Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Stefan Stahl
Seituoda Taskforce Command
1116
|
Posted - 2015.04.23 14:28:00 -
[1] - Quote
Hey,
Warpoints are awesome. Earning warpoints makes you earn SP, so apparently they exist to incentivise desirable behavior. If we want people to e.g. use nanohives more often, we can always increase the WP gains from restocking ammo and people will start doing that more often. On the other hand WP are often cited as a measure of contribution to winning a fight. If you're not earning many warpoints you must be an AFKer and should receive no rewards after the match.
These two definitions are slightly at odds with one another. On one hand WPs incentivise people to get into the fight and do stuff, on the other hand they sometimes make people flock to easy activities such as killing random turrets with a damage amped Soma instead of actually contributing.
I'd like to propose a method of how to measure actual contribution to winning (public) matches and discuss how this should be used.
If the goal is to capture how much a given activity ("perform a kill assist") contributes to winning a match I'd suggest going the statistical route. Grab a few hundred pub matches from TQ, filter the logs by each activity, add scores for the red and blue teams and correlate the number of activities performed with the likelihood of winning the match. The result is a value that can be (very loosely) interpreted as "Performing this activity contributes X% to winning a match". (*)
I'm very sure that this result will not mirror the current values for WPs per activity.
My personal suspicion is that it'd be possible to find many more activities that actually contribute to winning a (pub) fight. Such as "Spending 10 seconds within 75 meters of an enemy" or "Getting shot at without dieing within 5 seconds" etc. None of these should be incentivised through WPs, but I find it likely that you'd find they contribute to winning a fight.
This knowledge could be used to more appropriately detect AFKers. WPs are a very bad measure of player constribution. Killing an armor LAV doesn't equal contribution in a 20 minute fight. Enforcing higher requirements on WP doesn't help either since that'd just make people farm WPs instead of playing the game. If as per the above method we had approximate knowledge of how much activities actually contributed we could introduce a new - hidden - metric that is measured in contribution/minute and tie AFK-detection to this. Alternatively WPs could be brought half-way between being an incentive to perform certain actions (personal opinion: nobody likes going AV, but somebody has to do it) and measuring contribution.
What do you think about this? Do you have other examples of activities where the current WP-rewards don't reflect the actual contribution to the fight? Even for activities that aren't currently rewarded (and maybe never should). I think killing turrets is the most over-valued. Surviving getting shot at is probably the most underrated.
(* I cut a few corners here for the sake of simplicity. Statistics nerds, forgive me.) |
Dropship One
DROPSHIP ONE OPERATIONS
2
|
Posted - 2015.04.23 18:00:00 -
[2] - Quote
Please stop beating a dead horse. |
Kaeru Nayiri
Ready to Play
996
|
Posted - 2015.04.23 18:10:00 -
[3] - Quote
Excellent analysis Stefan, maybe there is hope for this idea. The only true obstacle is warranting the open surgery of the code to add this in. Definitely worth taking a look at. If Rattati wants to crack down harder on AFKers, this might just be the ticket.
I'd also argue that you'd need some other counter balancing element to prevent someone being labeled as AFK when they are in fact acting as a sentinel. Over watching an objective from far with a sniper rifle comes to mind. While extremely situational, it's not wrong to do if it achieves the goal and deters the ennemy's approach. I am sure there are other examples.
Know what cannot be known.
|
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |