It's been nearly a month since I last did some graphs and data. This week I've decided to use data to look at the impact of 1.121 on realm and class performance.
To do this I've taken three weeks of data post-patch (25 June - 14 August) and compared it to three weeks of consecutive data pre-patch (21 June - 10 July). Results are below, but first, some key points to consider before you draw assumptions.
There is a lot of data here - there are lots of assumptions you can make, but please don't jump to huge conclusions based on one data point
- The data I use is from @Vanesyra website and is available on a week by week basis. Patch 1.121 went live on 26 June, which means one week of the post-patch data includes 1 day of old pre-patch 'meta' (i.e., data from Monday 25 June). The servers were also down for an extended time on the 26.
- Data from the week 11-18 July was broken - the Herald stopped capturing data. I've omitted this week from the data.
- To get three consecutive weeks of pre-patch data, I've used three weeks from 21 June - 10 July.
- I've used various different measures (total rps, percentage share of total rps, percentage point change, and percentage change) - some of these different measures are appropriate for different figures. I've explained where relevant.
- This data is very limited - there are lots of nuances in just three weeks of data and it is very susceptible to outliers (see necromancers, for example). A more ideal study would look over a longer period of time, for example.
- This data is not definitive. It should be combined with played experiences of players, etc., and basic understanding of the game. Because the data sample is quite small in terms of time and number of players, other influences - for example, changes in zerg leaders, temporary breaks from key players, etc., can have a big impact on data.
- The patch is very new. Many of the effects of the changes will take months to be realised. The data below points to an initial reaction, rather than any long-term trend.
Key findings
- Midgard, already dominating RP totals prior to the patch, has extended their lead significantly after 1.121. The realm balance in recent weeks has been abysmal.
- Albion has seen little change overall, albeit is performing slightly better than Hibernia since 1.121 for the first time since I began collecting data.
- Hibernia has plummeted in RP totals since 1.121.
- Stealth classes have done extremely well since 1.121. Assassins are now the highest earning RP class across realms by a significant margin.
Graphs and charts
Population (number of characters) over time
Here are the basic graphs I include in all my data showing the number of characters over time. This data includes level 50 characters only. The number of characters is the closest proxy for population size, although one player could log in several characters in a week period. The light green dots are a three-week rolling average.
Midgard has 39% of the character population in the most recent week, followed by Albion on 33% and Hibernia on 28%.
Realm points over time
Realm points earned have stayed relatively steady for the last two months. However, the divide between the realms has increased since 1.121. In the most recent week, Midgard captured 47% of all RPs earned, nearly the same amount as the other two realms combined (Albion 28%, Hibernia 25%). For the first time since I started collecting data, the balance of realms has shifted and now Hibernia is bottom of the pile.
The effect of 1.121 on RPs earned by realm
The graph below shows the change in share of RPs earned by each realm when comparing two three week periods, one pre-patch (21 June - 10 July) and one post-patch (25 July - 14 August). The figure is the percentage point change (e.g., Midgard have improved 5.8 percentage points from roughly 41% to 47%. Likewise, Hibernia have declined from roughly 35% to 28%). Since the patch, Midgard appears to have improved significantly, while Albion has seen little change. The biggest impact, however, can be seen in the hefty decline of Hibernia.
The graph below shows the effect of 1.121 on realm balance (assuming all other variables are equal - as mentioned above, they aren't, although I can only speculate what other factors have influenced these changes).
The effect of 1.121 on RPs earned by classes
The following graph shows the amount of RPs earned by each class in the three weeks post-patch (orange) and compares them to the amount of RPs earned in the three weeks pre-patch (grey). The graph is arranged from by order of total RPs earned since patch (left to right).
The next graph is a little more complicated. This shows the change in share of RPs earned by each class since 1.121 (i.e., all classes earn 100% of all RPs. Shadowblades increased their share of RPs earned by 5.0 percentage points since 1.121). The data is presented in this way to show a relative change compared to other classes. All classes in the negative have reduced their share of RPs earned. All classes in the positive have increased their share of RPs earned. The net effect is 0.
For those interested in a bit more detail - the standard deviation of RPs earned by class has increased from 10 million/week pre-patch to 13 million/week post-patch (the average RP/week/class pre and post patch is roughly 19m). This suggests that 1.121 has so far reduced, not improved, class balance - i.e., there are more extremes in class performance.
Changes in RPs earned per character per class
Changes in the number of RPs earned by each class might not be because classes are better - it may simply be because more people are playing those classes. The following graph divides the total number of RPs earned by each class by the number of players for each class, for pre-patch and post-patch.
It shows the percentage increase or decrease in the RPs earned per character per class.
It's important to stipulate that this sort of data is quite volatile over a short period and is very susceptible to outliers (see Necromancers - there were only 20 level 50 Necromancers per week post patch, compared to 25 in the three weeks pre-patch. Necromancers earned roughly 700k per week post patch, compared to roughly 500k per week pre-patch. It only takes one player to skew the results on a small dataset. Similarly, this is likely why Bainshees are at the other end of the scale).
I'd argue that Bainshees and Mercenaries are at the bottom of this list because many people have tried the changes (i.e., have logged in), but have either not had much success because the changes were not good for them, or have given up and earned no RPs.
As always, happy to be challenged and look forward to hearing others' thoughts and interpretations of the data.