Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 12, 2014 12:34:23 GMT -5
Just to clarify, you don't have to get a character accepted for it to count, or even create a character. All you have to do is login with the password they email you. That is why some weeks, there are more new accounts registered than new characters created. This only backs it, week 26 of 2013, 142 new account logins, 106 pcs created (only 70ish of which were approved): www.armageddon.org/updates/index.php?week=26&year=2013 fakeymcfakerton. |
|
|
|
Post by lulz on Jul 12, 2014 12:39:14 GMT -5
Also, I get angry and snappish way too quick.  |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 12, 2014 12:42:02 GMT -5
Also, I get angry and snappish way too quick. Scorpio sun and taurus moon. I got horns and that venomous tail both. I can remain civil, but if it's not an obligation (someplace I am employed as staff), I am pretty quick to just stop talking to people when I get like that as I tend to say things I regret later. |
|
|
|
Post by BitterFlashback on Jul 12, 2014 12:44:07 GMT -5
I'm not clear on how you're getting that 41%. Could you ellaborate? in any case, I think i may plot some of this out tomorrow. I had an idea for a graph. 89 of 215 (the average weekly returning players in 2010). It's 41%. I don't believe that would be a count. if i understand what you said correctly, that would be hard cap on the maximum people you could consider vets who quit under ideal circumstances (waiting until the last moment possible in this context to quit). They're losing at a MUCH faster rate than Armageddon has ever lost, if you want to go there. Based on what? The data you've repeatedly insisted was utterly meaningless, except when it means Arm is doing well? talking out of two sides of your mouth isnt exactly a selling point for your credibility. And that is a problem given you keep claiming things without actually presenting a means to verify your numbers match what you say they are. Yeah see - they log into the game and type WHO. And they get actual counts. Just like you get, or just like your source (Nyr) gets. And that's where I found out how many people currently log in. The rest of the data I got from Elonka Dunin's history info (Elonka was a famous cryptographer who was also in charge of Gemstone for many years), the Simutronics data, the Krakiipedia, Wikipedia, a few logs of my own from when I used to play there, and various other sources that are directly related to that game. You got your data from - Nyr. You know, the guy who inspired this website, so that people could bitch about how wrong he is about everything. So "trust me, the data's legit, and no you can't look at it yourself because Im totes telling the truth and not just saying whatever makes my side look more supported." So then you are claiming that it's reasonable, and sane, and logical, and sensible, to compare Armageddon's "loss" of players, with Gemstone's loss of players. With Gemstone achieving a real actual number of loss, rather than just crunched numbers (10,000 during their heydayblah blah numbers blah Proof please. Except it wasn't an annual loss of 10%. And the reason it was 499, prior to the 2,000, is because their server couldn't accommodate more than 499 or it would crash. Prior to that it wasn't on the internet, and their server could only accommodate 49, and the 50th would crash the game (until they added a stop-gap to warn people attempting to log in, that they had to wait until someone else logged out). blah blah more unverifiable assertions without proof Still waitng to verify any part of your fanciful story.
A new account will not register as a "unique login" until it actually gets a character accepted and logs in.If that's true, there is no meaningful relationship you can derive from these numbers. As Ive said a few times now, you can estimate the conversion rate of account signups to players who made it in-game. This would indicate issues with Arm's application system, whether personal or technical. just because it wouldn't show veteran loss does not make the relationship meaningless. If you're going to keep claiming there is no relationship that can the two you're going to have to present something more compelling than just boldly rehashing the same statement. you are not recognized as an authority on statistics, and even if yuo were it would be an Argument From Authority fallacy to pretend you had no burden of proof to support your claim. |
|
|
|
Post by lulz on Jul 12, 2014 12:53:41 GMT -5
it would be an Argument From Authority fallacy to pretend you had no burden of proof to support your claim. I think at least one person on this board has been guilty of that on more than one occasion.  |
|
|
|
Post by BitterFlashback on Jul 12, 2014 13:02:14 GMT -5
it would be an Argument From Authority fallacy to pretend you had no burden of proof to support your claim. I think at least one person on this board has been guilty of that on more than one occasion. Oh, undoubtedly. But if they want to pretend this is a formal debate i may as well start holding them to that same standard. |
|
|
|
Post by fakeymcfakerton on Jul 12, 2014 13:11:00 GMT -5
A new account will not register as a "unique login" until it actually gets a character accepted and logs in.If that's true, there is no meaningful relationship you can derive from these numbers. As Ive said a few times now, you can estimate the conversion rate of account signups to players who made it in-game. This would indicate issues with Arm's application system, whether personal or technical. just because it wouldn't show veteran loss does not make the relationship meaningless. I don't believe you can estimate the conversion rate reliably either. It's possible that every single new account created either submits no character or the character gets rejected, i.e. a conversion rate of 0%. Alternatively, maybe every single new account created gets a character accepted, and the conversion rate is 100%. We simply don't know how many of those unique logins are from people who made a new account this week and how many are from older players. At this point we have to start making assumptions. Any model I can conceive of either makes too many assumptions to hold up to scrutiny, or has to start involving the previous weeks' numbers into the calculation of retention for the current week. The problem is that from the data alone, we have no way to tell "returning logins" at all, really. Imagine a scenario where Armageddon has 400 active players, split into two groups who play on alternating weeks. In week 1, there are 200 unique logins, and the same in week 2, but overall, 400 people in total played, and the "returning player" number for week 2 is 0. This is entirely unrealistic, but the point is the data alone doesn't give us any indication one way or another. If you want to present a model to derive conversion rate by the weekly update numbers though, be my guest. I don't claim to be an expert, just good enough to shoot holes into some of the conclusions being made here. Anyway, a much better way to calculate conversion rate would be the data from Nyr's monthly posts. For June, 182 new accounts, 51 logins, for a conversion rate of 28%. No assumptions are needed in that case, besides the assumption that Nyr presented the data accurately.
Anaiah, We had already established not every new account translates to a PC a while ago. Good job catching up. Also, you wrote "new account logins", not "new accounts created". I don't normally nitpick, but that's the very essence of what we're discussing, and the difference is key. The question at issue is at what point a "Unique Login" gets registered. It could be the first time a new account is created, the first time someone logs in with a password Armageddon sends, the first time they submit a character, or the first time they log in with a character. I firmly believe it's the first actual login with an accepted character that ticks that "Unique Login" counter up. This is based on the fact that I just logged in and entered my password with my existing account, having not logged in at all this week, and the "Unique Logins" number stayed at 211. If the trigger to increment that counter was logging into an account, it should have jumped to 212. I don't see why this behavior would be different for freshly created accounts versus existing ones. |
|
|
|
Post by BitterFlashback on Jul 12, 2014 13:18:16 GMT -5
As Ive said a few times now, you can estimate the conversion rate of account signups to players who made it in-game. This would indicate issues with Arm's application system, whether personal or technical. just because it wouldn't show veteran loss does not make the relationship meaningless. ... The problem is that from the data alone, we have no way to tell "returning logins" at all, really. I don't believe you can estimate the conversion rate reliably either. It's possible that every single new account created either submits no character or the character gets rejected, i.e. a conversion rate of 0%. Alternatively, maybe every single new account created gets a character accepted, and the conversion rate is 100%. We simply don't know how many of those unique logins are from people who made a new account this week and how many are from older players. You have gone with a meaning of "conversion rate" that has nothing to do with what I said. I statd that you could use this to estimate how many people who signed up for new accounts versus how many made it in game. you have pivoted to talking about how many accounts got a character accepted and stuck around, which is a much narrower definition than the one i stated this could be used for. |
|
|
|
Post by fakeymcfakerton on Jul 12, 2014 13:26:36 GMT -5
I don't believe you can estimate the conversion rate reliably either. It's possible that every single new account created either submits no character or the character gets rejected, i.e. a conversion rate of 0%. Alternatively, maybe every single new account created gets a character accepted, and the conversion rate is 100%. We simply don't know how many of those unique logins are from people who made a new account this week and how many are from older players. You have gone with a meaning of "conversion rate" that has nothing to do with what I said. I statd that you could use this to estimate how many people who signed up for new accounts versus how many made it in game. you have pivoted to talking about how many accounts got a character accepted and stuck around, which is a much narrower definition than the one i stated this could be used for. Is "making it into the game" not the same as "getting a character accepted"? If not, please let me know the difference. (Actually, these two are slightly different since I suppose not every accepted character is guaranteed to log in, but I don't believe either of us were making a distinction there.) I'm not saying anything about whether a character sticks around or not; simply if they made it into the game with their accepted character. The way I define conversion rate is as you said: (new accounts who had an accepted character log in) divided by (new accounts created) in the same time interval. From weekly update stats, we know that denominator exactly for a given week, but we don't know the numerator exactly. In fact, I can't even see any way to reliably estimate it. But if you're saying you can, please elaborate. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 12, 2014 13:30:21 GMT -5
Uh... yes, it is different. When 70ish of over 100 characters are approved, you still count as logging in with any of those characters, whether or not the character gets approved.
|
|
|
|
Post by BitterFlashback on Jul 12, 2014 13:39:56 GMT -5
I'll start from scratch and ellaborate. At this point talking about old conversation is just making things more confusing. and since the word "player" seems to have become ambiguous Im going to drop it for now.
We have the number of people creating new accounts. and iirc you can immediately apply for a character upon creating a new account. Theres no significant time between the two barring some technical issue. So it is safe to presuppose anyone creating a new account will move on to at least start a character application.
What the numbers allow us to estimate is the number of people connecting to the Arm with the intention of playing who do not convert into an active player. Becuase of the nature of the data we cannot narrow down where the problem occurs with converting people wanting to play Arm into people playing a character. it could point to problems with the application process in-game, poor turnaround times on approvals, etc. But the numbers do lend themselves to that relationship.
There would be a threshold of error as a result of the activity level of people who already have characters. But I dont claim these estimates are perfect.
i also agree with you that the new accounts who successfully logged in Nyr mentioned would make for more compelling results. I actually agreed with that the first time you mentioned it. that data would allow me to get the original aggregations i was trying to calculate. And I offered to add that data to the CSVs if someone else compiled it here for me.
EDIT: I believe some of this confusion is coming from the fact Id previously used the numbers to estimate pure attrition. i have revised the meanings of my estimates since your original post with Nyr talking about the more refined numbers. the post I did that in was a few pages ago so it probably got buried by all the bickering going on.
|
|
|
|
Post by fakeymcfakerton on Jul 12, 2014 13:41:09 GMT -5
Uh... yes, it is different. When 70ish of over 100 characters are approved, you still count as logging in with any of those characters, whether or not the character gets approved. I could submit a character today and get approved. Would the "Approved" number on the weekly update site tick up? We would expect yes. We already established that connecting to an existing account but not logging a character into the game does not tick up the Unique Logins counter. So assuming it was 211 when I submitted, and nobody else was playing, it would still be 211 when I was approved. (Unless the act of submitting a character or a staff member approving the character ticks up the Unique Logins counter, which I very much doubt.) Now imagine I wait until next week to log in and play said character. The totals for this week won't show me as having logged in, because I didn't. It'll still remain as 211. Unique Logins seems to be a measure of how many unique accounts, whether new or not, brought a character into the game that week.The "whether new or not" bit is key. We have no way to differentiate between the two, so the only thing the "New Accounts Created" metric gives is a possible upper bound on how many of the Unique Logins were from a new player, but it's no help beyond that. |
|
|
|
Post by fakeymcfakerton on Jul 12, 2014 13:50:16 GMT -5
What the numbers allow us to estimate is the number of people connecting to the Arm with the intention of playing who do not convert into an active player. Becuase of the nature of the data we cannot narrow down where the problem occurs with converting people wanting to play Arm into people playing a character. it could point to problems with the application process in-game, poor turnaround times on approvals, etc. But the numbers do lend themselves to that relationship. There would be a threshold of error as a result of the activity level of people who already have characters. But I dont claim these estimates are perfect. i also agree with you that the new accounts who successfully logged in Nyr mentioned would make for more compelling results. I actually agreed with that the first time you mentioned it. that data would allow me to get the original aggregations i was trying to calculate. And I offered to add that data to the CSVs if someone else compiled it here for me. I'm just not seeing how we can tell whether they did or did not convert to an active player with any accuracy at all. From the weekly update data alone it's possible they all convert, none do, or anywhere in between. Unique Logins tells us nothing by itself about how many new accounts converted, which is why I kept saying it's hard to relate the two. There's just too many unknowns. Player conversion using Nyr's numbers would be much more clear-cut anyway, as we both agree, and it would probably be a more useful discussion than debating whatever it is Anaiah is trying to prove. Since I've already spent so much time discussing this maybe I'll just head over to the ol' GDB and compile the data from Nyr's posts. |
|
|
|
Post by BitterFlashback on Jul 12, 2014 14:12:01 GMT -5
I'm just not seeing how we can tell whether they did or did not convert to an active player with any accuracy at all. From the weekly update data alone it's possible they all convert, none do, or anywhere in between. Unique Logins tells us nothing by itself about how many new accounts converted, which is why I kept saying it's hard to relate the two. There's just too many unknowns. This is one of those areas where opinion creeps in. i mean for everyone. Because the number of unique logins is relatively steady it comes down to if you believe 1) people are constantly quitting and constantly joining the game or 2) people want to play the game and either can't get in or quit quickly, but people who are already playing there are mostly sticking around. There may be other interpretations but those are the extremes. Obviously Im #2. I don't believe it to be likely that Arm is losing longer-time players at almost the same rate as getting new people to replace them. It seems much more likely that the bulk of the dead accounts are people trying to get in who are giving up and people quitting shortly after getting in. Older players are more subject to escalation of commitment (a.k.a "sunk cost fallacy") and sticking around because theyve made friends, so i find it harder to believe they make up the bulk of inactive accounts. but we are agreed that this isn't something we can be definite on from the numbers. EDIT: It's occurred to me there are other uses for the comparison which are extremely accurate. i hadnt thought of them before because i was stuck in the rut of thinking of useful-but-plausible interpretations of the data. The easiest one to explain is this one: They can be used to show how many people there were with the intent to play Arm versus the number actually retained over time. this ratio is only useful in demonstrating that the influx of new users is an insignificant thing to point to as a sign Arm is "doing well".
the problem lies in the fact that where the dropoffs are occurring doesn't make it clear where most players are getting turned off to the idea of playing. Id also need to make new calculations to get ratios i didn't previously derive. Which is probably another reason why I didnt think of it earlier.Player conversion using Nyr's numbers would be much more clear-cut anyway, as we both agree, and it would probably be a more useful discussion than debating whatever it is Anaiah is trying to prove. Since I've already spent so much time discussing this maybe I'll just head over to the ol' GDB and compile the data from Nyr's posts. That would be most appreciated. thank you. I'll plug them into my database and dump some CSVs so everyone has them in a usable format. |
|
|
|
Post by fakeymcfakerton on Jul 12, 2014 14:44:09 GMT -5
I'm trying to avoid calculations based on opinion or injecting my own opinions on Armageddon into my analysis. This board seems to have enough color commentary on its own without me needing to add any. I'll just stick to the facts.
I was only able to find about a year's worth of self-reported data comparing new accounts to logins on the GDB. There may be more, but it was tough enough to fish this much out of Nyr's recent posts and I lost interest. Below is what I have.
The third column is new accounts created that month, minus duplicate/staff accounts, as reported by Nyr. The fourth column is players who logged in for any amount of time, as reported by Nyr. The fifth column is the conversion rate, i.e. column 3 divided by column 4, to three significant digits.
June 2014 182 51 0.280
May 2014 224 51 0.228
April 2014 209 47 0.225
March 2014 279 72 0.258
February 2014 322 58 0.180
January 2014 191 45 0.236
December 2013 250 50 0.200
November 2013 305 61 0.200
October 2013 365 93 0.255
September 2013 434 88 0.203
August 2013 421 89 0.211
July 2013 406 83 0.204
June 2013 464 82 0.177
May 2013 488 97 0.199
April 2013 452 105 0.232
TOTAL 4992 1072 0.215
Overall, since April 2013, Armageddon has converted roughly 21.5% of new account creations into players who got a character into the game and logged in at least once.
Interestingly, although new account creations are down significantly in 2014 compared to the same months last year, conversion has actually improved. Every month in 2014 except February is beating the average conversion rate.
By the staff's own admission, many of these players who log in play for an hour or less. So the percentage of players who make it into the game and actually stick around for a significant amount of time (i.e. become something you might be able to call a "veteran") is lower - probably much, much lower.
|
|
