Wednesday, April 7, 2010

Statistics Galore(y)!

Ravious is talking about leaderboards, data collection and elitism today, so I thought I’d throw my two cents in. I think data collection is a key to long term survival for an MMO, and its fortunate byproduct should be player accessible stats, both about themselves and the community at large. Its been argued before that players don’t care about the MMO’s fictional story as much as they care about their own stories and memories from the game. Player stat devices like the Tome of Knowledge and even WoW Achievements reinforce the players’ own history, reminding them of what they have achieved.

Unfortunately even a system as mindless as WoW Achievements can be perverted by players’ darker sides, as seen in the current trend of only allowing players to join a PUG for a dungeon if they can show the achievement for the dungeon.

As long as stats have been recorded, they’ve been used for comparison, its just natural human behavior. Its also human nature to pursue what’s best for yourself, and when these two instincts are combined its only natural for people to exclude based on statistical thresholds. It becomes a problem when there is no way for people to improve their important statistics. In this case it means that players who are just reaching level cap cannot improve their gearscore or complete dungeon achievements since no one will group with them. Ravious (and Sanya) mention a solution to the problem, tracking everything. I agree, and offer two more solutions.

First, don’t just report all stats, reward all stats! Johnny Newbie may not be able to kill the UberDragon within his first week of play, but if he manages to fall off cliffs and get himself killed more than any other player, he deserves some recognition. Give him the weekly title of “Johnny the Fragile, Chief Inspector of Gravity” and watch him be proud to show off his title. Even if he’s showing his title off to other new players with equally lame titles, they’ll be happy because they are unique, for that week they are the absolute best (or worst) at something. With any statistic there can be only one leader, so if no other statistics are shown (or are shown but without possibility of recognition or incentive) they are naturally going to gravitate towards the one or two stats they are told matter.

Second, don’t show statistics that can be easily improved simply by playing the game a long time. Total number of times a player has killed a dungeon boss is a number that should really be more depressing the higher it is, because its just a neon sign screaming LACK OF CONTENT, but instead it would be (and is) used to gauge how worthy a player is. Why not instead track a players success rate at a dungeon, or fastest time to clear a boss, or largest damage healed/dealt/tanked in a single boss fight? These still measure the success and relative skill of a player, but do it in a way that at least limits the impact of longer time /played. And if you absolutely must show quantity stats, show them over a standardized period, like the week or month. That way players who suffer from EPeen Redundant Incessant Comparison Ketosis (EPRICK) will still have something to show for it.

I’ll throw in a third solution just for good measure. Role Proficiency Ratings! See my previous post for details, but it addresses these same types of problems of socialization gated by arbitrary statistics.

1 comment:

  1. Puzzle Pirate "standings" are a measurement of player skill as compared to other players who have been doing the same activities, and your standing goes "offline" if you don't play for a while. Those players with "Ultimate" standing in any given puzzle are therefore fairly reliably above competent. Players still use those standings to filter applicants (especially in Gunning), but since they are a measure of skill as compared to the rest of the playerbase, they have significantly more meaning and relevancy than some set of measurements based primarily on time (leveling, grinding through dungeons repeatedly, whatever).

    Standings and puzzle performance are also somewhat fuzzy measurements. Almost all of the numbers of the system are hidden behind the standings (just labels), which does have a tendency to reduce the social friction. The increased granularity of a pure numerical system makes comparison overcompetitive, but when your *performance* is measured in fuzzy categories (Booched, Poor, Fine, Good, Excellent, Incredible), you can gloss over the details and just get a sense of the general usefulness of a player's performance.