FF14 player count reached the lowest point in 4 years. We lost 75% of the playerbase.

[BigCheez]

Sample size can be relatively small (1000 people or less) and still be quite accurate at predicting general trends, or it can be relatively large (over 100k people) and be next to worthless for making predictions. What matters is whether the sample you are looking at is statistically random or not. Unless we can confirm that people playing the game truly choose the method they use in a way that would be considered "random" (and/or that the "profile" of Steam gamers matches the profile of those who use other login methods), then it doesn't matter how "large" the sample is - it's next to worthless.

Are you actually trying to argue that Steam players are more likely to stop playing the game based on the fact that they bought the game on Steam?