well this makes perfect sense to me.
Also, it suggest you do not understand Bayesian statistical analysis.
Says the person willfully misreading my posts…okay. 
Edit:
BTW, you remind me of people who have a good understanding of statistical analysis see a result they do not like, then spend a great deal of time looking for any and all flaws to dismiss the analysis out of hand. People are very, very good at doing this. As I noted people do not like to admit they are wrong or made a mistake.
1 Like
That presentation and comments like these make it clear that CCP is convinced that the explosion of new players is not significant factor in their retention. Take that for what it is worth, but certainly the simple hypothesis that wars or ganking drive new players away from the game is not supported by the observed correlation nor the other data CCP has shared with us.
2 Likes
I see a lot of aggressivity and personnal attacks in your post. I suggest you take a deep breath and leave your keyboard for a while.
Your lack of self control is making this turn into a waste of time.
BTW, you remind me of people [didn’t read]
I could not care less about this kind of opinion.
You’ll end up talking to yourself, Anderson; o wait…
2 Likes
Ahh yes, I was thinking of that quote. I am bookmarking that. It is further undermines that the hypothesis: suicide/“griefing” new players drives them out of the game.
1 Like
And now back to our scheduled programming… or at least a related topic.
Concord kills are a fairly good indicator of the level of suicide ganking, the data being recorded by zkill.
A quick look at the numbers reveals that in August of this year Concord killed less people than they have in any month since the killboard data was consolidated in 2013.
Suicide gankers are a very very large percentage of those kills, that Concord kills hit an all time low recently and are still in that region is a very strong indicator that suicide ganking is in decline.
The following post in another thread expands on this decline using additional data.
Which begs the question, if suicide ganking is so cheap, why aren’t more people doing it?
Is it because it’s not as easy as the OP would have us believe? Is it something else entirely?
Suddenly, graphporn.
9 Likes
If you start with a hypothesis like “ganking hurts new player retention” and try to back that up with data and you find out that:
- Only 1% of all new players get ganked.
- Players who get ganked are more likely to subscribe…
Then yes, that hypothesis got completely falsified. This has absolutely nothing to do with causation if the data does not support the hypothesis at all.
2 Likes
What CCP did was try to find out a correlation between ganks and retention.
The experiment showed there was a correlation between the retention duration and the presence of a gank : they both increase together (simplified).
This does not prove any logical implication between the two, neither does it invalidate any.
That’s something you seem to not understand : the presence of a correlation between two events is NOT a proof of an implication between the two.
I showed you ANOTHER explanation that does not invalidate the hypothesis CCP tried to back up.
You do NOT invalidate that hypothesis. CCP only exhibited a correlation, nothing more : saying " ganking hurts new player retention is false " is also an hypothesis, you can not prove it from a correlation.
Please note that it does not show that ganking hurts retention either. It just shows nothing.
And even if CCP showed a correlation between ganking and reduced retention, would not mean there is a causation. You can not prove anything from exhibiting a correlation from data. it only suggest where you MAY prove a causation with a proactive study.
I can provide an additional explanation (which I don’t know if it is true) : Maybe there is an attrition war between gankers to get to the more juicy targets first, so the gank in itself is very easy but the competition with other gankers makes the survival difficult.
Another one : people enjoy the graphisms in low sec and stop ganking in HS.
Another one : people don’t like to wait for hours and enjoy incursions waiting queues more ( ^ _ ^ )
I’m sure people with more imagination than me can provide additional explanations.
Ever since Nietshe stole his thread, and CCP Falcon made a bad call and locked Ralph’s “duplicate”, i haven’t seen him anymore.
It’s a big loss, and Nietshe deserves all the hate for it.
2 Likes
Don’t look for what’s possible, but for what’s likely. Then apply reason to why one thing might be more likely than the other, by looking at actual reality instead of what you believe is true.
The first one is really bad, btw.
1 Like
Correlation is not “proof”(I dont like using that word in science since it actually only has meaning in mathematics) per se, but its also not “nothing”. Science uses correlation all the time as evidence, they just dont use it as a standalone piece. The fact of the matter is, the evidence is leaning towards the claim that ganking helps retention, and opposite the claim that it hurts retention. This doesnt mean that ganking necessarily helps retention, but it does mean that the claim that ganking hurts retention, is an unfounded and currently false claim.
2 Likes
Well this is not what I said. I said correlation is not causation.
Never as evidence of a causation. That is exactly the opposite of “correlation is not causation”
no, it means CCP did not find any correlation between “more ganking” and “less retention” in their data.
I agree however that the claim that ganking hurts retention is unfounded. It is as much unfounded as the claim that ganking increases retention.
Trends in data can only suggest possible relations, which then need additional work to prove any causation.
The fact of the matter is, the evidence is leaning towards the claim that ganking helps retention, and opposite the claim that it hurts retention
No. There is no evidence at all.
If I find a correlation between the sales of an illness medication and the deaths by said illness, does it mean that the medication make people die of this illness ?
No, it does not. The correlation does not show any causation.
1 Like
That is BS.
I’ve started several new characters out of boredom to try NPE or mine and everyone one of them had a CODE. alt attempt to gank them within the first 30 days of playing the game, most within a week or two.
And more to the point, it depends on which characters are considered.
All new characters including alts of veteran players, or only first time players. Half of my alts have never undocked.
You said “the presence of a correlation between two events is NOT a proof of an implication between the two.”
Actually, yes it does, to the point that proponents of AIDS have claimed that the problem with connecting Aids and HIV is that correlation does not prove causation, when infact the scientific community is almost unanimous with regards to their connection. The same can be said of the acceptance of Smoking and Lung cancer, despite the tabacco industry using the same adage as what you are proposing.
Youre confusing fact with evidence.
Something can be considered evidence to something else, without that thing being considered a fact or true.
Causation is a fact, but no one is saying that it is causation, and that weve somehow proved it. But the claim here is that player retention is hurt by ganking, and the evidence that we currently have, contradicts this claim.
With regards to your medication analogy, if the claim here was “This medication extends the average lifespan of the people with this illness, which is only 5 years”, and it turns out that 90% of the people who took the drug dies of the same symptoms as the late stages of that illness within an hour of ingesting the drug, sure, its completely possible that it was a mere coincidence, a fluke. Maybe those 90% just happened to have an aggressive form of that illness, and they all just happened to die one hour after taking the medication. Correlation doesnt show causation. But it is evidence against the claim that the medication is safe and extends the lifespan. Does this mean that its a fact? Of course not. But things can be evidence of something else, without demonstrating it to be true.
2 Likes
They are because they made more tests.
People who tell you that correlation between smoking and lung cancer does not mean causation are correct, and you are wrong.
more informations here (article of 1956 on lung cancer and smoke) :
Note they accept they may be wrong, even if they try to prove the relation , and statistical data show correlations.
evidence noun
facts or physical signs that help to prove something
You cannot prove anything from a correlation, it’s only a result of the point of view you selected to observe data. You can exhibit thousands of correlation, it will not prove anything.
CCP did not find any correlation, but finding or not a correlation is NOT an evidence.
It does not help to prove anything, so it is not an evidence.
You can not be honest and consider a correlation as an evidence of anything. I won’t describe all the reasons why this is impossible, it’s just basic science. A correlation is commutative, a causation is not.
A correlation is only an observation that tell you where you may find a causation relation. It does not help to prove, only to search. So a correlation can NEVER be used as an evidence.
If you ever try to justify something form a correlation(other than said correlation), then you are wrong.
That’s why people fall so easily for this fallacy : they can’t agree that a correlation implies, proves, NOTHING.
Actually it is evidence it is just not proof. So far, according to CCP Rise they have been unable to find evidence that ganking hurts new player retention. Further, their analysis shows postive correlations between retention and social activities.
You are again contradicting yourself here. Yes, observing a correlation by itself tells you very little. However if you have a theory that gives rise to hypotheses and you then go out and gather data and see that yes indeed the data show a correlation in line with the hypothesis it is seen as support for the hypothesis and thus the theory behind it.
Using a rather simplistic approach given that this is a message board. Suppose we have a hypothesis H, and ~H (not H). Then when we observe evidence (E), we can use the following:
P(H|E) = P(E|H)*P(H)/P(E).
Now, with regards to the Fanfest presentation P(E|H) is the probability of observing people being killed staying longer given the hypothesis, H, that suicide ganking drives out new players. If P(E|H) is low, then that reduces P(H|E). That is we reduce our belief in the hypothesis that ganking drives new players from the game and as a result we should put more belief in ~H, that ganking has little to no effect and may actually help retention.
Now, none of this is “proof”, but it is certainly “evidence”. Could it still be the case that ganking is “bad”? Sure. But unfortunately there is no empirical evidence, according to CCP Rise, to support this view. Is it proven? No. But if you are going to stand on statistical hand waving…well you won’t get far here. You might as well stop wasting your time and ours.
2 Likes
But Jonah…none of that is evidence. Data schmata…you can’t prove anything with that.
/end @Anderson_Geten mode.
How does CCP define griefing here? I know plenty of guys who quit once they learned about suicide ganking the hard way. Oops there goes several days/weeks/months worth of effort because the game didn’t sufficiently inform you that concord won’t always blow them up before they get you. Its one of the biggest things that make EVE a joke of a game for so many people. If it was something that was reserved for the hardcore and affluent instead of for new players like it is now, it could be a different story.
matter of fact day i unsubbed was the day i was griefed ganked while autopiloting a mammoth with an empty cargohold. Does that fall unders CCP definition of griefing? Definately a “well, F this ■■■■” moment for me. One of the best things about EVE is that you can play it passively and still have a life and gankers foolishly and selfishly have been taking that away from people and making EVE not worth playing.