« August 2005 | Main | October 2005 »
September 30, 2005
Guild Recruitment Location
Server Sample: RP (High), PvE (Medium), PvE (High), PvP (High), PvP (High)
Sampling Period: 8/07/2005 12:00 am - 8/14/2005 12:00 am
Sampling Resolution: ~12 minutes
Parsing Method: The sample unit is each unique character. Each character was tracked across the server logs. Total playing time, lowest observed level, highest observed level, guild affiliation, and zones seen in were parsed.
Data Filter: None
Sample Size: 148,846 characters
We had competing hypotheses for where guild recruitment most often occurs. On the one hand, the amount of guild spamming in main cities made them a likely location. They are also centralized locations to begin with that would seem to encourage guild recruitment functions. On the other hand, characters spend more time in outdoor zones than in main city zones and grouping might encourage guild recruitment as well
To explore this question, we tabulated the location where characters joined a guild:
Guild Recruitment Location: If a character is not in a guild in snapshot X and is observed to be in a guild in snapshot X+1, then the character location at X+1 is noted down as the guild recruitment location and categorized as either being in 1) an instance, 2) a main city, 3) an outdoor zone, or 4) a PvP area (i.e., WSG or AV)
The resulting data let us look at 2 different things. First of all, it shows the distribution of guild recruitment events. Most guild recruitment occurs before level 10, drops sharply, and then spikes at level 60.
But the data also allows us to address our initial question. Across all levels, guild recruitment is most likely to occur in an outdoor zone followed by the main city zones. Interestingly, guild recruitment is more likely to occur in a main city zone as level increases. For example, city zone invites are almost comparable to outdoor zone invites at level 60, whereas outdoor zone invites are about 6 times higher than city zone invites at level 6.
Posted by nickyee at 09:12 AM | Comments (1) | TrackBack
September 27, 2005
Switching Guilds
Server Sample: RP (High), PvE (Medium), PvE (High), PvP (High), PvP (High)
Sampling Period: 8/07/2005 12:00 am - 8/14/2005 12:00 am
Sampling Resolution: ~12 minutes
Parsing Method: The sample unit is each unique character. Each character was tracked across the server logs. Total playing time, lowest observed level, highest observed level, guild affiliation, and zones seen in were parsed.
Data Filter: None
Sample Size: 148,846 characters
We wanted to get a sense of the pattern of migration from guilds to one another. Also, we were interested in how often people left guilds and whether this changed over the level spread. For each character, we calculated the following variables:
1) Unguild Event - for each time a character is observed in a guild in snapshot X but not observed to be in a guild in snapshot X+1, we increment their unguild event score by 1.
2) Guild Switch Event - for each time a character is observed in a guild in snapshot X and then observed in a different guild in snapshot X+1, we increment their guild switch event by 1.
3) Guilded - whether a character is guilded or not at the end of the sampling period. In other words, whether a character is guilded or not the last time we see them in a snapshot.
The results challenged some of our gut intuitions. For example, we would have guessed that guild switch events would be more common than unguild events. This is because with a sampling window of 12 minutes, almost all characters who had the intention of switching guilds would probably have done so during that window. We found the opposite, unguild events were far more frequent than guild switch events and this effect magnified over the level spread. Between levels 21-40, unguild events are 3 times more frequent than guild switch events (4% vs. 13%), and between levels 41-60, unguild events are 7 times more likely than guild switch events (3% vs. 21%). In other words, when characters leave a guild, it seems that they usually do not have a target guild they are intending to switch to. This leads us to believe that many characters become so frustrated or unsatisfied with a guild that they would rather leave and be alone.
The other interesting finding was seeing the familiar spikes. This time, they occur just before every 10th level, most noticeably the 39th and 49th. It's not clear to us though what it is about these levels (apart from their being close to big skill levels) that is causing the level of discontent with one's guild. Any guesses?
Posted by nickyee at 04:03 PM | Comments (10) | TrackBack
September 17, 2005
Fastest Leveling Times
Server Sample: RP (High), PvE (Medium), PvE (High), PvP (High), PvP (High)
Sampling Period: 8/07/2005 12:00 am - 8/14/2005 12:00 am
Sampling Resolution: ~12 minutes
Parsing Method: The sample unit is leveling event. We tabulate the time between a character's level and when we observe them at a new level. Only a player's online time is counted. We exclude the first leveling event from every character because it doesn't constitute the total amount of time to make that level.
Data Filter: None
Sample Size: 149,240 leveling events
To get a sense of the fastest leveling times, we calculated the top 1st percentile, 5th percentile, and 10th percentile of times it takes to get from one level to the next. We summed these times to get a lower bound on the time it takes to get to level 60. This is a lower bound because we're not tracking any actual characters from level 1 to 60 but the fastest times it took any character to get from one level to the next. Also, significantly lowered leveling times may be due to turning in quests where the work was done from the previous level.
Here is the plot for the time it took the fastest characters to reach the next level.
The graphs below show the time to level by class and top percentiles. Interestingly, several classes that overall have long leveling times had much lower leveling times in the top percentiles (i.e., Druids and Hunters).
Levels 1-20
Levels 21-40
Levels 41-60
The table below shows the estimated time to reach level 60 by class and top percentiles based on the data above. The lower bound produced by the top 1st percentile times hovered between 4 and 5 days. For the 5th percentile times, the lower bound was between 6 and 7 days. And finally, the lower bound was between 7 and 8 days for the top 10th percentile.
Finally, we chart the estimated time it takes for characters of different classes to reach level 60 by the top percentiles.
Posted by nickyee at 11:24 AM | Comments (6) | TrackBack
September 06, 2005
Distribution of Leveling Times
Server Sample: RP (High), PvE (Medium), PvE (High), PvP (High), PvP (High)
Sampling Period: 8/07/2005 12:00 am - 8/14/2005 12:00 am
Sampling Resolution: ~12 minutes
Parsing Method: The sample unit is leveling event. We tabulate the time between a character's level and when we observe them at a new level. Only a player's online time is counted. We exclude the first leveling event from every character because it doesn't constitute the total amount of time to make that level.
Data Filter: None
Sample Size: 149,240 leveling events
To get a better sense of the distribution of leveling times - whether most people level fairly quickly and the overall mean is skewed by laggers - we calculated the normalized score for every leveling event based on the character level. We then plotted all the normalized scores.
Normalized Score = Z Score = (X - Mean) / Standard Deviation
The plot shows that most people actually do beat the overall mean (a z-score of 0). While the average time it takes to get from level 1 to 60 is now 15.3 days, the median is 13.9.
Posted by nickyee at 11:22 AM | Comments (3) | TrackBack
September 01, 2005
Dances With Farmers
With my 57 mage nearing the end game in WoW, I recently had my first two encounters with farmers in the wild. Of course, I've had the typical encounters in Ironforge where someone is spamming the trade channel or yell channel, or the situation where you contact someone about an item they've advertised and the subsequent discourse is limited to "Farmer: [Lag Eliminator of the Auctioneer] 29g. okay?".
Now, by "farmer" I'm meaning the more restricted definition of someone who habitually uses RMT to monetize the things they do in-game. Which means I'm making assumptions about the two farmers I met, since IGE is not letting me look into their internal records. As players, it's very easy to make assumptions too quickly. My own trading alt, a 21 rogue, is listed by another player on the server forum as a "known farmer".
I stumbled across Nepili while questing in the Blasted Lands, needing to free some tormented soul at the back of a cave filled with crazed followers of Allistarj. Nepili was going from mob to mob outside the cave, and I started to -- very slowly -- make forward progress into the cave. Did I mention Nepili is a 60 rogue? (Did I need to? Ah, assumptions.) At one point she simply stealthed past me as I made my slow progress, and I would come on short stretches of cave filled with dead bodies. I died a few times, and after running back from the cemetary would find Nepili back outside again, or back inside. At one point inside the cave, Nepili pulled a mob, ran toward me, and then vanished, at which point the mob turned on me, while I was already at low health. Intentional? I don't know.
And then, about five minutes later, Nepili sent me a group invite, and through broken English and animated emotes I conveyed that I was trying to accomplish something at the back of the cave, beyond all Allistarj's crazed follewers, which we quickly dispatched working together. Paranoid that I am, I wonder if she tried getting rid of me by training, and then decided the fastest way to get me out of her hair was to help me.
Copfe is a different story. My trading alt met Copfe in Ironforge, trying to buy Arcanite Crystal, and when she couldn't speak very good English, I labeled her as a farmer. The chat started to run a little on the long side. She (and I use "she" because the avatar was female) said she was from Xi an, and asked how something was said in English, which I explained, and by this time I was curious enough about how farming looked from the inside that I tried to keep the relationship going. We added each other to our friends list, and exchanged simple greetings each day, and each day I'd try to say at least one thing that would stretch her English. Over the next couple weeks, each time one of my alts logged on, I noticed she was either in the Eastern Plaguelands or Azshara.
Eventually, the quest my mage was working on took her from the Blasted Lands to Azshara, and Copfe happened to be in that zone when I showed up. My mage could solo the first couple quests, but then died to a hydra I needed. And then I got a tell from Copfe,
[Copfe] whispers: omg
To [Copfe]: what's up?
[Copfe] whispers: you die
To [Copfe]: lol. yes. how did you know?
[Copfe] whispers: i see you
Well, it turns out she was swimming around in the ocean near where I was questing, and sure enough, happened to witness my demise. Curious, I ask what she's doing, and she says she's mining. Taking my opportunity, I ask if I can watch, and she says that's just fine. So I spend the next hour as Copfe runs an incredibly efficient loop around Azshara, mining veins and picking up chests, eluding large numbers of mobs, but going for the jugular for certain mobs (who are apparently too near chests and veins for stealth to work). During the hour that I followed her, I didn't see any good loot or even arcane crystals drop. The largest value would have been in the thorium and mithril mined. But the efficiency with which she worked was mind-boggling.
So what impact might Nepili and Copfe be having on players in the field, apart from any economic impact? For Nepili, very little. She's killing mobs with a very high respawn rate, which is probably part of the reason the site works for her. If she intentionally trained a mob on me, well, that's some degree of impact. The largest impact she's likely to have is that questers may have to slash through somewhat fewer zerglings on their way into the cave. Copfe's a different matter. I don't know the respawn times of the various chests and veins she looted. If they're a significant part of an hour, she could single-handedly be having a sizable impact on the loot available in Azshara to the "more casual" player.
Posted by eric at 10:05 AM | Comments (17) | TrackBack