Testing the Testing

Get or give advice on equipment, reloading and other technical issues.

Moderator: Mod

Message
Author
wsftr
Posts: 202
Joined: Tue Jan 30, 2018 12:58 pm

Re: Testing the Testing

#31 Postby wsftr » Sun Apr 19, 2020 8:57 am

pjifl wrote:wsftr wrote

here is one - 1000 yrds - shot during a comp. first shoot at 1000 after 100 yrd load dev.
Do ya reckon its case capacity throwing the shots high? what would people do. The group isn't centred as I was testing the load so apart from calling the wind and adjusting the scope no change were made. A lot of F class keep vert under control by clicking.

What to do - was it me, the load, conditions, hold, what?


There is very little that can be deduced from that group. Perhaps a plot of elevation V time will show a trend.

Why not repeat it while measuring V and also recording the wind setting for each shot. You then have something to go on which may preclude certain causes.

BTW, after a few hour's searching I cannot find my work on measuring V from different volume cases. I have not entirely given up. Ironically, the critical plot was actually part of a previous post on this site I did ages ago but I cannot find it in the archives.

Peter Smith.


Thanks for the reply(Peter, Albert, Gsells) - I posted it up as I thought it was a group that represented what many often see and discount in either load dev or the comp and which I think is one example of why you posted originally Peter.
For me shot 10 was very concerning - I know myself, my gear and my load pretty well and instinct said - that wasn't some anomaly....but what to do (assess the test). It would have been very easy to discount shot 10 and focus on the little cluster and think I had a hammer and I muffed shot 10.

Your suggestion is in essence what I ended up doing (plotting vertical over time - if I understand you correctly) . I'm always interested in the process of how methodical, data driven shooters solve problems rather than jumping to a specific conclusion and trying this that and the other in a scatter gun maybe this will work approach or worse assigning it to some random variable such as pulling the trigger and conditions and then on the big day it all goes to custard because there is a fundamental that's not right.

Rich4
Posts: 534
Joined: Sun Feb 24, 2019 2:33 pm
Location: Chinchilla

Re: Testing the Testing

#32 Postby Rich4 » Sun Apr 19, 2020 9:05 am

:lol: Gee Gyro I didn’t realise you had sunscreen in NZ, didn’t think there’d be much call for it from what I saw

Gyro
Posts: 764
Joined: Sat Jun 10, 2017 2:44 pm
Location: New Zealand

Re: Testing the Testing

#33 Postby Gyro » Sun Apr 19, 2020 9:35 am

You'd be surprised cobber.

pjifl
Posts: 883
Joined: Fri Jun 17, 2005 12:15 pm
Location: Innisfail, Far North QLD.

Re: Testing the Testing

#34 Postby pjifl » Sun Apr 19, 2020 10:18 am

Just to clarify a few things.

30 data points is often regarded as a minimum for reasonably high Statistical confidence levels. But obviously we cannot do that every time we explore variables in rifle shooting. Barrels are expensive items and an exceptional barrel needs protecting from excessive wear as much as possible. Luckily, we often have a lot of information from past experience whether from ourselves or others.

There is a real difference between 'proving' something and trying some possible tweaking. Proving something allows one to always subsequently rely on some general rule. This is where maybe 30 shots or more is called for.

As to ES and SD. One difference is that ES records a few events that have actually happened. SD on the other hand is a far better predictor of what is likely to occur during future events, as well as being better to establish a confidence level.

Not sure if everyone knows, but Hexta shoot results also gives a group size in terms of a 'Berdan String'. Well, HEXTA actually reports the average size of a Berdan String. This sums EACH SHOT'S actual distance from the centre - not simply the worst few as in our unfortunately flawed scoring system when high numbers of Xs and Vs are involved. The standard of accuracy has increased so much that, at the top level, targets have not kept pace.

Peter Smith.

Wal86
Posts: 319
Joined: Sun Nov 20, 2016 5:10 pm

Re: Testing the Testing

#35 Postby Wal86 » Sun Apr 19, 2020 2:45 pm

pjifl wrote:The standard of accuracy has increased so much that, at the top level, targets have not kept pace.

Peter Smith.


If only the target actually had a solid white X, results would be much better..

Cheers Alan

williada
Posts: 969
Joined: Wed Jul 22, 2009 12:37 am

Re: Testing the Testing

#36 Postby williada » Sun Apr 19, 2020 2:54 pm

Agree Peter, there a few issues to sort out with the electronic targets. That's why I only trust what's on paper. Reckon a few get their pockets picked in big events because of the small error margin or sensor issues. But the use of the electronic targets has kept our club alive on two fronts. One, it keeps those of us with a few disabilities in the game longer and secondly; it attracts the young people to our sport. I think without electronic targets our club growth would have not occurred and they have proven to be a useful training tool and assist record keeping. We deliberately have manual marking days to keep people in touch. Its at the butts were a lot of discussion about gear and technique takes place while waiting for shots to come in on manual days. Seems a lot of shooters never learn to predict where shots will go so they wait and wait to shoot on the same condition. Myself, I would like to see electronic targets geared to the old 45 second rule to develop wind reading skills which still sets shooters apart.

williada
Posts: 969
Joined: Wed Jul 22, 2009 12:37 am

Re: Testing the Testing

#37 Postby williada » Sun Apr 19, 2020 5:27 pm

To conclude my input in testing the testing for marginal improvements using statistics, my advice at basic level is to change only one variable at a time and if you are serious use 30 shots in a sample to determine whether for example, one primer is better than another. A fallback method rather than a one off test is keep good records from clubs shoots and overlay those to increase your sample size over time so that good gear can be utilized in the competitions for some benefit. Get a mentor who may have done the experiment properly. That is a shortcut and will save thousands of dollars over time. If you enjoy tinkering and want to know why something works use statistical methods to backup what you think. Making mistakes is the best teacher. One swallow does not make a summer, but it makes a good story. Have fun.

KHGS
Posts: 934
Joined: Thu Oct 20, 2005 12:46 am
Location: Cowra NSW

Re: Testing the Testing

#38 Postby KHGS » Sun Apr 19, 2020 8:46 pm

Wal86 wrote:
pjifl wrote:The standard of accuracy has increased so much that, at the top level, targets have not kept pace.

Peter Smith.


If only the target actually had a solid white X, results would be much better..

Cheers Alan


A solid white center wouldn't be any use to me :) :) :), I rarely aim at anyway!!!
Keith H.

GSells
Posts: 798
Joined: Sun Aug 13, 2017 11:04 pm
Location: Qld

Re: Testing the Testing

#39 Postby GSells » Sun Apr 19, 2020 9:17 pm

For me , running a extremely busy plumbing firm , family love for cars makes it very tough to test the testing. For example I’ll fire 5-7 to get a sample Sd , then group shapes and learnt experience and known nodes . I take a lot of short cuts because I have too! Either that or don’t shoot . Like I haven’t shot at Dalby must nearly been 12 months. Been too busy gallivanting around the country side shooting elsewhere. However once I think a load will hammer , I’ll test it at a local Opm . That’s where most times you will fire more than 30 shots and under match conditions. A great place to test and usually on ET’s , great to look back on . I know heart of Hearts that I don’t and cannot put in the time , money and effort that it deserves!! But better than not competing!
In conclusion of testing the testing.
I very much rely on learnt experience and then prove it in competition not at club ( I try to prove at club if I can ) and test the test in local competition!! That scope failure last year absolutely destroyed me for some time and my confidence still was shattered coming into Perth. This was very much highlighted with a 58.9 at Belmont!!
But Family, business, shooting don’t really mix and are always at each other ! Can be tough !
Regards Graham Sells Dalby / Tara RC .

Brad Y
Posts: 2181
Joined: Fri Jun 26, 2009 8:21 pm

Re: Testing the Testing

#40 Postby Brad Y » Sun Apr 19, 2020 10:22 pm

Out of interest, how many people develop their load for their barrel and prove it then test other combinations to find something better?

Gyro
Posts: 764
Joined: Sat Jun 10, 2017 2:44 pm
Location: New Zealand

Re: Testing the Testing

#41 Postby Gyro » Mon Apr 20, 2020 7:05 am

Have I got this right : the SD gives no significance to those ‘outliers’ in the group the way that ES does ? Just thinking here ( not easy ) but guess it does because those “ES outliers” do raise that total summed figure that gets used in the maths to determine the SD number. I’m dimly aware the SD will give the better prediction for future outcomes. But what about the outliers ? My brain tells me they need to be given more significance in any analysis of a group ?

I know a guy who has spent some time testing. Lots of time I suspect. Anyway one thing he did was number/indentify all his cases so he could see if it was the same cases repeatably causing them ‘stupid’ shots, as I believe My Boyer calls them. Now I absolutely don’t have a proper understanding of all his results but I do remember him telling me that he did notice some repeatability of some ‘feral’ cases and a good annealing of them ( to within an inch of their lives ! ) was beneficial.

A warning here re annealing : DO NOT ANNEAL ANYWHERE NEAR THE CASE BASE. If u want to anneal ( many don't ) just do the neck and shoulder and [ maybe ] a VERY SHORT DISTANCE past the top of the shoulder. Absolutely no more than that and remember you are allowed to quench the case in water directly the heat is removed, just so u can be sure the heat does NOT migrate down to the case base.

pjifl
Posts: 883
Joined: Fri Jun 17, 2005 12:15 pm
Location: Innisfail, Far North QLD.

Re: Testing the Testing

#42 Postby pjifl » Mon Apr 20, 2020 7:17 am

A calculated SD is of course raised by outliers. By all means look at them. But how does one decide if they are outliers unless the whole data set is examined. If you can localize a cause for the outliers the SD then becomes an even better predictor of future events.

Of course repeating tests on marked cases may be useful. It may show that you are below an acceptable tolerance or not.

Does Mr Boyer shoot past 200 yards ????

Gyro
Posts: 764
Joined: Sat Jun 10, 2017 2:44 pm
Location: New Zealand

Re: Testing the Testing

#43 Postby Gyro » Mon Apr 20, 2020 7:20 am

pjifl wrote:A calculated SD is of course raised by outliers. By all means look at them. But how does one decide if they are outliers unless the whole data set is examined. If you can localize a cause for the outliers the SD then becomes an even better predictor of future events.

Of course repeating tests on marked cases may be useful. It may show that you are below an acceptable tolerance or not.

Does Mr Boyer shoot past 200 yards ????


No he doesn't so the 'rules' for his work are very different to ours. But still related.

..... " localize a cause for the outliers " ..... now that's some really good advice I reckon. Then carry on with the statistics, which btw does make more sense to me now as a very good way to collate and analyse a bunch of information. Gotta be more thorough than some/lots of notes of what happened when ya went to the range ! Regards.

AlanF
Posts: 7495
Joined: Wed Jun 15, 2005 8:22 pm
Location: Maffra, Vic

Re: Testing the Testing

#44 Postby AlanF » Mon Apr 20, 2020 8:23 am

Peter and others interested. I have moved the scoring/SD etc discussion to a new thread HERE

wsftr
Posts: 202
Joined: Tue Jan 30, 2018 12:58 pm

Re: Testing the Testing

#45 Postby wsftr » Mon Apr 20, 2020 8:42 am

Brad Y wrote:Out of interest, how many people develop their load for their barrel and prove it then test other combinations to find something better?

Not me. But the next barrel I take what I have learnt and improve on that. Time, money and diminishing returns are the drivers.


Return to “Equipment & Technical”

Who is online

Users browsing this forum: No registered users and 41 guests