Mean radius or group size -- which one actually tells you more?

[2labs]

Genuine question, and I do not think the answer is obvious.
Group size is measured off two shots -- the two furthest apart. Everything else you fired gets thrown away. Mean radius uses every shot: it is the average distance of each hole from the centre of the group.
The argument for mean radius is that it is far less jumpy. Two five-shot groups off the same load can come out a third apart just from which five shots you happened to fire, because one wide one moves the extreme spread and nothing else does. Mean radius moves less, because every shot gets a vote.
The argument against it is that nobody scores a match on mean radius, and every shooter's eye is trained on group size. You can look at a target and know what a half-inch group is. Nobody has an instinct for a 0.180 in mean radius.
I ended up scoring both when I built Loadscope -- group size stays on the screen because that is the number people actually read, but the ranking runs on mean radius and vertical, because the ranking has to survive a flier.
Where do you land on it? Do you use mean radius at all, or is it a fix for a problem you have not got?

 

[JFrank]

Mean radius might tell me if the rifle wants to shoot, yet overall group tells me if it does without a driver error or ignition error.

If that makes sense..

 

[jelenko]

I would add a 'con' to the mean radius. It hides the shots that are out of the groups.

All I really care about is the size of the group - that's what controls my score.

 

[chromatica]

F-class competition uses mean radius for scoring. Every shot lands in one of the rings surrounding the bullseye. Each of the rings describes a band that has an upper and lower radius. All of the shots are counted in the final score.

 

[Varminterror]

In the immortal words of Al Bundy... or something like that... "ya CAN'T have one without the... OOOTHer... group size and mean radius, group size and mean radius...."

Flyers jump both group size and mean radius. Tight mean radii also correspond to tight group sizes. Large sample sizes can dilute the influence of a flyer upon mean radius, but they can't hide in group size. Larger and larger sample sizes SHOULD normalize the mean radius where "flyers" repeat and the group fills in - OR - we find out there are some bad behaving outliers with inherent biases where the ammo isn't following "NORMAL" behavior...

Bigger mean radius in a small group size mean the groups don't want to behave, and bigger group size with small mean radius mean either the group is undersampled OR there's some systemic biasing influence.

"ya can't have one without the... OOOOTher...."

 

[SeabeeKen]

Sounds like mean radius was developed by a guy who wants to make his groups smaller with math rather than at the shooting and reloading bench. When i put 5 rounds into .030 the outliers mean little

 

[chkunz]

I use both but I value mean radius more for the reasons you state. I also look at the shape of the group and if is a nice well rounded group all is well but if shows stringing or other issues I also learn something from that. I also favor 10 shot groups and usually take the average of my best two ten shot groups for my judgement of the rifle and ammo combination.

The mean radius calculation starts to present problems once the groups start to get very small and you can not identify all the shots in the group. The ShorMarker statistical analysis feature helps with this and also eliminates the extra time in calculation of mean radius vs extreme spread.

 

[wkdickinson]

I think comparing "group size" to "mean radius" is like comparing "extreme spread" to "standard deviation" for velocity. ES only considers the two extremes, but SD considers all of the shots. Kind of comes down to your discipline, as illustrated in Chromatica's post above (mean radius), where as in BR "group" competition (group size).

 

[TC260]

I don’t do any competitive group shooting so the main reason for me to measure a group is for comparison purposes in load development. For that i prefer to measure every shot with mean radius. It does take a little while to recalibrate expectations and it’s a waste of time to tell anybody at the range what your MR is. Virtually guaranteed they have no idea what you’re taking about

 

[ShaunBrady]

The thing that's missed with rigorous statistical definitions is they assume purely random variations.

If I was procuring a million rounds, I'd certainly use mean radius of large samples for the QC.

If I'm trying to shoot smaller 5 shot groups more consistently than my buddies, I'm going to start with cheating on the random assumption and cull the obvious outliers from the components. Now the process isn't as random as it was because I've knocked the tails off the curve. Marking the cases that produce the outliers for later checks is simple and effective. No, it won't always be the cases or obvious if it is, but it's a good place to start. Making a note on a case if something seemed off during loading is another simple time/cost effective tactic. The outlier shots contain the densest information.

 

[M-61]

Sounds like mean radius was developed by a guy who wants to make his groups smaller with math rather than at the shooting and reloading bench. When i put 5 rounds into .030 the outliers mean little

Yes.
Mathematicians are like statisticians. They will show either what they want and always what the person paying for the service wants.
Not that long ago I was shown how 40%+10% doesn’t equal 50%, in a particular case.
For shooting just show me the target if you’re shooting for groups.

 

[Clancy]

Sounds like mean radius was developed by a guy who wants to make his groups smaller with math rather than at the shooting and reloading bench. When i put 5 rounds into .030 the outliers mean little

THANK YOU

 

[Clancy]

Mean radius is almost as embarrassing as a guy posting his chronograph readout instead of groups.

Group size is the only thing that matters.

 

[M-61]

Mean radius is almost as embarrassing as a guy posting his chronograph readout instead of groups

After reading this site for a few years I actually thought there was a shooting match where that was done. All talk about chrono speeds and consistency but I never read or see a group.

 

[M-61]

It does take a little while to recalibrate expectations and it’s a waste of time to tell anybody at the range what your MR is. Virtually guaranteed they have no idea what you’re taking about

Wow. What a shocker! Why do you suppose that is?

 

[TC260]

Come on guys, it's a perfectly reasonable thread topic with perfectly reasonable answers thus far. There's no need to start casting aspersions at each other. Grow up.

 

[M-61]

Come on guys, it's a perfectly reasonable thread

Not realistic. Another way to arrive at the answer YOU want

 

[Clancy]

I’ve played just about every sport and shot just about anything that shoots.

The two hardest things I’ve done are hitting that damn baseball and shooting sub .250 groups.

Lots of guys come up with ways to skirt a lack of precision vs developing the skills to get it done. Others try to buy it.

I’ve seen some extremely humble setups shoot very small. No need to spend 20K on a reloading setup.

My point is…. Tiny groups are hard. There’s a lot of pride in getting to the point where you can shoot them consistently. Don’t short yourself by finding an easier mountain to climb.

 

[ELR LVR]

I would add a 'con' to the mean radius. It hides the shots that are out of the groups.

All I really care about is the size of the group - that's what controls my score.

This ^^^
I believe mean radius is more meaningful for hunting.
Since mean radius gives you a good prediction where all your shots may land, it would also be good for Score type shooting
---
While Group size being more meaningful for Competitions that score by group size itself
----
Mean radius means nothing for guys that try to shoot the tightest groups
Where if all I really care about is landing inside the 10 ring, my group size is not so important, even though we are trying to still shoot a tight group. - Mean radius may be a useful tool there.

 

[TC260]

Not realistic. Another way to arrive at the answer YOU want

Oh Lord. Alright my friend, enjoy your afternoon