Digging Into Xfire Numbers

In the interests of delving into a point made by Openedge in the comments for the last post, let me say firstly that I am deeply reluctant to take Xfire statistics at face value. They only reflect the activity of Xfire users, for one thing, and not even all of those – like myself, sometimes, when I forget to turn Xfire on before playing (I don’t have it set to autostart when I load Windows, since I’m running Vista and have never bothered to work around the OS stopping it from running on startup.)

Also, the pool of Xfire users leans most heavily in the direction of strategy (RTS, mostly) and shooter players, or so it seems to me. For example, it comes bundled with Supreme Commander, which I don’t think has at all the same kind of audience as a typical MMO. Bread-and-butter FPSes like the Call of Duty series and CounterStrike peform very well there.

Nevertheless, for the sake of furthering the discussion, let’s assume that Xfire’s statistcal sample is a represenative one, and where it does not reflect the totality of activity in a given game, it subsumes a reasonable sample of said activity. So where we cannot extract hard numbers from the Xfire data, and shouldn’t try to calculate anything very far removed from them, regardless of how reliable these numbers are, they are unquestionably the best publicly available data, and one can probably take the order in which games rank on their list as fairly solid. So let’s look at what Xfire says about 16 select subscription titles today. I’m including all of the SOE StationPass titles because of the origin of the discussion, but I’m excluding WoW as an obviously aberrant example, Guild Wars and Runes of Magic for not being subscription titles, and Lineage II for having much of its player base in Asia, where people probably play by the hour rather than paying a flat monthly fee.

Title Xfire Rank Minutes Played Users/Day Minutes/User
EVE Online 14 1,084,441 4,546 239
The Lord of the Rings Online 18 823,793 2,996 275
Warhammer Online 26 656,376 2,426 271
Age of Conan 37 474,280 1,695 280
CoH/CoV (Combined) ~84.5 (Est.) 181,877 3,031 244
Star Wars Galaxies 87 172,048 559 308
EverQuest II 99 151,944 490 310
Darkfall 127 94,130 212 444
Dark Age of Camelot 168 54,657 182 300
EverQuest 199 41,854 92 455
D&D Online 203 39,956 118 339
Vanguard 239 30,219 138 219
Pirates of the Burning Sea 282 20,823 70 297
Planetside 319 14,612 72 203
The Matrix Online 426 6,330 48 132

I have made one simple calculation from Xfire’s raw data – the average number of minutes eacxh Xfire users spends in-game. Looking at all this, we can see where the doomed Matrix Online falls – dead last, and I made no effort to cut the list off at that point. (I cut it off where I judged the list of ‘proper’ western subscription MMOs to end.) Furthermore, in terms of the number of people logging in per day, and in the average number of minutes played per user, Matrix Online also comes dead last. Probably the ‘users per day’ figure most accurately reflects subscription levels, even though Xfire doesn’t rank games that way. Only 48 Xfire users were logging on to Matrix each day – I’ve run D&D games that big (or at least helped run games that big.)

The First Tier: The Strong
This figure (minutes played, actually) determines the overall Xfire ranking of a game. We can see here that EVE Online is performing very strongly, a very solid 31% higher than the lext-highest totle, LotRO, which in turn pulls 25% higher than the third-ranked MMO, Warhammer Online. Number four is, sursprisingly, Age of Conan.

Now, three of these titles (WAR being the exception) have run promotions on Xfire, which has almost certainly helped increase the percentage of active players who use the Xfire service. I think this is probably not an incredibly significant increase, and thus while it may have increased those games’ overall ranking, it probably hasn’t affected the placement among the captured titles (subscription, non-Asian MMOs.) All four of these titles have in the neighborhood of 500K minutes played in the designated period, or higher.

Age of Conan’s numbers look very solid overall – not as spectacular as one might hope, surely, but extremely solid – falling in the top third of the pack in overall ranking (hours/day) and users/day, with a somewhat lower value for hours/user – but I judge this last to be the least important statistic. I judge AoC and the rest of this bunch to be “Safe for the duration, barring an unlikely and catastrophic collapse.”

The Second Tier: The Solid
The next block of games falls into the range of 100K-200K minutes played, and includes (in order) City of Heroes and Villains, Star Wars Galaxies, EverQuest II and Darkfall. I combine the (strong) numbers for City of Heroes with the (weak) numbers from City of Villains. Most players of the CoX titles are probably using CoH as their portal, regardless of what faction they play, and Paragon Studios now considers the two to be one single product. Flatly combining the numbers is a questionable statistical practice, but I don’t think I haver enough data to feather the CoV numbers in properly. I calculate that the consolidated title would fall between Xfire ranks 84 and 85, so CoV-exclusive play is enough to kick the aggregate up a full ten places.

The surprise here is Star Wars Galaxies. It comes in 13% higher than EQ2, which I find shocking, given that it seems to be held in general disregard post-NGE and is surounded by negligible buzz. EQ2 performs strongly, I think. All three games pull in the neighborhood of 500 users per day or better (the top 4 titles studied pull over 1,500 per day.)

I am both surpised (in a good way) and pleased at where EQ2 falls. It’s numbers look nice and solid. Not as high as I feel it deserves, but it’s also several years old and while it gets reliable talk, it lacks buzz and it perhaps an unexciting title at this point, depsite being a very good game.

Darkfall’s performance is also pretty good, consiering that it’s still not officially launched in North America (although a lot of NA players are playing it anyway,) and that it requires significant effort just to buy the thing. Also, its “hardcoritude” number, the minutes each user put in on average, is really high, second overall next to EQ.

The Third Tier: The Weak But Safe
We come next to a cluster of titles between about 20K-50K minutes played, and 100-200 users per day: Dark Age of Camelot, EverQuest, D&D Online and Vanguard. I’d call these titles “Safe unless they slip significantly.” Two of the oldest extant titles in the hobby fall here: DAoC and EQ. The latter’s numbers are the most interesting: far below EQ2, whereas I’d always assumed (based mostly on the MMOGChart data, which I am now willing to discard entirely as both out-of-date and generally innaccurate,) EQ to retain some number of subscribers roughly in parity with EQ2. From this data, that’s pretty clearly not the case, even if we assume that Xfire skews the number very low: EQ brings in less than one-third the minutes played of EQ2, and less than one-fifth the users per day. That’s very weak. But on the other hand, in terms of minutes players per user, EverQuest ranks first. But this should not be a surprise: EQ’s remaining population must be (it seems to me) mercliessly hardcore.

D&D Online falls about where I expected it, but probably higher than others might have guessed, and lower than it deserves.

Vanguard’s numbers are weak, but nowhere near as bad as Matrix Online’s are. Vanguard users clocked almost 5 times as many minutes as Matrix players, and close to five times as many users logged in per day. If we assume (a chancy assumption) that the same factor applies to subscriptions, and if we further assume Matrix has 15,000 subscriptions (which seems a reasonable guess to me, plus or minus 5K,) that puts Vanguard somewhere just north of 50K, which is lean but certainly viable, and about what I would have guessed. Vanguard looks weakest in the minutes per user statistic – but I can believe it given fairly unforgiving (compared to WoW or even EQ2) gameplay in a great-looking world, with people coming in to sightsee for a while, but not getting much play in. While I don’t think it’s in imminent danger of cancellation, it has little ground to lose – but I also judge that Vanguard has always (barring the immediate post-launch period when subscribers didn’t know what they were in for,) been in danger of going on life support. So I’m not sure that this data shows us anything new.

The Fourth Tier: Next Against the Wall
The next titles down, Pirates of the Burning Sea and Planetside, are pretty clearly in trouble. Pirates comes in well under Vanguard across the board, and worse in the users per day statistic, which is why I put it here instead of in the next highest category: It may be safe but it has no ground to give. But it, like Vanguard, can offer a promotion or content update and get a little spike, drawing numbers upward, if only briefly.

I doubt very much that this is possible for Planetside. If it were anywhere other than in the StationPass program, I’m guessing it would already be dead, and when The Agency Draws near, I think it will be, if not before.

Then, of course, we have Matrix Online. Dead last. The only reason people weren’t loudly predicting its cancellation is because nobody cares. Deservedly or not, nobody was playing it.

EDITED: to clean up some typos, add a link, make some clarifications and additional obsevations, and reorganize the whole thing. There’s some buzz about this one, and I want to be as clear as possible.

9 responses to “Digging Into Xfire Numbers

  1. Since I started tracking those Xfire numbers last year when WAR came out, I had noted about the same relevance for these players, except for EvE and LOTRO.
    They both ranked considerably lower. AOC also was in the 50’i.e: low.
    Then EvE was the first to advertise events and prizes on Xfire (out of that list…amazingly WoW HAS been doing that regularly and it shows), and then EvE skyrocketed on the numbers.
    LOTRO shortly after that…same thing. They jumped up into the top 10 even for a while. But, have settled into that cozy slot it is at.
    AoC did it, and made it into the top 20, and has finally found it;s place.
    If we look at the chart at this point, we can plainly use it to show which game probably has more scrips and popularity. I think it looks quite accurate.
    It is not perfect by far, but to get an inkling of how well EQ2 does vs say Lineage 2 (missing on your list, but ranks much higher) in player pops, I think it is useful.
    The best surprise on that list? Guild Wars which still ranks in the top 10 ALL THE TIME (Pay once, play often has it advantages). Having logged several times this week to look around again, city hubs are STILL full of people…what a great game.

    Now we need GW2..


  2. I left out Guild Wars becuase it’s not a subscription game and thus was less relevant to the numbers game I was playing. I left out Lineage II becuase I consider its numbers less well understood than those of other games, since its primary player base is in Asia and likely not subscription-based.

    I think we can trust the Xfire numbers to give us a general picture of what the popularity of different games is, relative to one another. I think it’d be very dangerous to can draw further quantified conclusions from those numbers. In other words, I expect that the order in which the various titles fall is likely correct except maybe where two games have very close numbers, but I wouldn’t want to try to calculate actual subscription figures from this data.

    I confess to being very curious to see what Guild Wars 2 looks like, and further suspect that it may be an even bigger blockbuster than GW. I don’t really consider GW an MMO, but it’s no less remarkable a title for that.

  3. I think that Xfire is a good sampling method for SOE station pass games. Most players that have an SOE game are probably aware of other SOE games, and are the most likely customers of them, so if SWG is beating EQ2… ouch. I’ve heard plenty of grumblings about player populations in EQ2 and every single player that I knew playing EQ2 is no longer doing so.

    Nice to see WAR up there with LotRO. Will be interesting to compare them after LotD for WAR launches. My guess is that the WAR user numbers won’t move much, but average minutes played will increase.

  4. Given that a lot of EVE is played AFK — when going on long trips, I just set the path and play another game or watch a DVD — play-styles may influence total minutes played for EVE in ways not true for other games.

  5. Play time per user is more linked to hardware demands of the game than anything else. For instance, tons of people never fully quit out of WoW because you can minimize it and not take a performance hit. DarkFall (which on Xfire I have ‘played’ over 900hr since launch) is left on all day for me, since I just leave my character at the murder wall and the game never afk-kicks you. LotRO on the other hand cripples your comp if you minimize it, so people generally quit out if they plan to do anything else.

    Still, the numbers are interesting, I just wish they community as a whole was not so massively dominated by FPS players.

  6. I’m afraid planetside may have some additional issues stemming from Vista. Last summer I was taking advantage of all the EQ/EQ2 promotions and decided to try out the demos for some of the other Pass games on a lark. I had the hardest time getting Planetside to run, and when I finally did, ran through the tutorial and tried out some of the weapons on the firing range.

    That night, I got home from work to discover my EQ/EQ2 accounts had been banned. I’ve had that EQ account since ’99!

    The CSR I spoke to confirmed that they were having a problem with Vista machines generating false hack alerts and unlocked my account. I immediately deleted the demo, and swore it’d be a cold day in Ro before I messed with that again.

  7. Pingback: West Karana » Nightly Blogroll — Though Amaryllis dance edition

  8. @Edge: Screw GamerDNA data. It may be as good or better than Xfie’s but it’s much more trouble to extract. Xfire is good enough for what I’m doing here.

    Also, I dont consier a 50ish ranking for AoC particularly low. AoC could drop 40 ranks and still be the #5 MMO (including WoW.)

    @Syncaine: I doubt very much that this is statistically significant, since many games do boot you after some period of inactivity, and along with the people who stay logged in while they’re at work, you have the people who log in for two or three minutes to check mail or the AH. If I were attempting to do real statistical analysis with better data, I might consider it significant, but for the kind of back-of-the-envelope analysis I’m doing, I’ll call it a wash.