[Suggestion] Publishing WvW Player Hours
1 2
There are so many conversations in this forum about how everything is coverage, coverage, coverage. I don’t disagree with that, but I would sure like to see how important a role it plays!
Actually, I don’t know that we even need to worry about AFK’ers. Here are some of my thoughts on the matter:
1) It’s probably reasonable to assume the AFK’ers across all servers is relatively equal. (And, even if it isn’t, the difference is probably not statistically significant.)
2) Any efforts to track their time (via substitution or any other calculations) seem unnecessary if my first point is at all accurate.
3) This complicates the request for publishing player hours. I can appreciate a specific request (in fact, personally I prefer specificity), however, ANet is more likely (even if it’s unlikely to begin with) to do something that’s easier rather than something that may require the construction/implementation of any kind of calculations.
I think emphasizing the points in your TL;DR, beporter, is more important than trying to control for all the variables. I say, let’s try to get our foot in the door before we get the door slammed in our face.
I think emphasizing the points in your TL;DR, beporter, is more important than trying to control for all the variables. I say, let’s try to get our foot in the door before we get the door slammed in our face.
Agreed. Any additional information along these lines, however imperfect, would be a huge step forward.
OP, your language suggests that you think there’s a direct (and strong) causal relationship between player-hours and score. Even if Anet did publish player-hours, at best, you can establish statistical significance, but you still won’t be able to measure the causal relation.
That’s just a more technical way of saying correlation doesn’t imply causation. What contributes to the number of player-hours logged is much more important than player-hours themselves: skill, organization, morale, etc.
You might think that normalizing score by player-hour lets you measure skill, but the reality is probably a lot more complicated than that. Playing well generally keeps people playing longer (how many people want to wipe repeatedly for 3 hours?) and builds morale/momentum. So a server may have more player-hours simply because it’s more skilled, and the normalized metric might lead you to think otherwise, especially since many skill groups don’t really PPT.
At best, your normalized metric tells you how efficient a server is at PPT (though, it probably doesn’t even do that). Let’s not pretend it’ll say much about a server’s skill or organization.
Second Child
OP, your language suggests that you think there’s a direct (and strong) causal relationship between player-hours and score. Even if Anet did publish player-hours, at best, you can establish statistical significance, but you still won’t be able to measure the causal relation.
Of course not, but can you think of a better “next” metric to make available? All (or much) of the bookkeeping is in place already to track this proposed addition. And while I could link to literally hundreds of posts in these forums railing over how war score and PPT alone don’t imply skill (just participation!), at least with the addition of how much effort is being expended to earn it we’d be able to make that relationship explicit instead of simply entertaining lively, repeated (and largely pointless) discourse about it right here in these very threads. We could make things constructive by moving to discussion about what strategy, training and tactics has a measurable effect on the ratio of score to effort expended. My proposal is exactly about helping servers become better at WvW at the macro level.
That’s just a more technical way of saying correlation doesn’t imply causation. What contributes to the number of player-hours logged is much more important than player-hours themselves: skill, organization, morale, etc.
Yes, certainly. I don’t disagree with that at all. But by the same token I could ask you to provide “proof” that your claims about skill organization and morale actually matter. My point is; both of our arguments continue to be entirely rhetorical without actual data. We’re sitting here saying, “WvW is this or that,” without any data other than a single war score to support it, all of us. What I’ve tried to do here is provide a specific theory (“war score is strongly correlated with participation”) (something we all seem to suspect or believe already anyway) and a way to measure it (“divide war score by the number of hours logged by all players in WvW for the match.”)
At best, your normalized metric tells you how efficient a server is at PPT (though, it probably doesn’t even do that). Let’s not pretend it’ll say much about a server’s skill or organization.
Yes, exactly! But its seems like you’re reversing my intent. This is not intended to score a server’s skill (well, not any more than war score already does, in any case) as much as act as a catalyst for it. No, I want something that can be used at the social level.
This isn’t altruistic by any means. I’m loyal to my fake digital “world” and I’d like to see us play better. Personally I get frustrated watching easily defended towers fall because nobody was around to man a single arrow cart. What I’m personally trying to get out of this is a single number I can point to and say, “Look, if we’re serious about this, we need a shift of volunteers on watch these hours of the day where our participation normally tanks.”
We can measure the effect that specific change in our behavior has with this number! Crudely of course, since as you say, skill and organization and other things may contribute, but what if they don’t? What is different commanders on the same day in the same time slot with “the same” amount of player participation produce a PPT that is statistically “the same”? What if I’m right and those variances end up NOT being statistically significant in the resulting efficiency ratio? (I think this might be why ANet will never provide what I’m asking for— because it might reveal that WvW is really a complete joke as an actual game mechanic. Note that I’m not saying it’s a joke, but the data might.)
But regardless, from that social perspective the value of feedback is measurably profound. The amount, quality, frequency, nature and characteristics of the available information has a direct impact on people’s choices and performance. Providing more transparency into the supposition that, “if your world wants to win and WvW, you need to make this (as yet theoretical) ratio go higher,” seems to me to be a valuable thing to at least test, don’t you agree?
Right now with war score our only available bit of feedback, when we ask questions like, “Do we just need more people? Or do our people suck and they need more experience?” we can’t really use anything other than anecdotal battlefield experience as a guide, which obviously differs for everyone. We may not be able to answer very (m)any more questions definitively, but some of them will become more measurable with this. We can control for the “how many people?” part of the equation. If you read only a couple of forum posts here, that seems to be a huge part of WvW so I can’t help but think this is a particularly worthy candidate to be considering.
I’m not saying this is the end-all, be-all of improving WvW, but it seems like an easy first step, no?
No one has said this up to this point so I will: data is neutral; data is neither good nor bad. Data merely encourages informed decision-making. With this in mind, more information (data) could only be a good thing.
Yes, sometimes our reactions to data are severe (“I’m sorry, you have cancer” is terrible news but it still helps you make informed decisions). And it’s also true that data can be misinterpreted but at least that encourages discourse on the topic to which the data is relevant.
From my vantage point, there is no harm in publishing some kind of statistic regarding the effectiveness/efficiency of a server.
If I’m being honest, I think there actually may be a reason Anet wouldn’t want to publish this. Being able to see when your opponents are at their weakest (in terms of participating players) could be exploited. Although, probably not significantly more so than first-hand in-game experience of not seeing any enemies anywhere.
1 2