by Jeff Woleslagle on Jul 20, 2006
Since World of Warcraft owes much of it's success to it's friendliness to solo players, it comes as no surprise that WoW players overwhelmingly favor solo-friendly classes over "support" classes, whose abilities are structured to benefit other classes especially in the endgame (like druids, paladins, and priests). While many players believe that Rogue is the clear favorite for much of the player community, owing to its stealth, survivability, and high DPS, the WarcraftRealms stats show that the Hunter class edges out Rogues for the most popular class, except by a meager 1% margin on PvP servers. Warriors beat out Rogues as well on PvE and RP servers. Since Shamans and Paladins are faction-specific classes, and the stats are heavily skewed toward Alliance (see above), we have to go to faction-specific numbers to see that among the Horde, Shamans rate with Rogues in terms of popularity. Paladins, however, are tied with Priests as the third least popular Alliance class.
Table 2 - Level Distribution Data
Table 3 - Level Distribution Statistics
No World of Warcraft census information lends itself better to statistical analysis than level distribution data. It's not just an academic exercise; if population is spread fairly evenly over all levels, players have a better chance to find players to group with, socialize with, join battlegrounds with or against, or pick a fair PvP fight with. In most new MMORPGs, the distribution looks something like a bell curve cut in half at the peak as players begin at level 1. Statisticians call this a "high normal" or "positively skewed" curve. The bulk of the population moves through the game like an ebbing wave, ultimately piling up at the level cap, and this is a "low normal" or "negatively skewed" curve.
In analyzing the level distribution of active players, we arrive at two important indicators of population distribution: skewness and kurtosis. The mathematics is beyond me, but I understand kurtosis to mean the even-ness of distribution (e.g. are there a similar number of active level 15s players as level 31 players? as level 52 players?) and skewness as an expression of the "weight" of the distribution at either end of the spectrum (e.g. are there more players at level 10-19 than at 50-59?).
When people talk about World of Warcraft's success, they cite accessibility, availability of quests, directed play, and other qualitative mechanics directly attributable to… well, fun. But one of the best kept secrets of World of Warcraft's success can be expressed quantitatively, Numb3rs-style, through math. Check out the descriptive statististics to the right. K values from zero to two indicate meso-kurtosis, or (literally) a flat level distribution, and low positive skewness indicates a gentle downward slope in population as levels increase - both signs of as healthy a MMORPG population that can be hoped for. Omitting level 60 - which I'm comfortable doing since, by most accounts, the level 60 game is an entirely different game than the 59 levels preceding it - World of Warcraft meets these conditions more than a year and a half after its inception. This mathematical fact makes WoW a rarity among MMOs, most of which coping with vacant low-to-mid level zones, snaring these games in a developmental tailspin of catering exclusively to players heavily concentrated at high levels.
So, long story short, your active World of Warcraft character is statistically likely to be a level 40 Night Elf or Human Hunter on a PvP realm. There's nothing wrong with wanting or having a humanoid hunter, so long as you're having fun. But if you really want to go against the grain, try rolling a troll priest on an RP realm. That might just be my next character!
Return to Page One...Questions, comments, or suggestions?
Please email me or post in the
World of Warcraft @ Ten Ton Hammer forums!