The maintenance is complete and the PTS is now back online. The Gold Road Chapter – which includes the Scribing system – and Update 42 is now available to test!
The Gold Road Chapter – which includes the Scribing system – and Update 42 is now available to test on the PTS! You can read the latest patch notes here: https://forums.elderscrollsonline.com/en/discussion/656454/
Maintenance for the week of April 15:
• [COMPLETE] ESO Store and Account System for maintenance – April 16, 8:00AM EDT (12:00 UTC) - 6:00PM EDT (22:00 UTC)
• [COMPLETE] ESO Store and Account System for maintenance – April 16, 8:00AM EDT (12:00 UTC) - 6:00PM EDT (22:00 UTC)
Found some interesting ESO population data...
-
FluffyBird✭✭✭✭✭
✭mpicklesster wrote: »FluffyBird wrote: »200k, half of population swings in one day? All points within 1k of either 200k or 400k? Someone either gathered the data, or processed it, or made this graph with their butt. Most likely, all three.
Also, with such drops it averaged around 350k, not 400k before the last drop
If I had to guess, I'd say that whatever graphing app they use has some smoothing artifacts. (I do stats for a living, so I've seen it before.) Sometimes, in lower quality apps, the smoothing function tries to fit the data points to the increments of your Y-axis. So, for example, a data point of 247k would be smoothed to 250k; a data point of 253k would also be smoothed to 250k; etc.
It's not an uncommon problem for apps designed to make graphs for web browsers. The goal for those apps is usually to make a graph that will fit within a small window like a phone, tablet, or laptop. Unfortunately, some of those same apps use cheap smoothing functions (like the one I described above) to make the graph fit on small devices. So this means figures like the one in my initial post are only good for looking at trends, not absolute levels.
If you go to the original graph on the website, you can hover over points and see the actual numbers (which are slowly going up, btw). I forgot most of math stat from the university, but to me this whole thing looks like one big smoothing artifact. Also, if they "snap" to 200-s, that last drop doesn't hold more meaning than any before that and there's no trend then.
EDIT: even 5-year history is more informative, but it's more like "the thing is generally growing, here people sat at home, here they went about their business again"Edited by FluffyBird on July 30, 2022 5:42PM2 -
mpicklesster✭✭✭FluffyBird wrote: »mpicklesster wrote: »FluffyBird wrote: »200k, half of population swings in one day? All points within 1k of either 200k or 400k? Someone either gathered the data, or processed it, or made this graph with their butt. Most likely, all three.
Also, with such drops it averaged around 350k, not 400k before the last drop
If I had to guess, I'd say that whatever graphing app they use has some smoothing artifacts. (I do stats for a living, so I've seen it before.) Sometimes, in lower quality apps, the smoothing function tries to fit the data points to the increments of your Y-axis. So, for example, a data point of 247k would be smoothed to 250k; a data point of 253k would also be smoothed to 250k; etc.
It's not an uncommon problem for apps designed to make graphs for web browsers. The goal for those apps is usually to make a graph that will fit within a small window like a phone, tablet, or laptop. Unfortunately, some of those same apps use cheap smoothing functions (like the one I described above) to make the graph fit on small devices. So this means figures like the one in my initial post are only good for looking at trends, not absolute levels.
If you go to the original graph on the website, you can hover over points and see the actual numbers (which are slowly going up, btw). I forgot most of math stat from the university, but to me this whole thing looks like one big smoothing artifact. Also, if they "snap" to 200-s, that last drop doesn't hold more meaning than any before that and there's no trend then.
1.) The whole thing can't be a smoothing artifact. All smoothing functions have to start from an actual data point. Smoothing artifacts are ultimately just graphical deviations from an actual data point.
2.) While I guess it's technically possible that the smoothing function could snap to the 200s, it's highly unlikely. In all of my years of doing stats, I've never seen a smoothing function that vacillates that much. The cheaper ones tend to snap to the increments of the Y-axis (e.g., 50k)--which is probably the likeliest of scenarios.
As to why the actual data vacillate so much, I can only hazard a couple of guesses. One possibility could be the scaling coefficient they use to estimate player activity from activity in forums and threads. If the scaling coefficient is too large, then the differences between data points will be larger. Another possibility is that user activity actually does vacillate in a sinusoidal function--and that's not a remote possibility. Look at the data-to-day activity on Steam Charts. It also shows a sinusoidal pattern. To be clear, I can't confidently say which answer is correct. These are just my best guesses from experience.0 -
FluffyBird✭✭✭✭✭
✭mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless3 -
mpicklesster✭✭✭FluffyBird wrote: »mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless
I totally missed the figurative aspect. My bad.
As to the possibility of the graph being a complete processing artifact: that's technically a possibility, but it's arguably a remote one. As you said in a previous comment: "If you go to the original graph on the website, you can hover over points and see the actual numbers." Given that we can see their actual numbers, that eliminates the possibility that their graphing tool erased all meaning from their data.
This means the best candidate explanations for the data pattern in their figure are: a) imperfections in the scaling coefficient, b) a genuine pattern of intense vacillation in user activity, or c) a combination of the 2.0 -
MaraxusTheOrc✭✭✭✭✭FluffyBird wrote: »mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless
As the old saying in stats goes:
Garbage in, garbage out.0 -
mpicklesster✭✭✭MaraxusTheOrc wrote: »FluffyBird wrote: »mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless
As the old saying in stats goes:
Garbage in, garbage out.
That's not a saying in stats. That's an adage in computer science. It also hinges on the notion that the inputs be entirely made up. If that's the argument you're making now, then you really are grasping at straws.0 -
MaraxusTheOrc✭✭✭✭✭mpicklesster wrote: »MaraxusTheOrc wrote: »FluffyBird wrote: »mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless
As the old saying in stats goes:
Garbage in, garbage out.
That's not a saying in stats. That's an adage in computer science. It also hinges on the notion that the inputs be entirely made up. If that's the argument you're making now, then you really are grasping at straws.
I don’t know what kind of stats you practice, but *every* epidemiologist and PhD statistician I work with in clinical trials and real-world data uses that saying.
Update 35 very likely will impact player counts based on the qualitative information emerging around it, but it’s so weird an actual statistician would die so proudly on *this* hill with *that* data.1 -
mpicklesster✭✭✭MaraxusTheOrc wrote: »mpicklesster wrote: »MaraxusTheOrc wrote: »FluffyBird wrote: »mpicklesster wrote: »1.) The whole thing can't be a smoothing artifact.
I was speaking figuratively. I meant that even if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless
As the old saying in stats goes:
Garbage in, garbage out.
That's not a saying in stats. That's an adage in computer science.c
I don’t know what kind of stats you practice, but *every* epidemiologist and PhD statistician I work with in clinical trials and real-world data uses that saying.
Update 35 very likely will impact player counts based on the qualitative information emerging around it, but it’s so weird an actual statistician would die so proudly on *this* hill with *that* data.
I hear computer scientists say that phrase quite a bit. I've seldom heard it from statisticians (unless they lean more towards computer science).
But, in general, I find your claims highly dubious given your post history. You've demonstrated over-confident naivete and borderline statistical illiteracy several times. Now, all of the sudden, you're dropping aphorisms like "As the old saying in stats goes.." and talking about the epidemiologists and PhD statisticians you work with? It looks like you're just trying to sanitize your poor first impression with empty claims. Your first post in this thread was an act of lazy, hyperbolic trolling and now all of the sudden you're trying to cultivate an air of credibility. Your motives couldn't be more transparent.
In case you forgot, here's your initial comment in this thread: "I found data that says ESO has 20 billion players. I can call it data because it has numbers. It doesn't matter how the numbers were collected, what sort of biases exist in the methods or what limitations their are in the discussion of these numbers because the numbers support the narrative I am trying to forward."
How am I to believe much of what you say? As I said before, that's lazy, hyperbolic trolling.
P.S. You still didn't address my last comment that "[Garbage in/garbage out] also hinges on the notion that the inputs be entirely made up. If that's the argument you're making now, then you really are grasping at straws." So do you really think the data are entirely made or not at all correlated with actual ESO user activity? If so, that's a rather conspiratorial line of thinking.0 -
FluffyBird✭✭✭✭✭
✭Someone clearly has a problem of taking things literally...Edited by FluffyBird on July 30, 2022 7:49PM0 -
mpicklesster✭✭✭FluffyBird wrote: »Someone clearly has a problem of taking things literally...
Someone has a problem with being articulate. You previously said something nebulous like "if their data held any meaning, whatever they did during processing and visualisation, erased it, and the graph is useless."
I then 1) apologized for missing the figurative bits and 2) explained in detail why the facts don't support your assumptions. Did you show the same maturity and apologize for being wrong about something you said? No.
So maybe don't cast stones in a glass house.0 -
MaraxusTheOrc✭✭✭✭✭“[Garbage in/garbage out] also hinges on the notion that the inputs be entirely made up.”
That is just extraordinarily wrong. C’s get degrees I suppose. Poorly collected data is still garbage. Unvalidated methods have poor external validity. I can go on and on. You can come to my class if you want. I’ll let you audit it.
I’m so done with this thread. GG you win the message bored! This is just a textbook example of how misinformation becomes mainstream.Edited by MaraxusTheOrc on July 30, 2022 8:11PM0 -
mpicklesster✭✭✭
If true, that's a poor defense of your credibility.MaraxusTheOrc wrote: »“C’s get degrees I suppose.
You say you got C's then try to say you also teach a class. Sus...MaraxusTheOrc wrote: »"you can come to my class if you want. I’ll let you audit it.
Typical trolling. 1) opens up with lazy trolling, 2) argues energetically, but not in good faith, then 3) ends by acting like they don't care.MaraxusTheOrc wrote: »I’m so done with this thread. GG you win the message bored!
Please explain to me how it's misinformation? You haven't up with a shred of proof that the data are made up; nor have you given a reasonable argument as to why they're uncorrelated with actual user data.MaraxusTheOrc wrote: »This is just a textbook example of how misinformation becomes mainstream.
0 -
mpicklesster✭✭✭
^How to telegraph to someone that you know you're wrong without having the courage to say it and face it.FluffyBird wrote: »Whatever.
0 -
Psiion✭✭✭✭Greetings all,
After review of this thread, we have decided to go ahead and close it down as it has become mostly back and forth, nonconstructive Baiting. We understand different opinions are bound to be shared on the Forum's, and of course we welcome respectful discussion, but Baiting and Flaming are both against the Forum's Community Rules.Flaming: It’s okay to disagree and debate on the official ESO forums, but we do ask that you keep all disagreements civil, constructive, and on-topic. If a discussion gets heated and turns into a debate, remember that you should stick to debating the post and/or thread topic. It is never appropriate to resort to personal comments or jabs about those participating in the thread discussion.
Trolling or Baiting: The act of trolling is defined as something that is created for the intent to provoke conflict, shock others, or to elicit a strong negative or emotional reaction. It’s okay and very normal to disagree with others, and even to debate, but provoking conflict, baiting, inciting, mocking, etc. is never acceptable in the official The Elder Scrolls Online community. If you do not have something constructive or meaningful to add to a discussion, we strongly recommend you refrain from posting in that thread, and find another discussion to participate in instead. It is also not constructive or helpful to publicly call out others and accuse them of trolling, or call them a troll—please refrain from doing so. If you genuinely believe someone is trolling, please report the post or thread to the ESO Team, and leave it at that.0
This discussion has been closed.