“Mind blow” I mean it’s all made up self deception!!!??
- Thread starter Utectok
- Start date
caorach
Well-Known Member
It is, and they make the point very well. It has been discussed on this forum many times, indeed we played about with the black dot of doom and other stuff and within the limits of our messing about we found that the stats and science was supported when people actually went out with rifles.
Fundamentally most reloaders are spending their lives fiddling with things they don't need to change, indeed their loads would be more consistent if they changed less random stuff, while measuring the wrong stuff in terms of performance. I've no idea why, or how it started, but it has lead to the weirdest of behaviour.
Fundamentally most reloaders are spending their lives fiddling with things they don't need to change, indeed their loads would be more consistent if they changed less random stuff, while measuring the wrong stuff in terms of performance. I've no idea why, or how it started, but it has lead to the weirdest of behaviour.
It’s captured by the term ‘Cargo Cult’ science.
I do statistics for a living, and almost everything I’ve ever seen about reloading on here is statistically meaningless. There are a tiny fraction who do it systematically enough, with large enough samples, to actually draw conclusions. The rest is chance.
The classic load development with 3 shots fired at each load in particular is laughably meaningless. I understand why people do it, but there is no way it can tell you anything at all.
Utectok
Well-Known Member
Clearly I’m Late to the party but it seems I’ve largely been following garbage advice on reloading which is kind of a relief really and it will save me a lot of time in the long run…..
Even seating depth seems to be a waste of time! Pick a good bullet good powder correct speed and shoot a 10shot zero!!
Last edited:
caorach
Well-Known Member
I started out doing what everyone else does just because it was what I was told to do and for a while I fiddled about before it dawned on me that what I was doing was completely pointless as I had gathered lots of pointless data and no information at all. Then I took to and shot 20 rounds at a target. The ammo was made up of a range of loads and bullet weights with, also, a selection of rounds seated to random depths. The bullet weights were between 110 and 165 grains and the loads were from book minimum to book max, with a few factory loads as well, and it was all shot from field positions including off sticks etc. From this I concluded that the group was probably about the same size as any 20 round group I was going to shoot no matter how many useless processes and outbursts of measuring I included when reloading. I also concluded that, almost without exception, the reloading advice on the internet and from so called ballistics or shooting experts entirely ignored physics and statistics. On concluding this I did some more critical thinking to find that some on this forum were claiming groups "from the bonnet of the land rover" that would be in the top ten smallest groups ever shot at bench rest. I also observed that one "expert" on the internet who was always keen to highlight that he had a PhD actually had a doctorate in social work and this explains why he used the wrong words for the basic physics all the time.
caorach
Well-Known Member
Spot on, saves a lot of pointless messing about with completely pointless processes and "data" that actually means nothing.
Yes.
The barest minimum sample size you can use in statistics and expect to be able to detect a ‘real’ difference between two samples is usually taken to be 6. With reloading, a single shot is not one of your 6 samples. Your group size is your ‘sample’ - so you actually need to shoot 6 groups at each load, and compare the average group size across loads.
It gets worse. You could get away with 6 groups at each load if you were only comparing 2 loads. But you’re not comparing 2 - you’re probably comparing 8-9. So you need to increase the sample size at each load (more groups at each load).
So - if you asked me to design an experiment to actually find the best load, I’d say you’d need to shoot 10 groups of at least 5, and preferably 10 shots, at each load.
And you would need to do that for every variable you changed.
Since no one does this (or nearly no one), you can safely say that pretty much everything they’re doing is statistically meaningless.
Very interesting, but to my mind it does tell me something significant. It says that if I'm doing my load development with 5 shot groups and I've found a 0.3" load, it's actually likely to be closer to 0.5". Is 0.5" good enough for hunting. Yes.
I'm not about to change my load development to 30 shots per charge weight, that would be madness and I would be ripping through cash barrels pretty fast - and probably not have enough time to actually go hunting. If I was a pro shooter looking to win the world championships then it would be a different matter.
I will continue with what I'm doing, but just be aware that 0.3" is not really 0.3" and probably most significantly ensure I do 5 shot zero's and another 5 after that before I twiddle any knobs.
I'm not about to change my load development to 30 shots per charge weight, that would be madness and I would be ripping through cash barrels pretty fast - and probably not have enough time to actually go hunting. If I was a pro shooter looking to win the world championships then it would be a different matter.
I will continue with what I'm doing, but just be aware that 0.3" is not really 0.3" and probably most significantly ensure I do 5 shot zero's and another 5 after that before I twiddle any knobs.
No.
It’s worse than that.
If you shoot a 5 shot group that comes in at 0.3”, it could really be anything. It probably isn’t, given the components you’re using. It’s probably somewhere in the range 0.3 - 1.5”. But with a 5 shot group you have absolutely no way of knowing. Really - none at all.
Chances are all the loads you shoot during load development are genuinely statistically indistinguishable (if you did the full 10 groups of 6 shots each at each charge, you probably wouldn’t find a significant difference between them). The likely exceptions are at the very top and bottom of the range.
So - each individual group is uninformative. But if you do your load development and they’re all somewhere in an ‘acceptable’ range (like 0.5 - 1”), then you can be reasonably confident that the sample is big enough to say that is the sort of group you can expect. So then just pick whichever charge you like.
Obviously almost none of us can afford to do load development in a statistically meaningful way. But that doesn’t negate the underlying maths. Your sample size is simply too small to say anything about your loads. So best stop caring.
Lies, damned lies and statistics…this is definitely not reason to change the way I do load development.
Every time I pick one of my rifles up and shoot 5 the group sizes are less than 1/2”. Don’t care what the statistics say, that’s consistent enough for me.
I do get the occasional 1/4” group but I’m sure that’s a statistical anomaly, but a nice one to have.
Utectok
Well-Known Member
That’s a pretty good summary of the pod cast
Freeforester
Well-Known Member
Seems to me that the barrel will probably have degraded by the time the ideal round load has been identified and verified…
Edit: for clarification, I am also a homeloader, having arrived at and settled upon my own preferred recipe thirty years ago, despite almost never living up to or realising its full potential myself
Edit: for clarification, I am also a homeloader, having arrived at and settled upon my own preferred recipe thirty years ago, despite almost never living up to or realising its full potential myself
Last edited:
Utectok
Well-Known Member
I remember once getting a load with an sd of 7 the i fettled the brass and got an sd of 2.5 both 6 shot tests and I thought well done me! Clearly I could have done nothing and got the same results. Or just kept shooting 6 shot groups with the same ammunition till I got an sd I liked lol
Probably because it’s just a summary of basic statistics re worded using reloading as a framing example.That’s a pretty good summary of the pod cast
First 5 minutes of first year statistics in any University course…
Ah - if every group is around 0.5”, and you’ve shot many groups (> 10), then yes, you can be confident that your mean group size is probably very near 0.5”.
What you have near zero confidence in is that it is any different to any of the other charge weights you might have tried.
You get your rifles custom made from the best components by someone who really knows what they’re doing, and you’re probably very consistent with your reloading. So it’s entirely probable that you have a 0.5” gun. But don’t kid yourself that your load development is playing any role in the accuracy you’re getting. You could almost certainly have chosen any charge below max and have got the same results over the long term.
What you're saying is that all load development is worthless, barrel harmonics count for nothing and charge weight/seating depth don't matter.
Sorry, I beg to differ. This is maths and statistics getting in the way of real life.
Utectok
Well-Known Member
I don’t think he is saying that at all. What he is saying is the differences are impossible to establish with a low margin of error using normal reloading practices. If you do test enough to get valid data the differences may be marginal.
“But don’t kid yourself that your load development is playing any role in the accuracy you’re getting. You could almost certainly have chosen any charge below max and have got the same results over the long term”
That’s the way it reads to me…
I’m saying that whether those things are important are not are irrelevant in the context of the way almost everyone does load development.
There may well be an optimal charge - but you certainly won’t find it with the procedure you use. It is also quite likely that the differences in group sizes as you move between charges are actually quite small.
As a thought experiment - imagine the best load for your gun gave a mean 0.3” group and the worst gave a 0.75” group. There will be spread around those means, and that spread will overlap. You could choose any charge weight between the very best and the very worst, and very seldom produce a group that was worse than 0.8 or better than 0.25. The majority would sit somewhere close to 0.5. And you’d be happy.
The point being that we imagine vast differences between the groups produced by the best and worst charge weights. That is unlikely to be true in a well made rifle. The difference may be so small as to be either undetectable or functionally meaningless. My guess is that in your rifles, the difference is tiny. Go on - what’s the worst group you can remember shooting (that wasn’t caused by pilot error or conditions)?
Similar threads
