EVE General Discussion

^^ Opening salvo detected, can the target tank the alpha or will he explode?


Place your bets please.

Actually it did. There is part where CCP Rise (?) says that alt accounts were not included. Maybe a few slipped through, but it sounds like they made an effort to weed them out.



Correct, but the correlation is exactly the opposite of what one would have expected:

Initial hypothesis/assumption/belief:

Ganked => leave the game early.
Not ganked => plays the longest.

What is actually happening in game:

Ganked => plays the longest.
Not ganked => leaves the game early.

The results were exactly the opposite of what many people assumed. The data and the results should automatically make one rethink the initial causal relationships.

Let me put it this way. A researcher comes up, intuitively or based on a theory, a causal pathway for something. He then goes and gets data and checks it against his initial hypothesis/beliefs. The data are exactly counter to what his hypothesis/beliefs are. The research should then re-examine his intuition/theory. Making up excuses to not do this is bad. Hell CCP Rise and his team were surprised just at the number of new players ganked in their sample. And then they were surprised again by the fact that those in the ganked group stayed with the game the longest.





Whatever. You offer no substantive criticism at all.

Apparently not. Salvos took a class you know.

Well maybe he did. We don't really know.

You can get beer delivered? Damn.

Actually, I'm predicting this thread will finally be put out of its misery. The actual topic was left behind a long time ago.

I'll do my best.

Salvos,

Your bourgeois attitude is reflected in your post-deconstructionist belief structure that is hampering your ability to view this issue from a neo-syndicalist feminist perspective and just further reinforces your textual desublimination as being unsustainable. You really should start over and use a post-marxian and prepatriarchial feminism as the basis of your argument.

Nice text, can somebody explain it to me? I'm over 50. *roflmao*

It is just some gibberish I threw together from a postmondern text generator. It is actually harder than it sounds.

IIRC McDonald's sells beer in France and Germany. That's what 2000 years of civilisation and the Romans do for you

*rubs some salt in*

I'm bavarian, sometimes I have beer with my breakfast!

Wow....

Yes, 1% of the sample were ganked. That right there should tell you something:

-That ganking is not a significant problem for new players. You have about a 1% probability of being ganked in your first 15 days.

And the result is not invalid simply because looking at the categories allocates a small percentage fall into one category.

And actually, what CCP Rise said in the presentation suggested that those ganked and killed legally stay about the same. No he did not produce any sort of statistical tests, but he did say that while those who were ganked stayed with the game longest, those killed legally stayed almost as long. Now, if we had the data we could take those estimates and construct confidence intervals (if we stay inside the frequentist paradigm) and see how much those intervals overlap. If the overlap is large enough--i.e. each interval contains the point estimate for the other group, then those two point estimates are statistically identical. So it may not be a mere 1% but more like 15% or about 12,000 out of the 80,000.

The conclusion, which is not proven here but is an opinion based on the numbers and statistics, is that player-on-player interaction is good for retention. This is even true for ganking. It would be nice to see if other interactions have the same effect but they are harder to isolate. You would probably need to construct a sample and then follow those players over time and collect data on their interactions with other players...essentially construct a panel data set. Then we could look into building fixed effects models and random effects models. My guess/hunch/belief is that player-on-player interaction has a positive effect on retention, even if that interaction is "negative" i.e. combat where there is a ship loss.

But don't mind me, I'm just a dummy when it comes to statistics.