LuckyBancho data shows that FF14 is in decline

[Noumenon]

If the thread is to discuss data and not the context behind the data, the reasons for the data, or what practical improvements can be made using the data, what do you expect to be discussed?

“Wow, look at that data.” “That sure is some data”. “This is some of the data of all time”

Or maybe you were expecting pages and pages of people saying “endwalker bad, game dying, look at graph” in creative ways as if that’s somehow worthwhile discussion?

I think if you perhaps studied some data science you would be enlightened by how much you can deduce from just the data alone, or perhaps you are ignoring Sindele's post and my reply completely.

[Snorky]

No offense but I don't think you understood that sentence or the data or anything really since that point is already addressed in the OP.

Omg no offense but I copy pasted a direct quote from that website. I'm sorry you think I'm to dimwitted to understand what I read but hey it's your thread. Have fun burning down the house. And.. data can be tainted in its collection the only true data or truthful data is held by SE period

[Saucy_Jack]

I think if you perhaps studied some data science you would be enlightened by how much you can deduce from just the data alone, or perhaps you are ignoring Sindele's post and my reply completely.

If you knew anything about data science yourself, you wouldn’t even attempt to imply that data should be interpreted without proper context. It’s an absurd notion that defeats the entire point of collecting it to begin with.

You never ever interpet data “alone” without context or rationale. That’s so comically wrong that anyone who knows anything about data analytics or statistics would scream

[remiff]

Although this kind of unofficial data should be taken with a grain of salt, honestly, it wouldn't surprise me if that were the case, given how there's practically nothing engaging for a certain target audience of players.

I wouldn't say that the entirety of the content is mediocre, but the longevity and rewards are lacking, which makes this kind of analysis credible.

The feeling of content weariness is something real and observable both on forums and outside of them through various means.

[Sindele]

If you knew anything about data science yourself, you wouldn’t even attempt to imply that data should be interpreted without proper context. It’s an absurd notion that defeats the entire point of collecting it to begin with.

You never ever interpet data “alone” without context or rationale. That’s so comically wrong that anyone who knows anything about data analytics or statistics would scream

To be fair, data analysts in games do that all the time.

It goes about as well as you might expect, but it does happen!

This is perhaps why I am so terminally skittish about trusting data that hasn't been cut into a thousand pieces so you can see all the weird angles. Data is suggestive and we have stupid lizard brains addicted to seeing patterns.

[Boblawblah]

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.

[Noumenon]

The situation is bad, but when it comes to making or predicting decisions based on it, you want to know what makes it this bad.

The biggest puzzle is the difference between Steam numbers and LuckyBancho numbers in terms of rate of decline (70% vs 20%). On a first order approximation, one can suggest that the explanation is due to players playing the game far less per day even though they still get at least one achievement per quarter. It is unlikely that Steam players are significantly different from all players so a sample bias cannot explain that discrepancy.

This explanation probably doesn't explain everything but Steam and LuckyBancho are basically producing two different summary statistics of player activity. Let's call player activity X. LB produces the summary statistic sum_1^N 1{X>=1}. That is, the number of players who cleared the minimum threshold of activity 1. Steam numbers are a proxy of X directly.

For Sqaure Enix, sum_1^N 1{X>=1} is probably the metric most relevant to their short term financials, actually probably most optimal, because they want people to stay subbed while consuming the least resources/content possible. A 70% decline in activity doesn't matter if everyone still subs.

However that is disastrous for the long term health of the game. X correlates with player interest and a binary check of whether X has passed 1 tells you barely anything for player participation and activity. While the decline in activity may be partially attributable to macroeconomic conditions, I don't think it's controversial at all to say that a large part of it is also due to declining player interest. I don't think any other expansion had such a large drop (I believe ShB was 40%) in activity.

The discrepancy between Steam and LB is something that only SE can fully analyze and explain, with their internal data, but nevertheless it paints a troubling picture for the long term health of the game: while players are still staying subbed, we can conjecture that macroeconomic conditions and player interests are increasing apathy in the game and declining activity. Such players have a higher propensity to quit once and for all in the future should things not change for the better.

And as a sidenote, many Silicon Valley firms overestimated the persistence of the pandemic shock and overhired, leading them to quickly fire once again. SE has done the same, but for servers. Let's be real: Dynamis has no reason to exist. The extra data center was a hasty decision made during the pandemic that is going to hurt the long term health of NA.

[Noumenon]

If you knew anything about data science yourself, you wouldn’t even attempt to imply that data should be interpreted without proper context. It’s an absurd notion that defeats the entire point of collecting it to begin with.

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. (Indeed, I would recommend any firm with data scientists whining about the lack of context to replace them with ChatGPT because your whole job as a data scientist is to figure out the context - typing lm(Y~X) is something a machine can do).

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.

And as a tangent, your attitude, along with popular slogans like "correlation =/= causation" are the hallmark of novice undergraduates who studied basic statistics for a semester and never studied much beyond it. Powerful inferences can be made even in the absence of enough context. The field of statistics has advanced much beyond the layman's understanding of the field.

[Supersnow845]

Why I’ve started looking at lucky less is that 1 achievement per quarter is far too little to be considered active, especially since the new ~~relics~~ award achievements for each step

If 1 million people log in for 1 hour a week in EW vs 800,000 logging in for 2 hours per day in ShB is that actually growth or disguised regression, this ties into a recent interview with yoshi p where he praised the fact that the manderville relics were the most completed relics to date and taking that line of reasoning further you could argue that they could get an even higher completion rate by having them as log in bonuses

When does “many people are doing nothing at all” become worse than “less people actively engaging in the game”

[NekiUno]

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.

But why are you so worry about it for? Many games will have these types of peak and valleys when it comes down to population. Besides if FFXI can continue to push on after 20 plus years and still going strong then FFXIV will be fine regardless of what happens.