones above and results shown here by this nationally renowned magazine are to me detrimental to the air rifle community. I believe anyone who reads this article will be turned off to air rifles in general. Who in their right mind would want to spend $200+ on a gun that shoots 2-3" groups at 10m?
Suffice it to say this gun is NOT an excellent example and this article will almost guarantee the reader will not consider air rifles in general to be worth spending money on.
-Chuck
B.B.
Pink gun….
Needs "Hello Kitty" decals.
twotalon
Twotalon,
I have the RWS 52 (dressed up 48) in 177 and find it's one of the more accurate rifles I own. Even with the beartrap mechanism, I keep my arm on the cocking lever when I insert the pellet into the breech, just for peace of mind. I like my fingers – all of them.
The downside to this rifle is it is heavy. To shoot it accurately from a standing position with nothing to rest your elbow or body on is a challenge. But shooting prone or from a table or bench that you're able to put your elbow on is the optimal solution, imo.
It's not a particularly loud rifle, I don't need earplugs when I shoot it indoors, unlike my Discovery, but it's certainly not as quiet as the weaker, underlever cocker RWS 46.
Hope this is of some help to you.
Fred
Mr.B
Tried a laser…still too far from the bore line for me.
Also legal issues if used for hunting game….
"No device for projecting a beam of light shall be attatched…"
twotalon
Chuck,
They rested that 99 directly on a sandbag! Of course they got huge groups!
It isn't the gun's fault–it's the fault of the testers who know nothing about airguns.
The sad thing is, "The American Rifleman" has promulgated such wrong information about airguns for decades. They think of them as "Fer the kids," just like most of the NRA.
B.B.
Fred
Extra weight helps steady the shot for me. I also improvise rests when possible….anything to up the odds.
An all day carry around the woods is another story.
twotalon
How about reviewing a few of the products that no has reviewed yet? I see you carry the Winchester 850XS22.
No one's said a word about it. I would imagine the RWS 34 combo package is the "gold standard" in the same price/feature range. How do they compare?
I agree with the comment above that indicates that some of the customer reviews aren't very well written.
There's a popular knife website back east that has an even worse problem with that. They've responded by editing the worse comments and appending text to the effect that "Thank you for your comments. We've updated the product information to reflect the correct materials. We have editted your comments to correct the spelling or other errors."
It's probably a full-time job.
Anyway, keep up the good work.
I intend to buy something soon. I like the looks of the Winchester 850, but no one has reviewed it so I'm inclined to make a "safe" decision and go with the RWS 34.
-Bob
Bob,
the RWS34 is an excellent rifle and considered "the" starter (I know I'll get some flack for that word) air rifle for the serious shooter who wants to get involved with spring-piston power plants. It has accuracy, quality, reliablity on it's resume. Plus you'll have lots of company out there and there are tons of modifications for this rifle, should you choose to start up that path.
Fred
I understand the pink cameras thing.
I ran a truck crane for 12 years. Rigging was always disappearing and keep 4 matched chokers or shackles was nearly impossible. I thhought it was just that they blended in with everything and were hard to find.
WRONG.
I decided to paint mine hot pink. None of our crews used it and it would stand out against anything we work with.
Suddenly none of my ironworkers or fitters wanted to use any of my rigging. They would suddenly bring out their own stuff. before they'd touch any of my girly stuff they would use gnarled old chokers that were shaped like streaks of lightening.
I went from replacing bits and pieces monthly or in the case of small chokers weekly to not losing anything for a year.
Pink is good
jim
B.B.
Nice gun but no pink for me.
TwoTalon, the RWS 48 is a fantastic rifle. Shooting offhand is made easier by resting the rifle on the body. It is really no strain to shoot 30 shots like that.
Kevin, the Eley Wasps have a great tin design. Bzzz bzzz. Fit to the B30 is excellent; better than the JSBs although the JSBs are really not bad. Shooting behavior seems harsher, and my impression is that the pellet is faster, but I can't tell without a chronograph. It took me a little while to adjust to the new shooting behavior, but it appears that accuracy is equivalent to the JSBs although this is what you would expect for a decent pellet. Any additional accuracy would not appear except at range distances, and I don't know when I will get out there. Anyway, it's definitely a good pellet. Thanks for letting me try it. It reminds me of how much pellet selection can alter the shooting experience.
Matt61
The statistics monkey is not off my back just yet. I returned to an earlier question about how to relate group sizes to each other; for instance, how does a 10 shot group compare to a 5 shot group other things being equal.
Supposing that shots within a group are distributed according to a bell curve that is fully established at 30 shots, it seems like it would be possible to find the radii for different group sizes within that 30 shot group.
For those interested in the details, the area under a bell curve corresponds to the probability of occurrence. You find the probability of a particular group size as a fraction of 30 e.g. 5/30. Starting from the middle of the curve, you establish a banded area and read off the standard deviation associated with that area using this calculator
http://davidmlane.com/hyperstat/z_table.html
In our physical situation, standard deviation corresponds to radial area. By taking ratios of standard deviations, you can find ratios of groups sizes to each other.
The values for common group sizes are
3/5 = .59
5/10 = .50
10/20 = .43
In other words, a 3 shot group will be .59 of a 5 shot group and so on. Qualitatively, this means that for a range of shots between 3 and about 17, the growth is almost linear which you can see from the shape of the curve at those points. Beyond 17, the size starts to expand faster. Between 25 and 30 the group expands rapidly indeed which throws the value of the 30 shot group into question. It is reliable, but it is not very representative of the preceding 25 shots.
So, it looks like the usual 3,5,10 shot groups have a good practical value after all.
Matt61
Wayne,
A suggestion for a t-shirt. Back of shirt: curved line on top saying Ashland Air Rifle Range; curved line on the bottom Wacky Wayne's Field Team; in the center of the cicle a sqirrel FT target holding and eating a nut. Perhaps the squirrel could be called Mr Wacky? Who knows maybe Mr Wacky will morph into a mascot for the AARR.
Just some thoughts of mine good idea/bad idea please let me know.
Mr B.
Matt,
Statistics is convoluted.
You've basically got the notion correct, but the correction factors are off. Group size isn't really a "standard deviation" it is really a range measurement.
Also these sorts of correction factors would be based on a "random" group. You can't shoot six groups of 5, and then select the best of the six to convert.
Believe me (Do I hear Jane talking to me??), more shots in a group yields a more reproducible result. You shoot a 1.40 inch group. Fine, let's assume that nothing changes. You shoot another group, are you likely to get EXACTLY 1.40 the second time?!? No!!
The value of 1.40 is meaningless by itself. You also need to have some idea of how much that measurement will vary.
All of this also fails to address a very critical point. Group size is HIGHLY biased by fliers. You have a million shots in 1/4 inch, then shoot one two inches off. Suddenly the group size goes from 1/4 inch to two inches based on one shot.
To just call a shot a flier is sloppy. You have to have some basis to make that decision. The point of statistics to to put your "gut feel" on a mathematical basis.
Let me be BLUNT. 3-5 shots groups survive because there isn't much information in that few shots anyway. So reducing that information to one measurement is about as good as it gets.
However, if you shoot 30 shots in group then you're basically throwing the information in the other 28 shots away. (not absolutely true, but close…) If you have 30 shots to be analyzed the relative standard deviation measurement technique is a much much better way to analyze the data. The RSD technique would also give you a mathematically sound way to handle fliers.
There are a lot of considerations…
Bob,
It would be very time-consuming to correct the product reviews. I admit that many are not well-written, but that's the nature of reviews written by people who are more interested in guns & shooting than writing.
As hard as it is to believe, Pyramyd AIR actually declines some reviews because it's impossible to understand what the writers are saying. I'm sure it wouldn't surprise you to learn that a number of reviews for youth guns and airsoft guns are written in text message format. Those are declined, too.
Edith
word verification: boast
Edith,
Like I said, it would probably be a full time job. Just to catch the worst.
I suspect the knife store I mentioned had a problem with repeat offenders that were making untruthful disparaging statements. Acknowledging them, but editing their remarks and politley saying so may be useful in making them go away.
I am very appreciative of the reviews, by the way. And this blog.
I believe it was a review on this very blog that tipped my decision to buy a product that unfortunately was no longer available at Pyramid because it had been discontinued. I am very happy with my RWS p5 pistol.
The review was spot-on.
-Bob
Matt,
I’ll add to what Herb said about group size and statistics.
First, it is important to understand that group size is just that – the size of a group. People use it as a means of comparing both the capabilities of different guns, pellets, and shooters, and it is commonly understood. Larger groups (in terms of count) will yield more spread, and are more representative of true capability. As long as people understand the limits of what they are talking about, and are comparing apples to apples, there is nothing wrong with it. But one thing that group size most definitely is not is an accurate assessment on the capability of the system (regardless of how many shots are in the group).
This is very different from a statistical analysis of the capability of the gun, shooter, or the gun/shooter “system.” It is a simplification that does have value because it is simple. An example of a statistical analysis would be answering the question “what would be the distance from the mean impact point such that 95% of all pellets shot will fall within that range?” Or more likely, “what is the accuracy of this rifle/pellet/shooter system at plus/minus three sigma capability?”
If you designed and ran a test, measured the data, and calculated this answer, it would not yield the same answer as a group size – even if the group was of thirty shots (in fact, it would certainly be larger than the 30 shot group size). That is because the total variance of the system will always be larger than the inclusive data in the test set.
One other thing to clear up – 30 is not some magical point where we have “legitimate” data. It is actually the minimum amount of data points needed to be able to perform a reasonable analysis on the statistical capability of a normally distributed system. I do believe that in most cases the accuracy of the “system” we are talking about can be fit to a normal distribution.
It is also important to understand that this kind of analysis applies to what is known as “common cause” variation, and not “special cause” variation, which of course leads to how to treat flyers. I would contend that their really is no such thing as a flyer – it is a result of an intended good shot at the time it was made. Of course, if one knows of a special cause source of variation and can plan for it, that could be eliminated form the system. A good example is first shot variation in some cases – for me, if I need to take one critical shot at a pest on a gun that has not been shot in a while, I first shoot one or two into my trap in my workshop before going out to take my real “first” shot. Of course this approach would not work for hunting. So the real question for flyers is to define the boundaries of the system that we are trying to test – if we know the data point was messed up by something outside that boundary, then we could discard it.
Herb is right in that group size throws out a lot of potentially valuable data, but not many people would be willing to collect it. The way to go about it would be to set up a target paper with at least 30 individual targets on it, and, with the sights properly centered, shoot one pellet at each target on the paper. Then each pellet’s unique POI would be measured for at least a radial dimension off the target, but preferably measured for both X and Y dimensional variation from the target. This data could then be analyzed in total for a true accuracy picture of the gun/pellet/shooter system.
I for one would not want to do all that work. I’m sure B.B. would drown under the added work load. And in the end, since practically nobody else would do it either, you would have nothing to compare your data to. So even with all the weaknesses, I happen to like group size because it is so simple and universal, and if done right, can provide confidence in relative capability of the system.
Man this was long . . . . class dismissed!
Alan
Matt,
I was thinking about this last night some more. We started out and wondered around in statistics.
I tried to related things to a standard deviation of an error source because an appreciation of the one dimensional case (ie X) works gives you an understanding of how the more complicated case in two dimensions words (ie X&Y). That is no doubt some of the confusion.
The other thing that I realize is that I have been basing my discussions on the standard deviation of a distribution. I did not really discuss “”group size” enough. Group size in statistics is a “range” measurement. Range measurements are based on the difference between the highest value and the lowest value measured. For group size that is the distance between the two widest spaced shots.
There is a subtle but important factor. If you have a larger and larger number of shots in a group, the group size will actually just keep getting bigger and bigger. It does increase without bounds. So for an infinite number of shots, the group size would be infinitely large. Think about the normal distribution. It stretches to infinity in either direction.
However the standard deviation of the normal distribution is a single number. When you measure the value, the measurement can be high or low, but the measurements would bounce around the particular singular value. So as you get more measurements the precision of the measured standard deviation increases, but thhe nominal value of the value does not. Unless God tells you, you’ll never know the true value of the standard deviation.
Basically a shooter is left in an awkward situation. When you measure group size, you are reducing all the information about the spread of the individual shots in the group, to a single number the group size. If you shoot groups of 3-5 perhaps as much as 7, this doesn't loose much information compared to actually using the more labor intensive method of RSD. However if you were to use 30 shots groups, then you're throwing away a lot of information since all thirty shots have been reduced to a single measurement.
Now we have our group size measurement. Let's take it a step further and measure 30 groups. Now you have enough groups to do a good statistical analysis on group size. However you've shot 5*30 or 150 shots! You could have arrived at the same information by using the RSD with only 30 shots. So that is the trade off. Do you want to do more work at the range, or more work calculating?
This whole thing isn't bad when shots from a pellet gun cost pennies. But assume that you were using some commercial ammunition where the rounds were $2 each.
Look at the following paper which has two charts:
http://www.shootersjournal.com/Features/WHICHONE.pdf
The first chart basically gives the "efficiency" of the measurements. In other words, what percentage of the information available does the particular measurement capture? You can see that for low numbers of shots, there is no real difference between group size and RSD.
The second chart, shows something else. Assume that you have two pellets "A" is the gold standard with another pellet "B" which is 10%, 20% and 40% better than A. What are the relative chances that Two group measurements (one for A and one for B) will find that B is better than A. The chart is a bit convoluted because it has been "normalized" so that the group size measurement is 1 regardless of group size (hence my reference to a "normalized" graph.)
Two basic points. First RSD is more efficient at finding the differences for larger (>5-7) groups sizes. Second, too find small differences you need a lot of shots!
BB,
I've got to have it just for the pink thumbhole stock! Who needs accuracy.
Alan/Herb,
A common theme seems to be that a group of any size tells us more than nothing, but less than anything definitive:). I prefer to think of groups in inductive terms, i.e., with similar-sized groups shot by the same person, we can infer (not deduce) something about the accuracy of the rifle/ammo, and most of the time, the shooter:).
For example, taking a benchrest rifle that has won several competitions and giving it to a mediocre shot is liable to tell you more about the shooter than the rifle's accuracy. I also see this online a lot — "my brandx rifle not accurate as claimed", on the basis of an afternoon at the range with 6 different kinds of ammo!
Matt,
So all in all, here is what I would recommend if you really want to use a reasonably rigorous technique.
(1) Shoot ten 5 shot groups.
(2) Determine the mean and standard deviations.
(3) Using +/- 2 standard deviations determine a 95% confidence interval for an individual group size measurement.
(4) For one time only – Throw away any measurements outside the 95% confidence interval as having "fliers."
(5) Recalculate the mean, Std deviation and confidence interval based on the remaining shots.
Let's call this your current gold standard ammo "A". So you shoot one group of a new ammo "B" that falls within you confidence interval for "A". Now you can't say that B is better or worse than "A". At the 95% confidence interval for "A" (since you don't yet have a Confidence interval for "B") there is no difference.
Now you go ahead and shoot 10 groups of B, to get mean, std dev, and confidence interval for “B.”
But you can also now use the “std dev of the means” to compare "A" and "B". So you can compare the average of the 10 groups of "A" with the average of the
10 groups of "B". This is much more sensitive a test and may find that "A" and "B" are different.
If you shoot more groups for "A", then you use all (including ones you may have previously discarded) and recalculate using original procedure. So what may have appeared to be fliers with only 10 group measurements, may not be fliers when you get 20 group measurements. Or the reverse could be true. You may now find a flier in the original data. But more data, more precision.
So with the technique I proposed you're in between. You've shot 50 shots instead of 30, but you got some of the benefits of RSD, and some of the benefits of using group size.
Note that for you favorite ammo "A" there may be variations between lots of "A." In other words different production batches of "A" may be better than others.
You could also test to see how many filers you get. More fliers is more inconsistent ammo. This is a whole different discussion in non-parametric statistics.
The other fly in the ointment here is “standard conditions.” You’ve measured your favorite ammo with the particular gun being tested. But your data was obtained on a nice day with little wind. Today on the range, the wind is gusty with gusts up to 15 mph. Now “A” isn’t going to group as well as it did under the better condition either. So there are other tests you can use. But under such awful conditions you’d have to be a masochist to be testing different kinds of ammo. The group size for your “gold standard A” ammunition will be so large that it will hard to figure anything out.
So on any given day, you’d need to shoot some groups of the “gold standard A” ammunition to verify that A was conforming to standards too.
So you can see that I'm not talking about the "best" group size for "A." The best groups size for "A" is unbounded too. Shoot more groups and just by chance you'll get a "better" group size. Example – Shoot 30 groups, what is the best? Now shoot a million groups what is best? Almost certainly one of the groups in the set of a million will be "better" than the best measurement in the set of 30. But it isn't "better," it is just chance. What you really need to measure is the "expected group size" which is the average of a number of measurement of group size.
——
With our 95% confidence interval, the "best group size" will be two standard deviations below the average group size.
Regards,
Herb
PS – BG_Farmer, you're absolutely right the shooter is an important part of the overall error. So just in simple terms there variations due to shooters, to guns, to ammo, to weather conditions. You can't really isolate one and measure it alone directly.
Geeze, I should leave well enough alone…
Of course if you really wanted to do the mathematical analysis right, invest in some software for your computer and a flat bed scanner. With mouse you register the target paper in computer scan, then click mouse on each shot. The computer then does the heavy lifting with the math.
—————–
Tom !!
This might make a really good program for your air-gun show… A great many shooters have a computer and scanner now..
Let the software guy show an excerpt of the show on his website. He has to leave "Now from the Ammerican Airgunner show on the sportsman channel.." in the clip and have a link back to your website in order to be able to use the clip.
No idea how all the contracts are written. Could maybe shoot some extra footage and summarize results in two minutes? Gets you more exposure…and link back on web to your show.
———–
Believe in karma? verification word "nothype"
RE: Moisture in air with HPA pump
Wondered about this…
Do any pumps have an inlet which can be connected to a hose?
I've wondered about making a tube to absorb moisture before the air goes into the pump. Use PVC pipe and hook it to inlet. Fill PVC pipe with something that will absorb water. That way you remove the water before it gets to pump and the inside of the pump and the inside of gun stay dry.
My thought would be to look at some of the silica gel cat litter. I'd wash it twice with water to get our deodorizers and perfume. Then dry it in an oven (When your wife isn't looking…).
Some litter had silica beads that would change color with water. If you had clear tube, (like to cover 4 foot fluorescent light), you could see color change and know that it is time to dry gel again.
Of course you'd need caps for drying tube for when it wasn't in use….
twotalon,
Thanks for the heads up on the game laws and lasers. Responding to your comment about a laser mounted on an AirForce rifle being too far above the rifle's bore has be puzzled.
I have the laser on my SS on the dovetail that's part of the barrel's shroud as close to the gun's "handle" as I can. You cann't get it any closer to the bore than that.
Alan,
Thanks for your math class.
Herb,
Thanks for your link. Between you and Alan some of this math "stuff" is making a little sense to me now.
Mr B.
RE: Drying tube
Most of the readers are not probably aware of what I'm talking about:
Go to:
http://www.drierite.com/default.cfm
and then click on "LABORATORY GAS DRYING UNIT" in the left hand menu selection.
Mr. B….
It's still farther than I want for working close range.
I don't want to have to allow much for distances between a couple feet and 20-25 yds.
A tall front sight is about as far as I want to go above the bore.
twotalon
twotalon,
I just measured from the center of the bore to where the laser hits a ruler. The distance is about 1 and 5/16's of an inch. You cann' get near that close with iron sights.
Just went and measure some shots for you: @ 2 feet the POI is 1 1/4" low, at 10 feet the POI is 3/4" low and at 48 feet the POI is 1/2" high. My yard doesn't allow me to get farther away than the 48'. They were all shot with with the 12' .22 barrel runing on CO2.
RE: Point Blank Range & red dot
I do find the red dot easier to use than iron sights to pick up tree rats.
Ok, What I'd like to have is a red dot sight with a D shape. Flat part of D to dovetail. Bring center line down to about 3/4 of an inch which would be the same as the iron sights on many of my guns.
I wouldn't mind sawing off the front sights to get the clearance. That would give me point blank range from about 8-20 yards, fine for pest control.
Herb
PS – With red dot I also realized that my glasses were part of the problem. I have been using variable focus lenses. Switched to an old pair with just distance correction. BSA now has spot instead of a very elliptical blob.
Mr. B
The laser that I have measures 2 1/8" abave or below the bore.
Not acceptable.
twotalon
RE: Matt's fudge factors
OK…Let's assume that the group size measurement is governed by the "normal distribution." I know that this isn't true, but it probably doesn't matter for the small amount of data that will ever be available…
The appropriate control charts can be found here:
http://www.itl.nist.gov/div898/handbook/pmc/section3/pmc321.htm
———
So if you make N measurements of group size, then the control limits are:
GS +/- A2(N)*SD
Where:
A2(N) is the A2 factor from table for N groups.
SD is the measured SD for N groups.
This means that I have N groups, NOT N shots in each group…
No dependency here on how many shots are in the group. Obviously the groups should be of the same size. There are always other ways to handle the data…
So if you make say 5 measurements of group size, then you can put control limits of:
GS +/- 0.577*SD
Where SD is the std dev from the 5 measurements. Let's assume that my measured GS is 1.25 inches and std dev is 0.2 inches.
Now I shoot ONE five shot group of ammo "B" and it has a group size of 0.3 inches, what can I say?
My 99% confidence interval for my gold standard ammo is:
1.25 +/- 0.577*0.2 or
1.25 +/- 0.12
That is from 1.13 inches to 1.37 inches. Since my measurement of 0.3 inches is outside of this interval, I can say that there is a difference between my gold standard ammo and "B". So "B" is better than my gold standard.
——–
Now for a single measurement of group size, it is a different calculation. The confidence interval is asymmetrical.
Upper limit is:
UCL = GS*D4(N)
where D4(N) is the D4 value for N shots, and GS is the group size for the N shots.
The lower control limit is given by:
LCL = GS*D3(N)
GS and D3(N) follow as above.
Again, this is for 3 standard deviations or a 99% confidence interval which is pretty picky.
——-
Let’s say that I only know now that my "gold standard" pellet shoots an average group size of 1.25 inches. I do not know the standard deviation. I go out to the range with some pellets, named SGM's (Self Guided Missiles).
I shoot a single group of 5 which has group size of 0.50 inches. What can I infer?
THE UCL is
0.50 + 2.115*0.50 or
0.50 + 1.06
1.31
My LCL is
0.50 + 0*0.50 or 0
Since my gold star ammo falls within the control limits of 0 to 1.31, I must conclude at the 99% confidence interval that there is no difference.
So sticking with my 5 groups for the SGM's I have to shoot more groups to get a tighter control interval. One group won’t be enough to find the difference.
——–
The difference between the first case and the second is that I knew what the standard deviation for my gold star ammo was. For example, a lot of range data from the past.
However it isn’t so obvious that I would necessarily know the std dev of my gold star ammo. Let’s say that there is a bit of wind. Now obviously my average and std dev for the gold star ammo isn’t going to be as good as the range data that I have from calm days. However since there is wind and wind is detrimental to group size, it stands to reason that measurements on this wind day would be worse than 1.25 inches with a std deviation of 0.2.
But the new ammo seems to be shooting great! Now instead of shooting a whole bunch of groups of my gold standard ammo to define a new GS and std dev for my gold standard ammo, it seems more productive to shoot more groups of the SGM’s. 0.50 inches seems better than what I do with my current gold star ammo even on calm days. I’m guessing that if shoot about 4 groups of the SGM’s then I can conclude that the SGM’s are better than my current gold star ammo.
————-
Such control charts are very very restrictive for small sample sizes. In other words, you need very large differences to be about to say that pellets "A" and "B" are different.
Again, charts based on a 95% confidence interval would be less restrictive, but the same idea applies…
Ok, one last post on statistics. I listed my e-mail address the other day. If I ever post on statistics again, send me a nasty note to my private e-mail…
🙂
Given a group size with N shots. To convert the group size, GS to an estimate for the standard deviation, s:
s = D(n) * GS
N d(n) d(n)/d(2)
2 0.8862 1.0000
3 0.5908 0.6667
4 0.4857 0.5481
5 0.4299 0.4851
6 0.3946 0.4453
7 0.3698 0.4173
8 0.3512 0.3963
9 0.3367 0.3799
10 0.3249 0.3666
11 0.3152 0.3557
12 0.3069 0.3463
13 0.2998 0.3383
14 0.2935 0.3312
15 0.2880 0.3250
16 0.2831 0.3195
17 0.2787 0.3145
18 0.2747 0.3100
19 0.2711 0.3059
20 0.2677 0.3021
This is basically calculating an estimate for the radial standard deviation based on the group size.
Now the third column is D(n)/D(2). The point is that the more shots in the group, then the smaller the fudge factor for the conversion gets…
Help needed
I'm new to this hobby, other than while I wand a child. Got the Beretta 95fs air gun about a year ago and love it. Now my questions, I did lots of research and decided to get the Falcon Hunter w/Gas spring .22 and I read that the scope mount that comes with it don't hold well. So I plan to get a 1 piece mount. I'm 6 ft tall and weight aprox. 230 lbs, not sure if this matters, but added just in case. I need to know what the best mount to get. I would like to be able to see the gun sights with the scope attached if possible, maybe a see-thru-base? Is it better to get one with weaver or dovetail sense the gun has both, which would hold better? I plan on using the scope that comes with the gun. It got mixed reviews, but plan to use it mainly for hunting. Will this scope work for this? And last, which would be the best pellets to use for this gun? Thanks,
Jack
oops, 92fs
Jack
Jack,
You DON'T want the kind of mount where you can also see the sights. They never work like the advertising sounds. They are nothing more than an extrusion with a hole in the center for weight reduction. Some ad guy made up a story about seeing the sights to explain why the hole was there.
A one-piece mount will work, but a two-piece mounts will give you better options for scope placement. With a one-piece mount, you are stuck with just one scope position. Two-piece mounts hold just as firmly as one-piece.
There is nothing wrong with the two-piece mounts that come with the Falcon Hunter. Give them a try before spending more money.
The Falcon Hunter has a base that accepts Weaver dovetails (Weavers are dovetails, too. You must differentiate them from 11mm dovetails by using the 11mm dovetail designation) as well as 11mm dovetail mounts, though the base is not cut for Weaver mounts. But the slots in the base will allow you to butt the Weaver keys against the back of the slots in the rifle base.
Only two-piece mounts will work if you want to butt both rings because the rifle base is cut with an odd slot pattern. You can read about it in this four-part report:
/blog/2008/6/walther-25-caliber-falcon-hunter-part-4/
My advice is to buy the rifle as it comes and use it that way first. Then you will be better able to judge what you might like to add.
B.B.
Thanks B.B.
Sounds good to me. One last question for now anyways, 🙂 My other choice was the Norica Goliath 88 Classic Carbine Combo. I know some think it's ugly, but I like it. But it not have sights built in. And still not 100% sure I still won't got it, lol (really like the looks and power) and like the ideal I can pling with it too. I know it just came out, But have you had a chance to play with it? Any info you might have would be appreciated, and thanks again.
Jack
Jack,
I have a Goliath and I have shot it, but I haven't tested any of the new Noricas yet. However, I find them all to be very well made. I'd say get one if you want it.
I will be testing them soon!
B.B.
Thanks B.B.
That's what I wanted to know, if they where made good. Thanks again for your time on a Sunday. PA has a good employee with you, Keep you the good work.
Jack
Now if only I knew how to post pics!
Reworking of the H 490 is finished:)
Stock has been reformed,cheek piece was
smoothed off,forearm shortened,forearm
angle was removed with a curved swell at
the trigger guard for a forward hand rest.
Butt is 1" shorter with rubber pad put back on and fit was close enough that I
didn't bother to reshape pad.
After the red stain was removed(PITA)the
wood had decent grain so it was sanded
and light walnut stain applied,then
waxed to smooth the looks more.
Bbl was cut 5",bluing removed then
polished to a chrome like shine.The bore was noticeably off center,but with
a careful crown using a drill press and
small cutting stone then a polishing
stone accuracy is good.The sights were
taken off and I like it so well with a
scope that I'm not going to put them
back on.The rifleing is very nice,even
though the bore is a bit large.I
thought I might have to choke the bbl.
but I'm getting quarter sized 10 shot 1 holers
at 25 yds.with Daisy precision max,if
I concentrate.That's good enough for
me any day!
At 37" and 5.5# with scope this is a
neat little carbine now and very handy.
Don't worry Vince I didn't mess with
your handy work at all.And I'm still very happy that I got this rifle from you.Now it's got that personal touch
and if this rain lets up I'm gonna go shoot it some more:)
JTinAL
Herb,
The FX 4 stage pump has two moisture filters included. Can’t speculate as to whether or not they actually work.
Bg farmer,
While we have never met, in my mind’s eye I have a fairly vivid picture of you. To that I added the pink rifle with you explaining to some buddies how a little off season practice with it could help them with their ’06’s this winter. Big smile.
The Walther LG 55 arrived, and at first I was a little disappointed, but after some bonding we are quickly becoming friends. I had sold all my Springer’s, and suddenly it felt like “old times” as I sighted it in with 5 tins of various pellets by my side. The lack of a safety still has not sunk in as I push the imaginary HW button off every time. I will try and muster a full report sometime in the future.
B.B.
The Walther is shooting JSB 7.9’s at 523 fps. Does that sound right? I would guess that it is?
Volvo,
Thanks, for the pointer. I assume the moister filters have some sort of membrane that keeps liquid water from being blown into the rifle. The air driers I was talking about would remove the moisture before it got into the pump. The driers would be before the inlet to the pump.
Part of the problem is that the gas in the pump will be warm. As the gas under pressure cools, then moisture in the HPA can condense on inside of the rifle's pressure reservoir.
The moisture issue may be a strong recommendation for a scuba tank. The scuba air will be dry.
Herb
PS glad to see activity – was beginning to think that I had scared everyone away with all of the statistics posts 😉
Herb,
If I had to guess, we have lost the contributors’ to the Emmy award show. Who knew?
As far as the inlet on the pump, I have never figured that out. The filters are located on both ends of the hose, in contrast the Disco pump had no such animals.
Volvo
JT,
Sign up for a google account and set yourself up with a blogger ID and gmail. Then you can upload pictures really easily to picasa. Your 490 mods. sound pretty cool.
Volvo,
Sounds like a dare:). I do have a large pink 70's style comb (for when my hair is over a 1/4") that I've been waiting for years to have somebody comment on. I bought a pack of 2 dozen mostly normal combs one time and carried and lost them all except for the large pink one — so I decided I would just carry it for a joke and to teach myself to be more careful. That was 12 years ago. The only person who's ever made fun of it is my wife, even though I like to flash it around when I can find an excuse:).
By the way, I am not watching the Emmy's…Sunday is Iron Chef night:).
BG
Thanks for the tip
Hope I did it right,here's the link
http://picasaweb.google.com/home?hl=en&tab=wq
Never mind link didn't work 🙁
will try to figure that out later
JT,
Easiest way is to look at the picture and then copy the link — that will put the user into the album. I've been curious about the 490 mod's since you announced your intent, but you worked fast! Not sure I would have shortened barrel and removed sights, but curious how it works for you.
Volvo,
Forgot to say, will be looking forward to LG55 reports.
http://picasaweb.google.com/JoeTinAL
O.K
I think the link above will work.
If not someone let me know and I'll try again:)
BG
I shot better with the rifle after the mods,but a lot of it could be the scope.
It definitely handles better for me now,
and It's more comfy :)Since I'm only
~5'6" with short arms I usually go for
the youth size rifles.:)
HAH word veri. is viable,so true!
Volvo,
The 55 sounds close to right.
B.B.
JT,
You really redid it. I had thought about trimming mine up and staining it a more natural looking color, but you took it all the way.
Its hard to believe how much wood you took off the forearm — or that there was that much excess there before. I'm not a big fan of the flat bottom on the forearm, but it can be useful and that's a personal bias on my part. Getting rid of the cheekpiece was all good by me, and you got a nice straight comb effect from the shortening. I think you thought it out well, especially since it fits you better now.
I also notice you have screw covers in both the before and after — did you add them or did Vince?
Hi all,
I came in tied for third.. shooting in the open class against folks with harnesses and some with higher ft. lb guns.. (I'm shooting the 12 ft lb USFT and no jacket or harness… that's the international class)
Here is a link to a report and some pics of the Washington state championship field target shoot.
report:
http://www.network54.com/Forum/451309/message/1253461470/Washington+State+FT+Championship
pics:
http://www.network54.com/Forum/451309/message/1253514047/Some+pics.
Next Sunday, the group from the bay area and one or two of the Washington club will come here! Yeah.. I'm gonna make a fun course for them.
This is just toooooo much fun!
Wacky Wayne, MD. (match director) .. not doctor!
Ashland Air Rifle Range
Hey Wayne,
Do you remember about two years ago when you wondered what field target was all about?
B.B.
B.B.
Just barely! … what an incredible journey!..
It was that article on starting a field target course that did it!!
You and the folks here have led me through it..
When I bought USFT#44 as a collector piece from Billy Lo, (2005 national champ) #44 was the prize for the 2005 national champ, intended to go to the worlds with the best US shooter….
I had no plans to shoot it, my AAs410 was keeping me busy.
But, finally I got the bug, and wanted to learn to shoot it.. so I had to get into shape..
Six months of swimming later, I could comfortably stay in the sitting FT position… so practice began…
Then Rick Knowles the North West FT director, heard I was interested in starting a course, so he helped me. They needed a place half way to California, so the Ashland Air Rifle Range was born.
Now I'm so hooked on the sport, I'm down at our range practicing every spare minute! …
and NOW, going to Texas to shoot air guns… I must be Wacky!!
Wacky Wayne,
MD. Ashland Air Rifle Range
JTinAL,
I looked at your picassa web site and your wood working. I do not believe my eyes!!!
-Chuck
JTinAL,
Ditto what Chuck said. I am amazed!
B.B.
RE: Group size as a Range
Just to leave some footprints in the sand, in case anyone reads the above posts on group size and range, I wanted to clarify my remarks that the distribution of group size values is not a normal distribution.
In the above posts, I tried to used the range of N observations from the normal distribution in one dimension to give an analysis of group sizes versus the number of shots.
In part you have to back up and look at what the group size measurement is. Group size is sort of a odd measurement in that it is never negative. Think about a +/-3 sigma confidence interval around the mean. So long as this confidence interval does not include zero, the normal distribution would suffice. However when the confidence interval includes zero, then the normal distribution isn't a good assumption. The distribution must be skewed. The closer zero is to the mean, relative to the 3 sigma window, then the more skewed the distribution would be.
Rather than analyzing the group size from a theoretical distribution there is a paper which did a large simulation. I give a link to it in a later post on the next day from this blog.
http://www.the-long-family.com/Group%20size%20statistical%20analysis.pdf
However the Long data calculates correction for 1 MOA rifle based on number of shots. I wanted standard deviation. So I took the last three numbers of the range conversions for the normal data and thhe long data and calculated a "fudge factor" of 0.489478. In other words, the Long data * 0.489478 will now give standard deviations.
#../long./.norm./.long_c
02 1.597 0.8862 0.7817
03 1.158 0.5908 0.5668
04 0.979 0.4857 0.4792
05 0.876 0.4299 0.4288
06 0.808 0.3946 0.3955
07 0.759 0.3698 0.3715
08 0.720 0.3512 0.3524
09 0.693 0.3367 0.3392
10 0.667 0.3249 0.3265
11 0.646 0.3152 0.3162
12 0.630 0.3069 0.3084
13 0.614 0.2998 0.3005
14 0.602 0.2935 0.2947
15 0.590 0.2880 0.2888
16 0.579 0.2831 0.2834
17 0.571 0.2787 0.2795
18 0.562 0.2747 0.2751
19 0.554 0.2711 0.2712
20 0.546 0.2677 0.2673
You can see that the corrected experimental data from the Long paper agrees well with the theoretical data based on he normal distribution except for the case of 2 and 3 shots. 3 shots isn't too bad, but 2 shots is clearly off.
I believe the experimental data is good, thus the distribution isn't truly a normal distribution. Too give the reader something to chew on, I think the problem is related to the notion of group size, versus the size of the covering circle.
!@#$%^&*
I should concluded with "Too give the reader something to chew on, I think the problem with small numbers of shots is related to (1) the overlap of tail of normal distribution with zero and (2) the notion of group size, versus the size of the covering circle."