Quote Originally Posted by Noumenon View Post
I'd encourage you to actually read some papers because no one can get the full context behind any data.

If you can ever get the "proper context" handed out to you freely there is zero reason for you to exist as a data scientist - a machine can take over your job.

The challenge is in interpreting data while information is scarce. That's the difference between L6 data scientists and top school professors and reg monkeys.
No, that’s… not how that works. The fact that your dataset won’t ever be perfect doesn’t mean you should start trying to jump to conclusions instead of very rigorously picking apart possible bias or errors

For example, in your OP. You compare that luckybancho graph of players who gained new achievements with WoW sub data, pointing out that the shape of the graphs are similar. This alone is a giant red flag that you don’t know what you’re talking about here, because that doesn’t mean anything. It’s not an equivalent comparison, the time periods being compared don’t even match, and there’s no reason to believe the underlying factors are the same. This is the equivalent of WSB gamblers overlaying random graphs on top of each other to pseudoscience about what the next big market trend will be.

Even if you automatically assume that the datasets are directly comparable or that achievements are an accurate representation of playtime, which you really shouldn’t, you don’t actually take into account any context regarding the unusually sharp increase in post-shb player count (such as Covid, SL issues with WoW, end of a 10 year long story, as I’ve already mentioned). Should we be expecting EW subscriber counts to follow the same trends as previous expansions? Was the sharp increase in player count and sharper than usual decline expected due to unusual circumstances? If so how much? How does that compare to the actual rate of decline? Is there any rationale to consider expansion quality a cause of this trend, or is it just a correlation?

I’m going to be frank. I think you’re just spouting bunk pseudoscience and dressing it up as some genius statistical analysis.