well first off it's stupid, and secondly it could be abused..
"hey guys im bout to die, kill me pls!!"
so.. no
I quote myself below just because detractors of the idea seem to have missed this:
something similar can happen with the current system: if you are fighting against a gang, and you are about to kill one of its members, maybe another of his gang-mates can kill him, because he was gonna die anyway, and at least the kill is scored by this gang-mate and not by you...
that, however, sounds like a good idea. not adding teamkills to stats. leaving deaths the way they are.
There will be
no selective counting of deaths/kills (e.g. if KillType is TeamKill then don't add).. the data is to be kept RAW. Classifying is OK, stripping data out permanently based on conditions is NOT, the kills/deaths is simply KILLS/DEATHS.
Server: "Kills [kills_here] but on [date_here] it was decided to stop adding gang kills to these stats."
Player: "Oh, how many kills have I actually made in total?
Server: "Who knows, someone decided it was a good idea to selectively stop counting some kills. Although these stats do include them before the date they were implemented but now they're all inconsistently aggregated together. Woo hoo, this is actually now statistically impossible to make a proper analysis."
The kill/death ratio is not a rating. It should not be treated as one like many players think it does. Is a player who has a higher ratio 'better' than a player a player with a 'lower' ratio? No; because the ratio is not a rating and does not classify different types of kills.
The intermediate point would be adding a classifications rather than selectively counting. This was what my original post was getting at. This way we could also add a stat named "Kill Rating" which calculates a rating based on the different class weights (e.g. driveby has a kill rating multiplier of 0.5, team kills have a kill rating multiplier of 0).
