Pages

Monday, July 7, 2014

Graphing EVE Online History: From Retribution To Kronos

Every once in a while, I like to play with graphs.  Usually the graphs are related to the illicit RMT markets, but I do like to pull the data from Chribba's EVE Offline website as well.  With the calendar now open to July, I thought this is an opportune time to look back at the average concurrent user numbers (ACU) for the past six months.  Then I realized that I needed to go back to the Retribution expansion to put events into context.  So I did...




Retribution, which launched 4 December 2012, was a popular expansion.  Launch week saw the average number of accounts logged in jump 12.6%, and the activity stayed above that mark for most of the time until the launch of Odyssey.  Perhaps surprising to many, part of the success of Retribution is attributable to the pilots in factional warfare.  In the early weeks of the expansion, the Gallente militia made an all-out effort to gain 100% control of the war zone, succeeding in January 2013.  While the Caldari militia would gain control of half the war zone again within a month, the Gallente effort showed that a group that historically did not like to plex would if the rewards were high enough.

Factional warfare also had a hand in the biggest single event of Retribution: the Battle of Asakai.  The battle started as a small affair over a small POS orbiting a cobalt moon owned by the Liandri Covenant of the Caldari militia.  In the escalation that followed, Goonswarm (or at least DaBigRedBoat) chose to back the Liandri Covenant while Pandemic Legion backed the Gallente militia in the form of Drunk ‘n’ Disorderly, who were also allied with Lost Obsession.

The second half of Retribution saw a decline in activity as the big null sec powers began positioning themselves for war.  Well, at least TEST did.  The number of accounts logged in declined from an average of 35,752 two weeks after Asakai until Burn Jita 2, when the number logged in hit a Retribution-worst 30,604.  Immediately following Burn Jita 2 was Fanfest.  The excitement generated by the event, combined with the upcoming changes for Odyssey, spurred the ACU to 36,252 for Odyssey's launch week.

If Retribution was a success, then Odyssey was anything but.  The ACU fell by 7.4% in the first week and never recovered.  The numbers for the first week were probably inflated as the CFC immediately invaded TEST-owned Fountain for the newly rebalanced moons in the region.  Another factor that could have contributed to the decline was the introduction of dual character training.  

For the first ten years of EVE's history, players could only train one character on an account.  Players who wanted to develop multiple characters needed to obtain multiple accounts.  With dual character training, a player could pay with PLEX (the equivalent of one month's subscription fee) to train two characters.  Of course, that meant that some pilots would consolidate accounts, thus reducing the number of subscriptions.  How many did that is only known to CCP, but the ACU numbers over a week have never come within 2,000 users of the level recorded in the first week of Odyssey's launch. 

One known fact is that in the wake of the destruction of a big null sec war, the ACU goes down.1  That held true with the Fountain War as well.  Despite the ACU falling 12.7% from the first week of the war until the final major battle in 6-VDT-H, the ACU would continue to fall after TEST announced its intentions of deploying back to low sec.  From the first week of the war until TEST lost its last system on 19 September, the ACU dropped 24.2%, from 36,252 down to 27,463.







One interesting coincidence I was unaware of until doing the research for this post was that from the time TEST announced its pullout to Aridia until the CFC seized TEST's last system and the Goons made arrangements with Chribba to act as a trusted 3rd party in establishing rental agreements in their newly formed rental alliance, that the ISK sellers I monitor suffered from an ISK shortage that sent their prices sky high.  In the same week when TEST lost the last of its sovereignty, the median price charged by the ISK sellers was actually higher than if a player just purchased two PLEX from CCP and bought PLEX off the market in Jita.2

Late September was the low point for Odyssey, however.  Between increased Factional Warfare activity, the beginning of the Halloween War, and a live event that didn't go quite as planned, the ACU had again reached over 29,000.

Rubicon had a promising start, with the ACU increasing by 10.5% from the week before launch.  But like Odyssey, Rubicon was unable to hold onto its initial success.  Unlike Odyssey, though, Rubicon did not give back all of its gains in the initial month.  The participants of the Halloween War proved more resilient than TEST and the decisive fight of the war, the battle of B-R5RB, saw a weekly ACU of 33,400, or 1,000 more than recorded during Rubicon's launch week.

However, a loss as massive as the one suffered by N3/PL is impossible to just shrug off.  While fighting did continue, the battle effectively ended the war.  Although New Eden saw a surge of new players enter the game, the traditional falloff of players logging in began.  From the Rubicon peak weekly ACU of 34,201 the week after B-R, the ACU bottomed out at 25,669 the week the Goons held Burn Jita 3.

War weariness wasn't the only cause for the decline in ACU.  During this same time, the price of PLEX, at least in terms of in-game currency, went through the roof.  From a price of 619 million ISK on 31 January, the average price of a PLEX sold in Jita rose 19.5%, to 739.9 million ISK six weeks later.  One of the conditions that will draw CCP intervention into the PLEX market is if the rate of increase is so high that it affects the ability of players to pay for PLEX to keep playing EVE.  At Fanfest, CCP revealed that the EVE Central Bank stepped in, probably on 11-12 April, to stabilize prices.

Rubicon was the last of the old-style expansions that CCP plans to release.  Kronos was the first of the planned content patches that CCP revealed it plans on releasing every six weeks.  The results so far, if the ACU numbers are any judge, are mixed.  While the ACU did go up 3.7%, the number of accounts actually logged in did not reach the level achieved during Fanfest four weeks earlier.  Worse, the numbers keep declining each week, but as the World Cup is currently running, that is possibly a statistical outlier that happens every four years.

That is a brief history of EVE over the past 18 months or so, trying to match events to the number of users logged onto Tranquility.  I'm sure that others will look at the annotated graph and come to different conclusions.  But the graph is interesting all the same.




NOTES:

1.  Although the drop after the fall of the Band of Brothers was caused by the Unholy Rage anti-RMT operation.

2.  I almost expect Riverini to stop by and say, "I told you so."

25 comments:

  1. I like (or at least don't object to) basically everything CCP has done in the last 18 months or so, but the fact is that the fix for a stagnant dull 0.0 has simply been delayed too long for me. I along with others repeatedly warned CCP that the patience of the average ordinary nullsec dude (ie: the ones who don't have personal R64s and multiple supers) was not infinite.

    And so it is proving to be and so it has proved with me. My subs are due for renewal and rather than give the money to CCP, I'll be getting an Elite:Dangerous beta sub and I'll play that for a few months.

    If and when CCP's roadmap, the one announced this May, comes to pass and 0.0 is revitalised as a consequence, then I'll certainly be back, and I'll be very happy to be so. That roadmap has CCP publically committing to focusing on 3 of the 4 greatest issues that are crippling the game right now (the 4th being power projection). It's great and all, just... too late.

    But even though CCP are doing the right thing at last... I'm not up for paying £100 per 6 months to log in for a once-a-week POS repair op.

    So until the next *real* war... au revoir.

    ReplyDelete
    Replies
    1. says the guy that wont leave lowsec because he doesn't like risking his own stuff

      Delete
  2. Interesting to see that ISK resllers had an ISK shortage at the same time that null sec ISK generation was at a low. I am sure there is no relevant connection whatsoever.

    As for the graph, beautiful work. I would love to see this revisited in about 3-4 months when Crius' full impact is being felt, as well as the in-the-loop Eve readers will have a real good idea of the upcoming high sec invention nerf.

    regards, Dinsdale

    ReplyDelete
  3. Will income from high sec / low sec production with industry alts allow players to PLEX multiple accounts?

    Probably not, if the number of industy alts stays about the same.

    Which means that the number of industry alts in HS/LS will probably decrease.

    Will they go to 0.0, as CCP hopes, or will they just go?

    ReplyDelete
  4. Put another way: EVE logged-in player count is currently sitting exactly equal to the lowest nadir of the aftermath of the summer of rage. Put yet another way: to find another time period when so few people were logged into EVE Online, you have to look all the way back to late 2008, when the game was on its way up.

    ReplyDelete
    Replies
    1. No, you have to look back to 2011. Not that I was around then, but just looking at Chribba's graphs.

      Delete
    2. @Hespire - 2011 is the dip caused by the Summer of Rage.

      Delete
    3. I have asked it many times, and no one answers. How has the game changed in the past 12-24 months? What paradigm has changed that would explain the drop in PCU and subs?

      I have said it is 2 things, that are related.
      1. The ongoing assault on high sec, particularly via the null sec cartel dominated CSM dictating changes that wreck high sec.
      2. The total gridlock that null sec has morphed into. Stagnation breeds boredom. Boredom leads to dropped subs.

      I have yet to hear any other reasons.

      regards, Dinsdale

      Delete
    4. To answer Dins - nothing has changed in the past 12-24 months. And, that is exactly the problem:

      1) CCP has done nothing to introduce the eye-popping, attention-grabbing new content which attracts new players. The game is old, and getting older. I think I have an Atari 2600 around here someplace, but I seriously doubt that I could interest my nephew in playing the games on it.

      2) CCP has done nothing to prevent excessive and increasing victimization of the many by the few. This includes most of the high-sec nonsense: wardecs, ganking , scams, Burn Jita, etc. One ganker can easily drive 5-10 players to quit the game each week - do the math.

      Delete
    5. No need for the "lowest nadir" expression since the term nadir is already defined as the lowest point of something (or direct opposite of another observable point).

      Delete
  5. This comment has been removed by the author.

    ReplyDelete
  6. If you look at the Last Year graph on EVE-Offline, there is a very recent 20% burst in activity. Both in EVE and DUST. This seems very odd. What happened in the last few days in the EVE Universe to cause this bump in both DUST and EVE all of a sudden? No bump that large has ever occurred in EVE without some major factor (war, expansion) attributing to it. At the moment there is nothing newsworthy that could attribute for it.

    ReplyDelete
    Replies
    1. This comment has been removed by the author.

      Delete
    2. I think you're misreading the graphs. The data points on the different graphs represent averages over different numbers of samples, and if you look at the 1-month graph you can see that the peak weekend numbers have been around 40K for the last month. And it's up around 10% from last month, not 20%

      Delete
    3. Such spikes are a consequence of the algorithm that Chribba uses to generate the graphs for past three months, six months, and year. They do not reflect reality. Check back in a couple of days and those spikes will be gone.

      Delete
    4. I generally just look at the 24 hours, past week, and all time weekly average graphs.

      Delete
    5. Jester is correct - the spikes are entirely an artifact of Chribba's bad math. Much like the in-game market history graphis, Chribba plays games with the raw data. Unfortunately, this sort of nonsense tends to produce the graphs you expect or want to see, and filter out what is really happening.

      Delete
  7. BNI is the largest corporation in the game. BRAVE is the third largest alliance in the game, and HERO is the third largest Coalition in the game. Not one mention of these entities.

    ReplyDelete
    Replies
    1. Did they do something of note that would have influenced the ACU numbers? Because I didn't see anything.

      Delete
    2. No one give 2 fingers about BNI. You guys are only important to yourselves. We tried to tell you dip shits on Podside but ya wouldn't listen. Welcome to null sec

      Delete
  8. From each of the dips, isn't a lower ACU typically indicative of more action and content?

    At each of the downslopes, we've got wars or conflicts.

    To my naive eye, it seems that ACU count may currently be low, but a high ACU count may be indicative of even bigger problems (that of absolutely no conflict, like before the Fountain War began).

    Heck, couldn't this easily be attributed to a wide-spread shedding of alts because plex has hit 800M each?

    ReplyDelete
    Replies
    1. The lower ACU correlating with big wars in nullsec is because you don't understand how nullsec wars are fought. The way to win with them isn't to kill ships, but rather to kill accounts. Many of the major sovnull entities have absolutely enormous endowments as a result of the value of their space and careful investment when not paying SRP. If the money runs out, like it did for TEST, that does, admittedly, impact the ability to defend, but, short of that, the way to win a sovnull war is to break their morale. In the meantime, the ships will just keep coming.

      Delete
    2. You are absolutely correct. Indeed, I made this exact point to David Reid and to Greyscale, Fozze & Rise: the current meta means that it's nearly always impractical to "starve out" your opponent unless they're the bunch of astonishingly incompetent idiots that the TEST HC was ("Hey let's start off our blocwar by disbanding our bloc!"), so as you very succintly put it, you target the accounts.

      In short, nullsec warfare cannibalises the nullsec playerbase because its so trivial to resupply as long as you have the ISK (and as long as your bloc members are willing to log in, they can easily make ISK) that all the normal verities of RL warfare - interdicting supply points, seizing vital resources, etc - are irrelevent.

      So it's all very well viewing B-R type events as tremendous advertising, but they come maybe once per year at the expense of corroding your long term player base for the other 364 days.

      In CCP jargon, the null meta is great for Acquisition, but it's catastrophic for Retention.

      Incidentally, I don't want to give the impression that the state of null is the whole story here; soaring PLEX prices have obviously caused a huge cull of alts, and they have also put the squeeze on RMTers. And CCP have been persecuting botters really hard. Many bots were "casual" botters who basically used the bot to buy PLEX to sustain 2-3 accounts. Now that they can't do this and now that PLEX are 800M a pop, well... numbers go down from that too.

      Delete
  9. DCT and MCT affect the number of subs, but don't affect the ACU. Most of the players who were training up toons on multiple accounts were not multiboxing.

    ReplyDelete
    Replies
    1. Who knows? So sure of your generalization? For me, the simple fact that I had running accounts for character training, made me take advantage of them.

      Even such simple things like using one to keep chat open, while logging around doing PI in the wormhole. Tada, +1 concurrent user.

      Delete