Quote Originally Posted by anhaato View Post
I think the problem with this argument is that for a sample to be statistically valid, it needs to represent the whole well and not have any significant deviations.

I'm honestly not sure that Steam numbers do that. If you were to ask me why, I'd say that the Steam users are PROBABLY a lot of the WoW refugees whereas PS4/5 players (especially people that don't PC game) are likely a very different batch of people. Likewise, the people using the regular FFXIV launcher having bought it normally instead of through Steam are probably also a different mindset of player.

In order to use Steam as a sample, one would have to confront each of these arguments with data to suggest that Steam IS a representative sample, which I've seen no one do. People using the number are simply making an implicit assumption it is and not ever justifying this assumption other than "Surely Steam players are the same as the rest, right?" with no actual evidence to substantiate that - and yes, the burden of proof is on the people wanting to use the Steam numbers as a representative sample, not the people questioning that assumption.

Moreover, as people have noted, (a) this trend exist (using the Steam numbers, no less) in some form for all past expansions, (b) the measurement is of concurrent players, not total players (meaning it's weighted to people spending more hours playing and NOT the total number of individual people still playing the game just doing so fewer hours), and (c) even based on those numbers, the player activity is STILL higher than at this point in prior expansions and shows a net upward trend overall if you look at the entire Steam history of the game.

So, for those wanting to use the Steam numbers, not ONLY must you justify the claim that the Steam numbers ARE a good representative sample, you must explain each of points (a), (b), and (c) above somehow also leading credence to your hypothesis. Can you do so, I wonder...?