The Effect of Player Perception of Class Balance on Player Perception of Class Balance

Meters, Warcraft Logs & Statistics: Part 1

Player perceptions of class balance are based around a few different metrics: ranking & statistic pages from WarcraftLogs, stacked simulation results for all specs, or even in-game DPS meters. There are problems with all of these approaches, and these may be the reasons why players see under or over-powered specs in a different light to the dev team.

As implied in the topic this is effectively a positive or negative feedback loop on these metrics, and the largest of these is also the hardest to quantify: individual player skill. The following chart shows two randomly generated Gaussian Distribution datasets from

image (24)

The distributions are roughly the same, with Spec B having a 2.25% gain over Spec A on average, or a 3.87% gain for the top 5%.

image (23)

If we move every second data point in the top 100 for Spec A over to Spec B the chart gets skewed a little, but still looks fairly similar to the previous chart. The differences increase to 5.79% & 8.77% for average & top 5% average respectively. As you can probably guess, if the top end of players for one spec shift to another it will distort the overall dataset, so players look for statistics on Warcraft Logs and see that Spec A is nearly 9% better than Spec B. This will create an environment where more players (usually those at the higher end who want to min/max further) will make the same change in a negative feedback loop.

It’s not just the drive to min/max or skill that affects things like this. The population of players who would switch also tend to be in the higher levels of raiding (ie: halfway through Mythic Nighthold) which means that the fight times will be shorter, further increasing the DPS impact of bloodlust or burst DPS cooldowns.

As a more extreme example, this is what happens when the top 50% move.

image (25)

The midpoint of the bell curve has visibly moved towards the right, and Spec B has a much higher average (nearly 10% more) than Spec A, although the top 5% difference is only 7.66% as the only change is the spec, not the output value itself.

TL;DR Version

As more top end players move from one spec to another it distorts the charts on Warcraft Logs, which in turn causes more players to move. This is why it’s not always accurate to say “Buff A plz bc its bad on WCL”.

Update 12th May

After posting this on Reddit a lot of people have said that this is a simplistic approach, doesn’t take into account “lower skill” players shifting as well, and isn’t very realistic. These points are all correct, but they miss the point of this post: to demonstrate that the shifting of players from one spec to another has the potential to distort metrics displayed on Warcraft Logs & similar sites more than any actual differences between the two specs.

If you look at the number of parses per spec for “pure DPS” classes like Mages or Warlocks you can see a large shift from specs that were popular for Emerald Nightmare to specs that are now popular in Nighthold. A quick comparison of the final bosses of both instances on Heroic shows that Mages shifted from about 69% Fire to 61% Frost (see this comment for details) with most of the shift directly from Fire to Frost (Arcane had 13% & 10% respectively).


About binkenstein
I'm a geek, living in Christchurch, New Zealand.

One Response to The Effect of Player Perception of Class Balance on Player Perception of Class Balance

  1. Pingback: Simulations, Logs & Dummies | Blogenstein

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: