Solo game balance: Non-Season Era 11 and 12 Data and Analysis

Introduction


Here, I present data and its analysis for the final era 11 solo leaderboards (patch 2.6.6) as well as the current era 12 data (patch 2.6.7a/b). Era data is non-season and is used to exclude confounders caused by seasonal buffs.

The datasets being analyzed are imperfect in that there are biases in the number of players pushing each class. As an example, demon hunters and witch doctors did not receive any buffs or nerfs in patch 2.6.7; however, it is much easier (i.e. lower GR clears are needed to get onto their respective leaderboards). At the highest ranks of each leaderboard, the era 11 and era 12 data for these 2 classes is similar but as one goes further down, a gap broadens in each leaderboard where era 11 was more competitive than era 12.

Table of contents

1. Data Sources
2. Top 10 worldwide solo clears in era 11 and 12
3. Era 11 and 12 GR Tiers in relation to solo leaderboard rank (America Region)
4. Solo Greater rift efficiency in era 11 and 12 (America Region)
5. Conclusions


1. Data Sources

  • A. Blizzard’s API was used to download the entire final era 11 leaderboard and the current era 12 leaderboard. The era 12 leaderboard was accessed on January 3. The final era 11 leaderboard includes the top 2,000 players while the current era 12 leaderboard is top 1,000.

  • B. Zero empathy’s website http://ranks.zeroempathy.org/e11 was accessed on January 4 to obtain the data for the top 10 solo worldwide (America/EU/Asia/China) for each class for era 11.

  • C. https://hero.d.163.com/?page=0 was accessed on January 3 to obtain the top 10 solo worldwide (America/EU/Asia/China) for each class for the current era 12. Since there was a delay in updates to this website, I used the API to find a GR clear from China that occurred on January 3 that had not yet been included.

Back to table of contents


2. Top 10 worldwide solo clears in era 11 and 12

  • Top 10 Worldwide Era 11

Era 11 Barbarian Crusader Demon hunter Monk Necromancer Witch doctor Wizard
Rank #1 135 138 140 136 147 142 148
Rank #10 131 136 136 132 138 138 143
Median of Top 10 132 137 137 133.5 141.5 140 143.5
  • Top 10 Worldwide Era 12

Era 12 Barbarian Crusader Demon hunter Monk Necromancer Witch doctor Wizard
Rank #1 146 150 140 143 136 142 144
Rank #10 142 147 135 139 134 137 140
Median of Top 10 143 149 136 140 135 139 140.5
  • Top 10 Worldwide Era 11 and 12 (Sorted from highest to lowest median top 10)

Era Era 12 Crusader Era 11 Wizard Era 12 Barbarian Era 11 Necromancer Era 12 Wizard Era 12 Monk Era 11 Witch doctor Era 12 Witch doctor Era 11 Demon hunter Era 11 Crusader Era 12 Demon hunter Era 12 Necromancer Era 11 Monk Era 11 Barbarian
Rank #1 150 148 146 147 144 143 142 142 140 138 140 136 136 135
Rank #10 147 143 142 138 140 139 138 137 136 136 135 134 132 131
Median of Top 10 149 143.5 143 141.5 140.5 140 140 139 137 137 136 135 133.5 132
  • Graph of Era 11 & 12 Worldwide Top of Leaderboard Data Top 1000 (sorted from highest to lowest with respect to median of top 10 GR clear)

Back to table of contents


3. Era 11 and 12 GR Tiers in relation to solo leaderboard rank (America Region)

  • Era 11 Data (America Region)

Era 11 Rank Barbarian Crusader Demon Hunter Monk Necromancer WitchDoctor Wizard
1 132 137 138 132 138 142 143
10 125 133 128 127 130 132 138
50 120 126 125 120 122 125 133
100 116 124 124 117 119 121 131
200 114 120 122 114 115 119 130
500 110 115 120 110 109 113 125
1000 107 110 117 105 103 109 122
1500 105 108 115 104 100 105 120
2000 103 105 115 101 95 103 117
  • Era 12 Data (America Region)

Era 12 Rank Barbarian Crusader hunter Monk Necromancer doctor Wizard
1 143 148 135 140 135 135 142
10 136 142 126 135 125 133 132
50 131 137 122 127 115 120 125
100 130 134 120 124 110 116 123
200 125 130 117 120 105 110 120
500 121 124 112 114 93 100 114
1000 117 116 110 108 80 92 106
  • Combined Era 11 & 12 Data Top 1000 (sorted from highest to lowest with respect to rank #1 GR clear) (America Region)

Era Rank Era 12 Crusader Era 11 Wizard Era 12 Barbarian Era 11 Witch Doctor Era 12 Wizard Era 12 Monk Era 11 Demon Hunter Era 11 Necromancer Era 11 Crusader Era 12 Demon hunter Era 12 Witch doctor Era 12 Necromancer Era 11 Barbarian Era 11 Monk
1 148 143 143 142 142 140 138 138 137 135 135 135 132 132
10 142 138 136 132 132 135 128 130 133 126 133 125 125 127
50 137 133 131 125 125 127 125 122 126 122 120 115 120 120
100 134 131 130 121 123 124 124 119 124 120 116 110 116 117
200 130 130 125 119 120 120 122 115 120 117 110 105 114 114
500 124 125 121 113 114 114 120 109 115 112 100 93 110 110
1000 116 122 117 109 106 108 117 103 110 110 92 80 107 105
  • Graph of Era 11 & 12 Data Top 1000 (sorted from highest to lowest with respect to rank #1 GR clear) (America Region)

Note: Data for Rank #1500 and #2000 were only available for era 11 and not era 12 through Blizzard’s API. This data is affected by the number of players for each class as illustrated by comparing era 11 and 12 demon hunters and witchdoctors that were not buffed/nerfed in patch 2.6.7.

Back to table of contents


4. Solo Greater rift efficiency in era 11 and 12 (America Region)

This is a modification of the method initially used by Prokahn (see reddit https://www.reddit.com/r/Diablo/comments/cv6bcx/analysis_era_11_barbarian_grift_clearing/ and links therein and several of his posts on this forum). Briefly for my analysis, GR efficiency was defined as P=(A1,000,000,000)/(BC) where A = GR tier, B=GR clear time in milliseconds, and C=player’s paragon level at time of clear. Blizzard’s API reports clear time in milliseconds where 15 minutes = 840,000 milliseconds. This method corrects in part (but not entirely) for discrepancies in the number of players pushing each class (see Graph D). For simplicity in the graphs presented below for every greater rift tier, a single data point is used to represent the average “GR efficiency” for all players at that specific GR level. * denotes that the lowest GR clear for each class is not graphed as the data is “incomplete”, since it spans the leaderboard cutoff. Multiple graphs are shown:

  • Era 11 GR Efficiency (Top 2000)

  • Era 12 GR Efficiency (Top 1000)

  • Era 11 and 12 GR Efficiency (Select Classes)

  • GR Efficiency of classes at GR tier 125 and 120 (higher numbers at a spefic tier correlates with better efficiency)

Efficiency at indicated GR tier Era 12 Crusader Era 11 Wizard Era 12 Barbarian Era 12 Monk Era 12 Witch Doctor Era 12 Wizard Era 11 Witch Doctor Era 11 Crusader Era 11 Necromancer Era 12 Demon Hunter Era 11 Demon Hunter Era 11 Barbarian Era 11 Monk Era 12 Necromancer
125 44.58 43.41 40.98 40.65 38.13 34.03 33.27 31.54 27.58 26.43 26.15 25.83 25.49 21.46
120 54.31 59.29 56.00 50.74 40.64 48.08 41.42 40.17 33.12 37.86 38.94 33.45 32.56 31.09

Back to table of contents

5. Conclusions

Balance is an ongoing process where underperforming and overperforming classes remain. In era 12, crusaders are clearly overperforming. Necromancers need buffs. Barbarians appear to be overperforming, especially in light of Matthew Cederquist balance comments. It will be interesting to see Blizzard’s upcoming blog post on game balance and future changes in store as they improve balance between classes for their respective top builds.

Back to table of contents

11 Likes

What a waste of energy and time.

13 Likes

I think from the point I decided to do this, I spent a few hours on getting the data summaries and graphics while concurrently watching TV. Once you know how to get the data and use excel, it is fairly simple copy and pasting. Even the formatting (e.g., making clickable table of contents and clickable back to table of contents) was relatively straightforward and easy to do using a template that Morgan provided. With more knowledge and handson experience, the quicker a seemingly complex task can be accomplished.

Data-driven posts and analytics are definitely not everyone’s cup of tea. Personally, I think comparisons of era 12 classes to era 11 wizards is interesting considering the new nerf/buff cycle.

Given Matthew Cederquists comments on game balance, the data potentially has implications for crusaders, barbarians, and necromancers in particular.

Some appreciate evidence-based conclusions

4 Likes

There is nothing to compare between era 11 and 12 wizards, the macros are almost identical and for top leaderboard push bazooka was and still is better than vyrs. At least vyrs was nerfed so you will see a difference for top 1-2000 but outside of that it boils down to how much the r1 pushed.

For necro it is interesting tho.

4 Likes

You missed that 146 clear by Barb. CompletedTime timestamp is 1578023227000 or GMT : Friday, January 3, 2020 3:47:07 AM. I don’t understand why you would use any data source other than Blizzard’s API.

This means Barb is probably getting nerf next patch.

This was a blue post commenting on Crusader doing solo GR 146 on EU non-seasonal.

This barb from CH did 146 solo with 9477 Paragons.

5 Likes

The GR 146 clear for barbarians is not yet listed at
https://hero.d.163.com/?page=0

There tends to be a smidgen of delay in how current the listed website is. I will update this post when this website updates.

Do you have a screenshot of the current China or worldwide top 10 for barbarians?

From Blizzard’s API, it reads:
“API requests cannot retrieve global information from the China partition, and vice versa.”

Is there a work around? I get an error message when trying to download the China data.

4 Likes

Exactly. That is why you should not use it.

https://gateway.battlenet.com.cn/data/d3/era/10/leaderboard/rift-barbarian?access_token=

Put your access_token at the end and you have it.

7 Likes

Thanks for the link. It did work with my access token. I will update my post in a few minutes.

5 Likes

Clearly, barb is in gray. Without hard cast, it may still under perform a little bit. Only few players can properly use hard cast right now.

Bliz and players still need some time to see the true power of zodiac rend. The data of barb era 12 is pretty close to WD era 11. Frequently changes the numbers of certain build is not a reasonable choice. This is also why gray zone was introduced by Matt.

Currently, the only issue is crusader in NS. Fix that issue, buff DH Nec and fix bazooka, buff other wiz builds(channeling builds, etc). In my opanion, that’s is what they should do next season.

1 Like

@JustinFan,

I differ with certain elements of your interpretation.

With additional time, the top GR clears will only increase. It will be interesting to see what Blizzard’s blog post says. At the top end of the leaderboard, the top barb clear is +5 GRs over the target of 141. Within 1-2 GRs is considered good balance. The gray is 3-4 GRs. 5 GRs or more is most likely “correction” zone. Matt has made a couple posts already about this.

See:

The data is much closer to wizards of era 11 than witch doctors. Double check the median of the top 10 worldwide GR clears and GR efficiency (overall and at specific GR levels for example GR 125 and GR 120). You can also look at GR tiers as a function of leaderboard ranks. For example, rank #100 era 11 wizard cleared GR 131 while currently barbarians are at GR 130 and this era is a couple of months old.

4 Likes
  1. I do think Matt means this single data. They will balance it according to the entire database, distribution. Not as simple as your data.

  2. Top 1000 mean or similar, are not reliable. E.g. Top 1000 barb players may have higher avg paragon than WD. That’s why Prokhan’s analysis is more reliable than yours. Try his method.

1 Like

Prokahn’s method was used in section 4 that discusses “Greater Rift Efficiency”. This method takes into account paragon levels. You may want to look carefully at the graphs and the table. The conclusions between the different metrics are largely consistent.

Here is the graph again

4 Likes

Sorry. I can’t understand your “largely consistent”.

You already get the api. Is there any techinical issue for you try Prakhan’s method?

Prokhan method was used in section 4 of the original post. I have already posted the results in the first post in this thread.

3 Likes

If this greater eff is Prokhan’s method.

The green and red are close. Wiz is higher. right?

Let me walk you through section 4. There are 3 graphs.

  1. The first is GR efficiency for era 11. You will notice that the black dots are outliers. These dots represent wizard era 11 GR efficiency. It is clear that this class is an outlier and overpowered in comparison to all other classes.

  2. The second graph is era 12. The data is more noisy (presumably because the era is younger). However, even with all the data points you can notice that the yellow triangles (representing) crusaders are grouped toward the upper edge of the distribution.

  3. The third graph is a simple representation of only four classes where one can directly compare era 11 wizards and witch doctors to era 12 crusaders and barbarians. In general, era 11 witch doctors represented by red triangles remain grouped separately (for the most part) while the 3 other classes are more overlapping (where era 12 crusaders are the most overpowered in terms of GR efficiency).

4 Likes

Thanks for the explanation. Here is my suggestion.

In my opinion, people use a curve (e^ - x like functions) to approximate the data. Before we say two curves are close enough, first we have to focus on an interval on the horizontal axis. In your data, we should focus on the right half of the interval (Cuz we all care more about 130+ clears not 110 clears), then calculate the area between the two curves.

My personal conclusion is:

barb is close to wiz on mid-level grifts: e.g. 120-130.

is more close to WD on top-level grifts e.g. 135+. (This is what we care most)

1 Like

I do not know that it is fair to claim that “we all care more about 130+ clears”.

If you want to only consider 130+ GR clears, than it makes sense to look at the worldwide leaderboard to increase the number of data points that are available for analysis. At the very top end, you can look at the top 10 worldwide leaderboard as done in section 2.

Rank Era 11 Wizard Era 12 Barbarian Era 11 Witch doctor
Rank #1 148 146 142
Rank #10 143 142 138
Median of Top 10 143.5 143 140
4 Likes

Pick up data of 130+ clears. Still compare the eff curve. Not the mean/median. As I said before.

2 Likes

For ease, I took the worldwide data from the zero empathy leaderboard.

The area under the curve for GR clears 134 and above for era 11 witch doctors is 186.8 while for era 12 barbarians is 267.0. As you mentioned their difference in GR efficiency becomes even more pronounced at lower GR levels where barbarians are more efficient.

I could not go down to GR 130 because the ZE leaderboard only lists the top 200 worldwide (and the rank #200 barb cleared GR 133).

4 Likes