metric 252 equivalent

tabashir

Supporter
Supporter
Nice one Jon. Out of interest, not sure whether it will make a difference, but think the 'green' area (ie it ends up in the grass) maybe should be a larger area on the last one? You have a few fliers but these are limited to the corners not covered by the rings. Although you have to put a stop somewhere, it's perfectly possible (and maybe likely) that some will miss at the 3 o'clock, 9 o'clock etc positions.
 

little-else

Supporter
Supporter
AIUK Saviour
now the thing about missing the boss, esp with longbow is that there isnt usually an even distribution, the individuals errors or the spining of the arrows or any other cause makes the arrows all go off to the left or the right rather than evenly spread around the target until you get to a distance where the group is much bigger than the target to begin with. I'm not saying that anyone should be rewarded for flyers because that it an error outside the person's ability to hold a group but it would be interesting to see a representation of this and when chucking a flyer into the mix say once per 72 arrows how often other errors put that back into the near middle of the group. I certainly have done that quite often for real but not it seems as often as a well aimed arrow seems to do the opposite.
I thought I had posted it here but you can see some peculiar real groups shot by archers ( riflemen or whoever) of a particular ability so all in the 8 is not impossible and that can be achieved by all the arrows spread around the centre or by a small group in one place that has been achievd by pure fluke as the other errors counter themselves.
The NRA used to collect and use such data when it had a military as well as sporting purpose.
My original thoughts were that at short range the handicap system doesnt fit with the classification scheme and at long range the number of expected misses (regarding other bowstyles) will distort things so a best fit at 60yds to either 50 or 60m would be sensible to make things both workable and achieveable to someone who could expect to score 252+ at 60yds and what I have seen where such a scheme is used doesnt quite fit the brief.
Now any chance of agreeing a common standard? Plotting the fall of arrows by archers of different abilities could be very helpful, it is something we seldom do and amassing and collating such data will keep someone busy for a while.

In the last lockdown I shot in my garden at ranges between 18 and 30m on proportionate targets to represent distances of 30/50/60/70 and 90m and there are small changes in the scores achieved on say a 90m proportional target shot at 20m and one shot at 30m suitably scaled. I compared these to what I do in reality and there are again small differences but quite noticable when wind is taken into account (of course) but as I failed to keep a log of where the arrows went I cant say much that is useful to add to the equation.

in rifle shooting you sometimes have problems getting good groups at short range because the bullet hasnt "gone to sleep" before it gets to the target but stabilisies in flight and at a longer range the group achieved hasnt spread as the gyroscopic stability of the projectile is good as it slows down, particularly through the sound barrier. Not a problem we suffer from but vibration is.
 

ArcheryFox

Active member
now the thing about missing the boss, esp with longbow is that there isnt usually an even distribution, the individual's errors or the spining of the arrows or any other cause makes the arrows all go off to the left or the right rather than evenly spread around the target
Yes.
However...
In my handicap adventures I have considered trying to factor in what happens from having an off-centre group or an oval rather than circular group (i.e. enhanced L/R spread etc). I was also interested in seeing how much incorrect distance estimation could impact field archery scores*.
However, all algorithms rely heavily on the fact that everything is circular, and these circles lie exactly on top of one another. This allows us to reduce the problem from 2 to 1 dimensions**, eventually getting a nice simple equation that can be easily applied to many target sizes/distances/faces.
If we break from this symmetry then you'd need to do the calculations numerically using a computer - either like JohnLockley did in post #20, or with a numerical integration code - instead of having one direct equation. This is far less general as you would need to generate a different set of numerical tables for different combinations of face size, distance and handicap. As you allude to, this would take a lot of time and effort and likely be too daunting to users to be practically useful***.

My general approach to get around this is to acknowledge that archery does not reward tight groups in the wrong place.
At a competition you are rewarded on score - if you could have scored more by moving your group, well that's your fault!
As such, if we remove ourselves from their origin relating to group size (angular deviation) then the handicap scores can still provide a benchmark of your overall ability. You just have to acknowledge that your improvements in handicap can come from more than just a shrinking of groups.
As I allude to in footnote ***, there is a balance between a functional system that most archers/records officers can use, rather than a complicated system tuned to an individual archer. For a nationwide AGB scheme the former is the priority.


there are small changes in the scores achieved on say a 90m proportional target shot at 20m and one shot at 30m suitably scaled.
Aha!
Don't forget that if you were to be scaling everything correctly here you would also need to scale your arrow diameter (a key number in any handicap algorithm) not just the target size!!
If you wanted to compare properly you would need to be shooting arrows that were just 22% (= 20m/90m) of the diameter of your normal ones. Alternatively, if you were to take your scores from a scaled target at 20m you would need to be shooting arrows 4.5 times as thick at 90m to compare!
In fact - this is one of the reasons (compound especially) archers tend to shoot a noticeably higher handicap indoors when they use maximum diameter arrows on small targets****. If you want a better comparison across seasons I recommend experimenting with the 'arrow diameter' parameter in the spreadsheet above.


in rifle shooting you sometimes have problems getting good groups at short range because the bullet hasnt "gone to sleep" [...] Not a problem we suffer from but vibration is.
I have a little more to add here.
In archery if you have issues with contact with a rest/riser/etc. it will show up in poor groups at short range. The effect will disappear at longer ranges as the arrow stabilises, just like a bullet.
At the other end if your poundage is too low/arrows too heavy/distance too far, you will see the groups opening up excessively at long range as the arrows begin to parachute.
Comparing handicaps at a range of distances can help highlight any potential issues if there is a trend or large discrepancy.

*Fun fact I discovered - it is actually harder in terms of handicaps to get MB and GMB field (even if you shot all the targets on the flat) than it is to get MB and GMB target! (for gent compound at least, but also a few other bowstyles I expect)

**Happy to elaborate on this.

I really struggled with stats on my maths A level though, scraped it by doing well in the applied maths instead so may have to rely on your more knowledgeable folks in that area.
***One of the biggest issues I feel with this topic is that it is often poorly explained and the average archer finds it too daunting to even bother with. I try and point people to the slightly simplified explanation here so that they can hopefully get a broader idea of what the handicap means and why it can be useful.

****Again, if you use the spreadsheet and run the numbers you find that the only place max diameter arrows really make a difference is for compound archers on the indoor rounds where the size of the x ring becomes comparable to the size of the arrow. Anywhere else and you will generally pick up just one or two points across the entire competition. Given the difficulties of tuning thick arrows and changing setup this is why I usually recommend (non-compound) archers stick with one setup indoors and out.
 

JonLockley

New member
Interesting thread. BTW, the code does have the ability to make non-circular distributions and offset for wind etc. However, given that these are just arbitrary choices in the random number generator and not correlated to anything the archer is physically doing, I'm not yet convinced about what it's really useful for. I did think about including a distance parameter to see how scores might change from one distance to another, but as little-else says, there are other things going on which mean it's not just a straight forward scaling of the target.
 

ArcheryFox

Active member
Interesting thread. BTW, the code does have the ability to make non-circular distributions and offset for wind etc. However, given that these are just arbitrary choices in the random number generator and not correlated to anything the archer is physically doing, I'm not yet convinced about what it's really useful for. I did think about including a distance parameter to see how scores might change from one distance to another, but as little-else says, there are other things going on which mean it's not just a straight forward scaling of the target.
One thing you might want to do is try changing your Monte Carlo style simulation to use a proability function and integrate this over the scoring zone.
This would give you a closer estimate of the 'true' final score as if you shot an infinite number of arrows using your current code.
This is what I was getting at in the first section of my previous post.

Another thing is whether you have you thought about accounting for arrow diameter and linecutters at all, or are you using point data?
Adding something in here would be good.
Related to this would be to setting up the code so that you can drive it in a way to replicate the AGB or AA handicap schemes, and even invent your own.

I really like your visualisations. Matplotlib?
You could try a fill_between to get target face colours between your circles.

I think it should be possible to come up with some sensible parameterisations for a few effects.
Just spitballing here but:
- Splitting into a 2 axis Gaussian to examine L/R vs high/low effects.
- Adding a small random component in the -y direction to a central distribution to represent dropping the bow arm.
- Adding in the effect of wind - there must be some literature on how it scales with distance if it isn't just linear.
- Adding in a random x-y component to simulate the effect of shooting in gusty weather.

Even using is as it is and producing a plot of how much score changes as a function of distance off-centre of the group would be really interesting.
This could inform you how many points could be lost due to wind or incorrectly estimating distances in field.

I think there are a few interesting line plots that could be made in this way to give accurate descriptions of effects that we as archers usually discuss based on our 'feeling' of how much the effect has - you may find some unexpected results.

Food for thought, and all things I'd love to investigate myself if I had the time.
I'd just say keep playing!
 

JonLockley

New member
Thanks for the suggestions!

Yes, just point arrows at this stage, although adding a physical arrow size wouldn't be too hard.

The wind thing is a difficult one as it isn't a constant, and may also play a greater a lesser effect as the arrow decelerates during its flight. So I'm unsure of how to add an effect that's more useful than a simple offset function like you say.

I thought about how to use an integrated solution too. Maybe having it generated as part of a dashboard with all the parameters on, and then once your happy with the set up you can press a "Shoot" button to get the random number plot.

I'll look at the fill_between option - sounds like what I was looking for but couldn't find!

One thing you might want to do is try changing your Monte Carlo style simulation to use a proability function and integrate this over the scoring zone.
This would give you a closer estimate of the 'true' final score as if you shot an infinite number of arrows using your current code.
This is what I was getting at in the first section of my previous post.

Another thing is whether you have you thought about accounting for arrow diameter and linecutters at all, or are you using point data?
Adding something in here would be good.
Related to this would be to setting up the code so that you can drive it in a way to replicate the AGB or AA handicap schemes, and even invent your own.

I really like your visualisations. Matplotlib?
You could try a fill_between to get target face colours between your circles.

I think it should be possible to come up with some sensible parameterisations for a few effects.
Just spitballing here but:
- Splitting into a 2 axis Gaussian to examine L/R vs high/low effects.
- Adding a small random component in the -y direction to a central distribution to represent dropping the bow arm.
- Adding in the effect of wind - there must be some literature on how it scales with distance if it isn't just linear.
- Adding in a random x-y component to simulate the effect of shooting in gusty weather.

Even using is as it is and producing a plot of how much score changes as a function of distance off-centre of the group would be really interesting.
This could inform you how many points could be lost due to wind or incorrectly estimating distances in field.

I think there are a few interesting line plots that could be made in this way to give accurate descriptions of effects that we as archers usually discuss based on our 'feeling' of how much the effect has - you may find some unexpected results.

Food for thought, and all things I'd love to investigate myself if I had the time.
I'd just say keep playing!
 

tabashir

Supporter
Supporter
Minor tangent, I stumbled across a face generator for short distances that the folks at R-Core have made due to folks shooting in gardens that are shorter than a range would be:
I haven't measured yet to see if it is a standard "half the distance, half the face", but considering they also take into account arrow diameter and that they say that the faces are 'scorable', I'm thinking their computation may be a little more complex than that.
For a bit of fun, check out the 'Flat Earth Target Generator' at the bottom of the same page.
 

tabashir

Supporter
Supporter
The spreadsheet 'code' above is also available in python and I would be happy to share parts of it or ideas if you are interested.
Would there be any chance of the code in python please? I use the MyTargets app for tracking scores and recommend that to anyone that asks. It's a great well-designed app that also is open source. The major omission from it from my point of view is that it lacks handicap calculation so I'm aiming to add it.

Will have to convert from Python to Java or Kotlin, but I'm thinking that may be easier than starting from scratch. Even if I do end up writing from scratch rather than converting, it would give me a useful check to write some unit tests to make sure my code is working OK. Will obviously credit you.
 

tabashir

Supporter
Supporter
Hi @ArcheryFox , think I'm OK without the python code now. Was a bit daunted by the spreadsheet but making headway into understanding it now, despite as previously mentioned my less than great maths.
Awesome job on it though, really is a thing of beauty, nice one. Will still credit you with the calcs!
 

tabashir

Supporter
Supporter
Back to the original(ish) point of the thread, thanks to the help from ArcheryFox's pages, MyTargets now supports handicaps and will calculate the equivalent handicap for *any* round.

Since you can define your own rounds, this means you can make a 252 metric round and it will show your handicap, so archers will be able to track their improvement/progress towards whatever score that is defined as the target.

I added the 252 imperial into the list of predefined rounds as this is a more generally accepted round at this point, but can add the metric ones in too if there are a few that would find it useful.
 
Top