LuckyBancho data shows that FF14 is in decline

[Atelier-Bagur]

its definitely seeing some decline, but for me it comes down to how successful Dawntrail is. Im sure it'll sell well in its release but what happens after 7.0 and if the data still sees going downward without reaching the peak it had then yeah.

[Noumenon]

No, that’s… not how that works.
...
achievements are an accurate representation of playtime

So you picked the least consequential part of OP and laser focused on it and ignored the rest of the discussion? Which is clear because I've already remarked on how LuckyBancho's numbers represents a summary statistic that does not capture huge swathes of the activity distribution.

Actually yes, a comparison can be made between WoW and FF14, because there were other similar trends in WoW that now exist in FF14, like longer patch cycles, more bugs, etc.

And a comparison does not mean that the two trends are identical. It simply means that there is cause for concern. Hence "eerie".

We're not doing any rigorous statistical analysis here. There is not any regression or model or tree or whatever in this entire thread. We're merely discussing the data. And figuring out the context. Something that is currently done in the thread. This process always preceds actual statistical analysis. If this process is unfamiliar to you I don't think you have produced any quality data scientific work in your past.

And the fact that you think I'm conducting statistical analysis reveals a high level of non-understanding. I have not brought forth a single model or regression or MCMC or whatever. Anyone who actually understands statistics knows that I'm not actually conducting any analysis. Again, we're just discussing the data.

[Jojoya]

One important thing to note is this translated statement from LB's blog entry that contains the graph:

Therefore, rather than looking at the absolute number of the latest active characters, we have been using the changes in the achievement information of the characters whose achievements have been made public as a reference for looking at changes over a long period of time.

Not all players make their character achievements public.

For those of us who can't read Kanji, this is what the colors on the graph mean (translation courtesy of Google Translate - if you can read Kanji please feel free to correct the translations if needed).

Blue - New ("the number of characters whose first achievement was recorded during that period.")
Red - Continuation ("characters whose achievements were confirmed at that time, excluding new and final characters")
Green - Final ("the number of characters whose achievements after that period could not be confirmed.") (I interpret this to mean players that were active during the quarter but had no new achievements earned in the quarter as of the date the data was collected - I could be wrong about that)

I don't know if he collects the data only on the dats listed in his normal census or if he does additional collection on other dates. If it is coming strictly from the census dates, it explains the strange appearance of the 2nd QTR bar - the census for 2nd QTR would have been done on May 20th, 2 days before Patch 6.4 was released. The census prior to that was March 6th, just before patch 6.35 was released. It's not likely many veteran players would have earned new achievements between April 1 and May 20th because there was no new content with new achievements released during that period.

That there are fewer players active right now is known and it is normal for MMORPGs in the latter half of an expansion.

It is very possible that the game did peak with Endwalker's release and may be starting a slow decline similar to WoW.

So what? It's a 10 year old game, which is much longer than most games remain relevant. Think about the games you were eagerly playing 10 years ago (if you can even remember their names). Are you still playing all of them at least once a month? Probably not. Likely you got tired of them and moved onto other games then got bored with those other games and moved on again.

It doesn't matter how many other players are playing FFXIV. What matters is whether you personally are enjoying it right now. Some of us are, some of us aren't. Play according to how you feel about a game and not because of some information on a graph.

I think it's more appropriate to look at the population numbers during the previous expansions at the same time. As it stands right now, FFXIV is still seeing population growth (which Lucky Bancho even says). Now, I wouldn't be surprised to see a dip for DT, as EW was an end of a 10 year saga, and the more laid back vibe might turn some off, but we'll have to wait until then to see what happens.

I definitely expect Dawntrail to be lower than Endwalker. Many of those who were "testing the waters" following the WoW content creators have stopped playing.

That's something that's going to cause problems when it comes to predicting if there is a specific trend - the WoW Exodus turned the normal number flow into chaos.

[Saucy_Jack]

So you picked the least consequential part of OP and laser focused on it and ignored the rest of the discussion? Which is clear because I've already remarked on how LuckyBancho's numbers represents a summary statistic that does not capture huge swathes of the activity distribution.

Actually yes, a comparison can be made between WoW and FF14, because there were other similar trends in WoW that now exist in FF14, like longer patch cycles, more bugs, etc.

And a comparison does not mean that the two trends are identical. It simply means that there is cause for concern. Hence "eerie".

We're not doing any rigorous statistical analysis here. There is not any regression or model or tree or whatever in this entire thread.

No, you don’t even understand what I’m getting at. You definitely don’t know anything about data science

The time periods not being equivalent matters because even if the length of time is the same, the difference in actual time periods when the data was collected matters. 2013-2023 and 2004-2018 aren’t comparable, the industry has seen massive changes in game design, player expectations, quality, etc…. This isn’t even taking into account that the portions of the graphs that are similar have an even larger discrepancy

Sub counts and achievements are also not the same dataset, and you have absolutely no rationale for why we should consider one representative of the other - especially since achievement attainment tends to vary massively with how long the player has been playing and whether they’re doing new content they don’t usually partake in or are doing the same old usual

FF and WoW having similarities in other unrelated areas is not a good reason to believe the above two issues don’t apply in THIS comparison

“Eerie” doesn’t mean anything. It’s pseudoscience. Weak or strong correlation. Causation. Justify it.

We're merely discussing the data. And figuring out the context. Something that is currently done in the thread. If this process is unfamiliar to you I don't think you have produced any quality data scientific work in your past.

My first reply in this thread was that your post didn’t mean anything because you provided no specifics for any context, and your conclusions had no real substance for rationale. After which I actually brought up examples myself, that you proceeded to ignore, in favor of preaching some general nonsense about analyzing data “alone”.

You can’t gaslight like that when there’s a written record of the conversation we’ve been having, my guy

[Noumenon]

No, you don’t even understand what I’m getting at. You definitely don’t know anything about data science

Not sure why I'm inclined to listen to you when you have not engaged in any discussion beyond the WoW comparison, which I put in simply because the chart reminded me of it. There were a grand total of two sentences dedicated to it and it was also helpfully captioned as a "first impression". I then went on to discuss the details, the actual substance of the post. WoW was not mentioned at all after that. I don't think I shall take any advice from someone who seemed not to comprehend what an indicator function is. If you want to prove you know more about data science than me, feel free to engage in the actual discussion where I'm using stat jargon and reply to me that way.

Furthermore, it now occurs to me that you're intentionally laser focusing on two sentences in an entire post and outright ignoring the rest of the discussion, not even stuff from the OP like the different colors of each bar. I now think that perhaps, you want this thread to turn into yet another FF14 vs WoW debate, and unfortunately for you I shall not respond to you any further if you simply want this to become a WoW bash sess.

[Saucy_Jack]

Not sure why I'm inclined to listen to you when you have not engaged in any discussion beyond the WoW comparison, which I put in simply because the chart reminded me of it. I then went on to discuss the details, the actual substance of the post. WoW was not mentioned at all after that. I don't think I shall take any advice from someone who seemed not to comprehend what an indicator function is. If you want to prove you know more about data science than me, feel free to engage in the actual discussion where I'm using stat jargon and reply to me that way.

I’ve given you three examples from the very beginning of why the trends during the endwalker had exceptional circumstances you’re not taking into account (which you’ve ignored twice now), given you a whole list of questions regarding why it doesn’t make sense to equate your data with causation from expansion quality (which you’ve also ignored), and given you two examples of why the FF and WoW datasets you’re comparing aren’t directly comparable in my previous post (which you’ve again ignored)

All of these things above can be verified by you or anyone else here reading the thread by clicking two buttons, so I’m not sure why you’re trying to double down. You’re the one that keeps going on about data science experience and not actually addressing any points of substance

[Noumenon]

Not all players make their character achievements public.

You are correct. There is therefore a sample bias in the LB data. The rate of decline however seems to be similar to the overall decline (around 20%) according to LB's complete count (anyone who has a Lodestone change in general). So the total decline count trend is similar for the latest quarter.

The question is: do the proportions of the colors of the bar generalize to the overall population? I'm inclined to say absolutely not. Players who opt in to make the achievements public are of course more dedicated than the average player. So this data shows that the more dedicated players are more likely to "terminate" in Endwalker.

However, do the change in proportions of the colors of the bar over time generalize to the overall population? I don't see much reason to say no. For this to not be the case, we would need to violate the assumption that there is a sharp shift in the "innate" motivation or playstyle of players that is going in very different directions with a large magnitude between the two groups. While this could certainly be the case, I think it would take an extremely massive widening of the gulf between the two groups for us to not conclude that the number of terminations among the wider population is also increasing.

So what? It's a 10 year old game, which is much longer than most games remain relevant.

But does it have to be that way? No.

I’ve given you three examples

Already discussed the pandemic long ago in page 2. I've also already remarked that the WoW influx was a fluke. Your criticisms are already addressed in other comments and I am not going to repeat myself.

It's also quite obvious you didn't bother reading the OP anyways. The primary point of the OP post was about the change in distribution of "terminated" players over time. This has nothing to do with whether EW was the conclusion of a story arc. You are just whiplashing at the obvious story of player counts going down when I was exploring the details of the graph: the different proportions of types of players and how they're changing across expansions.

You've shown zero understanding of data scientific concepts, which is why I'm not bothering to refute you. You don't write like a statistician. Certainly my remark on indicator functions went completely over your head (and it's quite an elementary mathematical concept too). I'm not going to bother replying anymore unless you show aptitude.

[Jojoya]

However, do the change in proportions of the colors of the bar over time generalize to the overall population? I don't see much reason to say no.

As I said, the change in color appears to be related to the dates the data seems to have been collected. Lucky Bancho said nothing to indicate that he scraped the achievement data at the end of the quarter. He seems to have been working from the May 20th data.

That is something he would have to clarify. I couldn't find anything in the translation of his blog post that specific stated when he collected the data. If he was collecting on the final day of a quarter or the first day of the following quarter, there would be no need for a "Final" bar.

But does it have to be that way? No.

No, but I can't think of any 10 year old video games that kept increasing in popularity. It's really not that realistic when players are looking for the new best thing or improvements in graphics and game play thanks to technology achievements.

Even WoW's decline started in year 7. FFXIV's unexpected jump in popularity during Shadowbringers was clearly due to the content creator exodus from WoW, which swept players from multiple games in its wake. I can see Dawntrail starting off at the same numbers as Shadowbringers - and I don't consider that a bad sign.

The question I would like to ask you: what do you think you are accomplishing by posting this thread?

[Fawkes]

Thanks for TEDing to my come talk.

[Noumenon]

As I said, the change in color appears to be related to the dates the data seems to have been collected. Lucky Bancho said nothing to indicate that he scraped the achievement data at the end of the quarter. He seems to have been working from the May 20th data.

That only applies to the last bar which is why it's all green. I specifically discarded that outlier from the discussion. Even just looking at the first few bars of EW it's clear the ratios are changing.

In other datasets LB released, we see that the churn rate of FF14 is now at 20% per quarter.

And why I'm posting this is none of your concern.