Use median group size per kill instead of average. Or show both.

Looking at my own, it seems like I fight solo most of the time, but every once in a while there's a kill where like...six or seven people touched the guy who died. Usually these are in raid or retrieval situations.

Using median would produce a number that tends to de-emphasize these occasional "really large group" kills.

Like average numeric advantage per kill as below:

A kills B

Look at everyone who has been fighting B, and add them to A's group size.

then look at everyone who has been fighting A, and add them to B's group

Then look deeper recursively, so that everyone fighting someone that was fighting A is added to A's group, and people fighting those people are added to B's group and so forth.

Then you get a ratio X = GroupsizeA/GroupsizeB for each kill that can be easily aggregated. If you wanna be really fancy then to the pbf add a page when you round X to one decimal and make a frequency plot of kills per X.


Of course you have to go by unique players so each character can only exist once on each group size.

Is how many charges your character was worth. Maybe how many at their death, and the highest number during the characters life?

Cause really, that seems like a good system for telling how PK-successful a character is overall. But you never know, until you die to one of a handful of chars.

Also- I realize that for most people the median is going to be "1", "2" or "3".

That's fine. So maybe post both in the PBF.

Edit: an even better measure might be "% of PK wins that were solo".

Consider guy A who always hunts in a group of two. His gank-o-meter is 2.0, but 0% of his kills are solo.

Consider guy B who almost always hunts alone, but gets the occasional large-group kill. His gank-o-meter may be somewhere in the 1.75 range, but with a solo kill % of ~70%.

The difference between 70% and 0% is a lot more meaningful to me than the difference between 1.75 and 2.0.

you can do the math yourself.

I could calculate average-group-size-per-kill by hand too, but that's no fun either.

waaa?

My point was that the "gank-o-meter" score could also be computed by hand. So why do they pre-compute it for us? I'm just suggesting there are some other "interesting" stats that could also be pre-computed and given in an easy-to-digest form in the PBF posts.