A Deeper Glance at Some Quest Stats

by trace

As many of you might might be aware, I post lots of questionable statistics for community quests during their duration. They're pretty basic and straight to the point, for the most part, and basically serve to answer if we're going to succeed or fail, and nothing more than that.

I've always wondered if I could answer some more complicated questions with quest data. In that regard, I used this off-week to take a gander at some quest data I haven't had a chance to use.

First, a little background. Over the last few weeks, I've had the chance to give Daniel and Coonce my input on what kind of information would be useful for statistical purposes, both for the user's benefit and Whiskey's benefit. Finding out how quests are best enjoyed by the userbase makes it a whole lot easier to create better and more interesting quests in the future, after all. Anyways, Coonce threw together a nifty/freakin' awesome collection of data for community quests, and invited me to use it however I see fit. Said data can be found here, if you're so interested.

Right now, Coonce's data contains a very accurate graph of completions over time, along with the names of every user to have completed a quest set by order of completion. Given that I've already tracked most of the completions over time, I decided I'd focus on the user data, to see if there were any bits of information I could glean.

This is the point where my data collection went mad. You see, as of 2:00 AM EST on Friday, 14,646 users had completed one or more of the nineteen community quests that have been on the site. I put all this data on a single spreadsheet, which now happens to be part of an 8 MB Excel 2007 file. To put that in perspective, I used to work on month-end financials at a large company, pulling SQL data into an Excel workbook. Those financials were 2-4 MB large, and that was considered a risky size because it had the capability to crash several of the computers in the office, mine included.

So yeah. Eight megs is a nightmarish, crazy size for a workbook. Making large changes to this file can take thirty seconds to a minute on my computer, and I can hear the variable speed fans spin up to a fast pace during calculations. I regret nothing, but I'll freely admit I might be a bit mad.

Anyways! I haven't gotten any really interesting information out of this data yet, but here's a few preliminary charts to gander at based on what I've grabbed:

Completions by User - All Community Quests Here's an overarching pie chart showing the dedication of users to community quests. As you can see, about 75% of the 14,646 users who have completed a community quest haven't finished more than five community sets. There's also 101 crazies on this site (myself included, as you'll see later) who have completed every community quest.

(Considering that these stats are counting completions to date, well after many of the community challenges, I wondered if removing limited edition quests from the data would change anything. In short: No, not really.)

This alone isn't necessarily useful data, though. A few of the community quests fell flat very quickly, such as Stalker Status and Welcome Wagon. Even now, three months after its introduction, only 290 people have completed Stalker Status.

I decided that for further analysis, it'd be best to divide quests into the following categories:

Scavenger: Find something on the site, from a wiki page to someone's profile.
Video: Watch a video, be it specific or more general.
Participation: Contribute something to the site, from a comment to a bunch of quiz questions.

Obviously those are very general categories, but I think for purposes of comparing whole sets, they'll suffice. I then started comparing completions of quest sets based on their types.
Completions by User - Scavenger Quests Only Looking at quests with only scavenger components to them, we see that users are more apt to complete just one set before moving on with their lives, but there's a slightly larger group interested in completing as many of these as possible. It's hard to say for sure, considering I just removed ten quests from the picture and reclassified every pie wedge.

I also wondered if sets with both scavenger and video components had a different level of popularity with the community.
Completions by User - Scavenger & Video Quests Only In short, not really. The wedges in this chart are pretty comparable to the wedges for just scavenger quests, I'd say. Maybe there's a little more dedication to completing all the sets with video included, but with only four quest sets to examine, I'm sure as heck not going to jump to that conclusion.

Noting the lackadaisical portion of these piecharts, I wondered if users who had completed only one quest were more prevalent in the quest system's early days or right now.
Completions by User - First Seven Quests
Completions by User - Latest Seven Quests Seems like users are a bit more passionate about quest these days. Then again, 4,189 more users participated in the early days. I assume they've since moved on and made the newer stats a little more equal.

So, can I draw any really interesting insight from this data? Probably not, at least not from any of the charts above. I think I need to step back and look at the data later on, and see if I can come up with more specific questions to ask. I've got a few in mind already, but if anyone else has some suggestions, I'm more than willing to hear them.

I'll end this blog post with perhaps the most unsettling data I put together: The top 25 of the 101 people who completed every community quest.

Rank	User	Average Finish	Average Ranking
1	PsEG	140.47	6.2%
2	Video_Game_King	200.53	8.5%
3	Psycosis	214.74	11.2%
4	Milkman	263.53	10.5%
5	GoodAssLettuce	273.84	8.5%
6	KaosAngel	304.84	8.2%
7	Bozar	318.21	14.9%
8	arab_prince	332.05	11.5%
9	Hamz	347.05	11.6%
10	crono11	388.53	14.7%
11	groin	404.79	13.7%
12	Victor018	413.47	18.2%
13	mracoon	440.47	14.3%
14	onimonkii	447.84	19.3%
15	Meltbrain	452.00	16.7%
16	TurboMan	463.58	19.1%
17	Marino	471.58	13.5%
18	ZombiePie	500.16	18.2%
19	S1LENT_ASSASS1N	503.00	18.5%
20	xanavi	511.74	14.4%
21	BunkerBuster	517.84	17.7%
22	Atomasist	525.58	18.8%
23	MillerWakka	526.95	22.1%
24	Shini4444	527.74	15.5%
25	jadeskye	551.05	20.1%

Average finish is the average finish number, whereas average ranking is the top % on average a person finishes in. In my case, I finish in the top 6.2% on average.

Perhaps I'm a little too into these quests.