SD and Scoring

[AlanF]

...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.

Peter,

When ETs are good enough and importantly affordable enough to become the only type of target in use, then I agree that a more precise way of rewarding accuracy would be good. How about SD of MOA from the centre of the target? For the benefit of spectators, the SD would need to be updated and displayed on screens after each shot.

It would make for some interesting changes to strategy e.g. deciding whether to keep sighters .

[pjifl]

It may seem counter intuitive, but applying SD to Radial distance has little sensible meaning in shooting. It must be applied to the X and Y components (either as mm or MOA) of the distances independently since Radial position also depends on direction. And what does a negative R mean ?

The military has used other stats criteria for this like 'Circular Error Probable' but I suggest we don't go there except to know it exists.

In this case the sum of all Radial deviations - or its mean - is of far more use. SD of the length of Radial error actually measures the density of shots in a circular donut 'edge'. It takes a lot of explaining. SD can be applied nicely to V and for example the weight or volume of cases or to the orthogonal components of position from target centre.

See https://en.wikipedia.org/wiki/Circular_error_probable

The history of the Berdan String is quite interesting - and very practical - and quite understandable. I find it interesting that HEXTA report it in their online results. But we are being distracted from the topic.

Peter Smith.

[AlanF]

It may seem counter intuitive, but applying SD to Radial distance has little sensible meaning in shooting. It must be applied to the X and Y components (either as mm or MOA) of the distances independently since Radial position also depends on direction. And what does a negative R mean ? ...

With current targets (both paper and ET), direction of shot from target centre has no bearing on score. It is the subtended radial distance only that determines which score ring applies. So are you in favour of rewarding vertical and horizontal accuracy differently?

[pjifl]

Alan wrote

With current targets (both paper and ET), direction of shot from target centre has no bearing on score. It is the subtended radial distance only that determines which score ring applies. So are you in favour of rewarding vertical and horizontal accuracy differently?

For testing, yes, but for scoring definitely not.

The Berdan String, or the mean of all radial distances does that perfectly for scoring. SD here means something else.

Peter Smith.

[AlanF]

Alan wrote

With current targets (both paper and ET), direction of shot from target centre has no bearing on score. It is the subtended radial distance only that determines which score ring applies. So are you in favour of rewarding vertical and horizontal accuracy differently?

For testing, yes, but for scoring definitely not.

The Berdan String, or the mean of all radial distances does that perfectly for scoring. SD here means something else.

Peter Smith.

Peter,

I would definitely like to discuss this further, very interesting subject, but as you've pointed out, its off topic so I'll start another thread hopefully later tonight (some "lockdown projects" to attend to) ...

[pjifl]

I have written a short document on the meaning of Radial SD in Shooting. I hope it helps. It should be - hopefully - understandable to a fairly newcomer on the meaning of SD.

https://sites.google.com/site/etargetco ... statistics

or more specifically

https://docs.google.com/viewer?a=v&pid= ... dkNjQ5NjJl

Hope it works OK.

[Gyro]

Hey Peter do you have a link for just the plain raw maths for say ES, mean/average and SD ? That for me is what works to enable my tiny brain to get a handle on this stuff and may work well for others too ?

And thanks for yours and some of the other clever folk here for the input. My statistics knowledge is not great ( understatement ) but lately it’s improving. Regards Rob.

[AlanF]

Thanks Peter. I have used SD extensively in various accuracy calculations, but for some reason did not think it through in terms of radial distance. Thanks for the very clear explanation.

Now, I asked you what was at the time a flippant question "are you in favour of rewarding vertical and horizontal accuracy differently?". But it got me thinking, it would be easy to do with ETs, and may be worth considering in terms of fairness and also to address some recent trends.

I would like to see horizontal accuracy rewarded more than vertical, for two main reasons :

1. Horizontal accuracy is always dependent in a large part on the skills of the shooter, whereas vertical accuracy can sometimes be attributed mostly to others e.g. when using a borrowed rifle.

2. Recent trends with rate of fire have devalued the skill of wind reading. This could be offset to some degree by putting extra value on good wind reading.

Thoughts Peter, anyone?

[John T]

Hello.

Skilful accuracy is measured by the combination of vertical and horizontal dispersion.

It would brave to say which of equipment and handling is the more important.

But for the last decade and more "equipment" has demanded the most attention. This is the determinant of vertical, considered by many these days to be the more important.

Alan, forgive if I misinterpret you, but are you saying that there should be some kind of score adjustment to accommodate the normal verses the rapid shooters?

The hall mark of the very best shooters in any discipline has always been those who can "wait and blaze".

It seems to me that the examples of Bill Hallam and Matt Peroz of decades ago are being followed by an increasing number of competitors.

It is far beyond me now to learn this skill, but if I were to advise a new shooter, I would say "train to shoot fast and accurately".

Regards,
John T.

[KHGS]

Hello.

Skilful accuracy is measured by the combination of vertical and horizontal dispersion.

It would brave to say which of equipment and handling is the more important.

But for the last decade and more "equipment" has demanded the most attention. This is the determinant of vertical, considered by many these days to be the more important.

Alan, forgive if I misinterpret you, but are you saying that there should be some kind of score adjustment to accommodate the normal verses the rapid shooters?

The hall mark of the very best shooters in any discipline has always been those who can "wait and blaze".

It seems to me that the examples of Bill Hallam and Matt Peroz of decades ago are being followed by an increasing number of competitors.

It is far beyond me now to learn this skill, but if I were to advise a new shooter, I would say "train to shoot fast and accurately".

Regards,
John T.

It is never beyond you John, it just takes practice, patience and commitment. It is a skill and it can be learned by anyone.
Keith H.

[pjifl]

I suppose you could give a prize for the best wind reader. But often a shooter will be lucky and find a good patch of wind which can vary enormously over a day.

Also, whether fast or slow shooting, one has to make the best height adjustments possible. Drifting close to the top or bottom of the X ring leaves almost no margin for errors in wind reading. Barrels heat up during a shoot - often with some elevation shift. We do not get unlimited sighters like in some countries. And because of AJ, wind changes may change height on the target. It is all interconnected and must be managed as a package.

The most practical change would probably to follow various Olympic shooting disciplines and use eTargets to divide each ring into 10 bands but I can hear the shrieks from here !

On a white centre - whether you central aim or not during normal shoots, almost certainly in coached and team shoots a central aim would benefit a lot from a white centre. The use of a black centre has made it almost mandatory that riflescopes need to to be $3000 plus to compete with the best. A white centre would change that significantly. Some day a scope maker will come out with a viable illumination system usable in the daylight that contrasts against black. Another $2000 I am sure. A few cents worth of white paper would actually be more effective.

Just my thoughts. Peter Smith.

[pjifl]

Gyro wrote

Hey Peter do you have a link for just the plain raw maths for say ES, mean/average and SD ? That for me is what works to enable my tiny brain to get a handle on this stuff and may work well for others too ?

I am not sure that a deep study of Maths definitions etc would be very beneficial.

Most people these days simply use a Spreadsheet like Excel to calculate SD. The mean is simply the average, and most consider the Extreme Spread to be what it says.

Although SD is minetuly different, one may consider it as a Root Mean Square (of deviations)- ie.

The Square Root of the (Mean of the Sum of (deviations squared)).

It is a pain to work out by hand although there is a shortcut. But use Excel or an Open Office equivalent program.


Extreme Spread. Most simply assume Extreme Spread is Max reading - Min reading. But we can be cleverer than that.

If the extreme spread is due to obvious outliers, a Histogram (plot of the distribution) will show this. But if a tidy Normal Distribution shows in the Histogram then ES can be predicted from SD depending on the number of events. This is a very useful technique and it helps one decide if ES is simply due to higher probability with more events or some extra extraneous cause producing an outlier. Events further from the mean that 3 Sigma - ie three SDs, are usually outliers. This can often be very hard to decide in practice. There are mathematical tests for outliers but simply eyeballing a histogram is often more useful.

This link is fairly understandable.

https://en.wikipedia.org/wiki/68%E2%80% ... 399.7_rule

[Gyro]

Just shot out onto my front deck and aimed at a tree across my suburb with my Kilo2400ABS rangefinder, that accounts for AJ in its calcs, so long as you tell it what the wind is doing. So at 1000 yards with a 15mph full value crosswind the range finder tells me the AJ value is a half an MOA.

That's shooting a 180 g berger hybrid at 2750fps from an 8 twist barrel. I guess the atmospheric numbers need to be factored in too ? That's what it says for todays grey overcast day anyway. Temp is maybe 18 deg c. Altitude maybe 1600' because we look over a lake and it is 1170' above see level. May interest some.

[pjifl]

So that puts you past the very top or bottom of the X ring.

AJ seems to also be sensitive to some head and tail wind. It is often a ghostlike phenomenon largely masked by other wind swirls and curls on a Range. But superimposed upon other random variations it can make the difference between an X and a 6.

Peter Smith.

[AlanF]

I suspect there's more to AJ than just cross-wind direction and speed. On numerous occasions in challenging wind I've noticed some targets where the wide shots are all in a nice horizontal waterline, while other targets shot at the same time show the classic "right wind raises" slope. It suggests that something like bullets/twist rates/velocities etc. has an influence. The online Hexta data shows evidence of this.