Maybe ranking shouldnt be based on number of kills but on skill or kills ratio e.g.??

Good Idea.....any disadvantages of doing likewise? Btw does skill rating take into account what weapon was used to make a kill?

I dont find any. It really should be taken into consideration, becuase current rank is very imperfect.


AFAIK yes.

It shouldn't be too hard to show stats in a different way. The skill percentage
stored in the stats already calc's the shots vs kills. The weapons used is also
stored, and could be taken into account. Maybe I'll explore some of that for
the mod-servers, and see what results.

--- Moderator's note ---
This topic was split from the original thread as this is a separate discussion.

Nice split BTW BB.

May I also point to the most recent thread that discussed this.

http://scorched3d.co.uk/phpBB3/viewtopic.php?f=2&t=4442

Boy had some ideas that involved shots per kill, kills per round, round wins and rounds played. To rely on skill points only would be flawed to I think since if I'm correct skills also depends on the ranks of the players you are playing against. I may be wrong. If this is true the disadvantage would go to the folks that don't get to play with people so high ranked.

Mod-Servers currently testing Stats via Skill, instead of Kills.

So far, they show a fair level of the players actuall play ability, rather
than simply sitting day and night to collect kills.

Rather the reverse, actually.
After collecting higher numbers of kills, the Skill score will become
harder to increase due to the ratio needed to bring it up. Players
who are consistantly more accurate will be able to speak of thier
rank more proudly.

Of course, this is only a testing idea, and can be returned to the
older method just as easily.

Nice one Thrax.

How is the skill rating being calculated atm on the mod servers? (Always a controversial and tricky aspect)

All servers that record stats, collect many aspects of the game-play. The
database records the servers own Skill calculation, and I simply use it for the
ranking.
It's universal on all servers, and I believe its a ratio of #kills / #shots with
a few variances for weapon level used. You can look up on gavins stats, by
tapping the skill link on his own stats pages.

If you want a more apt description, we may have to ask gavin himself. I can't
find where in the source it actually calcs the "skill" stat.

I checked a few old stat series sorted by skill and to be honest I'd have to say that the word I would use to describe it would be arbitrary. Whether I am competent or not to say who is a good player or not might be argued over, but to me using the current skill system to decide rank would make even less sense than kills.

There are two major problems with using a skill based system.
1. What is skill and how do you quantify it?
2. How do you make it clear enough what you need to do to improve so that is doesn't just appear as some "random number".
I don't have a solution to either of these problems and the suggestions I have seen so far haven't really done it for me either. Until that is accomplished I'd rather have a flawed but clear quantifier to decide rank.

I know exactly what your saying man, but, as has created many a thread (and hijacked a few I dare say) how to actually work out skill is a very hard thing.

Kills/Death would be more relative than just kills, as long as that included resigns (be too easy to get a kill and resign etc..).

We definately need people to put their suggestions and some prime examples to show how it would work with known players and their stats.

To show how bad skill is <checks its still the same (and finds his example has changed somewhat> look at the current highest ranked player by skill. Seriously that arrogant 'person' is no way better than the majority lower down.
Then again, going by K/D ratio it's not much better.
BTW, Yes I know its very early days, but it still highlights the issues.

Hello,

Here in the U.S.A college football uses a computer to rank the teams in the top 25. One of the ways they rank is by strength of schedule(who you play). If you are ranked 10th and lose to a team that is ranked 50th in the country, your rank will drop signifigantly. On the other hand if you are ranked 10th and beat a team ranked say 3rd, you will rise in the rankings.

How does this apply here? If I'm ranked 8th here and kill noobs ranked 1000...my rank shouldnt get much better percentage wise. But if I kill opponents around my rank or better it should reflect that in a better percentage of rank. All I'm saying is this could be just one of the factors.

Heck, if that were one of the criteria...maybe noobs wouldnt be targeted so much. Thus they would stick around longer instead of giving up and never playing again.

Just one opinion... Bullseye

I am pretty sure that is one of the criteria B.E.

I am not keen on it though for this type of game as it's easy for a top player to cripple you (say u have shields etc..) and a newb to finish you. Also the newbs tend to use funks /hogs very early and get quick kills, they then end up last as they have no money, but you'd have to kill em, say 10 times in that game to make up for those 1/2 deaths. Not possible.

In theory it's a nice, uh theory, but in practice I find it a problem.

Currently the skill system already takes into account the weapon used for the kill and the difference in skill rankings between the killer and killed to calculate the skill difference.

But it can all get skewed depending whos playing at the time.