General class balance observations

Warbands Data
% Duelist in top 15000: 5.95
% Duelist in top 3000: 5.63
% Duelist in top 600: 6.33
% Duelist in top 120: 8.33
% Marauder in top 15000: 19.95
% Marauder in top 3000: 18.40
% Marauder in top 600: 18.83
% Marauder in top 120: 16.67
% Ranger in top 15000: 21.50
% Ranger in top 3000: 22.37
% Ranger in top 600: 20.17
% Ranger in top 120: 21.67
% Shadow in top 15000: 8.07
% Shadow in top 3000: 6.77
% Shadow in top 600: 5.33
% Shadow in top 120: 4.17
% Templar in top 15000: 5.58
% Templar in top 3000: 4.60
% Templar in top 600: 4.00
% Templar in top 120: 1.67
% Witch in top 15000: 20.40
% Witch in top 3000: 19.30
% Witch in top 600: 17.33
% Witch in top 120: 22.50
% Scion in top 15000: 18.55
% Scion in top 3000: 22.93
% Scion in top 600: 28.17
% Scion in top 120: 25.00

Warbands: increased/reduced chance to top 4%
1. Scion +51.9%
2. Duelist +6.4%
3. Marauder -5.6%
4. Ranger -6.2%
5. Witch -15.0%
6. Templar -28.3%
7. Shadow -34.0%

Tempest Data
% Duelist in top 15000: 5.03
% Duelist in top 3000: 5.23
% Duelist in top 600: 5.00
% Duelist in top 120: 4.17
% Marauder in top 15000: 28.45
% Marauder in top 3000: 29.80
% Marauder in top 600: 28.67
% Marauder in top 120: 37.50
% Ranger in top 15000: 12.67
% Ranger in top 3000: 11.50
% Ranger in top 600: 10.17
% Ranger in top 120: 12.50
% Shadow in top 15000: 5.54
% Shadow in top 3000: 4.40
% Shadow in top 600: 5.33
% Shadow in top 120: 5.00
% Templar in top 15000: 9.05
% Templar in top 3000: 7.23
% Templar in top 600: 5.50
% Templar in top 120: 1.67
% Witch in top 15000: 18.83
% Witch in top 3000: 18.70
% Witch in top 600: 19.17
% Witch in top 120: 14.17
% Scion in top 15000: 20.43
% Scion in top 3000: 23.13
% Scion in top 600: 26.17
% Scion in top 120: 25.83

Tempest: increased/reduced chance to top 4%
1. Scion +28.1%
2. Witch +1.8%
3. Marauder +0.8%
4. Duelist -0.6%
5. Shadow -3.8%
6. Ranger -19.7%
7. Templar -39.2%

Overall
1. Scion (1st, 1st)
2. (tie) Duelist (2nd, 4th)
2. (tie) Marauder (3rd, 3rd)
4. Witch (5th, 2nd)
5. Ranger (4th, 6th)
6. Shadow (7th, 5th)
7. Templar (6th, 7th)

My conclusions regarding class balance
Duelist, Marauder, Ranger, Witch - These classes are fairly well balanced right now. In all of these cases, either there is a balance in the mass population compared to the top 4% of players, or there is an apparent imbalance which the very highest-level players (top 0.8%) seem to have sorted out, or even converted into an advantage.
Shadow - This is a bit of an awkward situation from a data perspective, as (surprisingly) the Shadow is showing very poor performance in Warbands while showing roughly balanced (but slightly suboptimal) performance in Tempest. Although the Warbands data is troubling, since the Tempest data points towards a relatively balanced situation, I wouldn't recommend any major changes to passives in the Shadow area at this time. Maybe a minor buff or two, and definitely nothing which could be construed as a nerf.
Templar - The data shows clear evidence that this class is underpowered at present (especially at the very highest levels of play). Since most of the Templar's skills are available to Witches and Marauders, both of which are balanced well enough in the current meta, I would strongly recommend improvements to the Templar passive tree, especially the first few nodes which the class cannot avoid, in order to make Templar a stronger option.
Scion - The data shows clear evidence that this class is overpowered at present. Most likely this is due to the high concentration of jewel sockets available in the Scion's starting nodes, as little else about the Scion starting area seems appealing. The available options seem to be to either limit the Scion's access to jewel sockets relative to now, or to reduce the power of individual jewel affixes to make the jewels themselves a little less overwhelmingly powerful; of these two options, I feel the former destroys the Scion's passive tree identity, and thus would favor the latter.

Where does the urge to make all classes equal come from? Especially making characters equally viable in SC and HC... why? If there is a class that works better for HC and less good for SC, so be it. What is the problem?

explain the Duelist OP please. stuff like 50%+ etc

and what you compare is classes when what you should be doing is comparing FOTM builds

stuff like iron will incinerate (incinerate in general), bows and general cheap and powerful builds determine class distribution, not class stats (that are pretty moot in general as one frequently visits 2+ starting areas thus making a 'marauder' build a duelist and templar build as well

"

sidtherat wrote:

explain the Duelist OP please. stuff like 50%+ etc

TLess than 6% of WB played Duelist but it made up 9 of the top 100.

This is about general passive tree balance, strengths of various nodes, etc, more than pointing out specific OP builds. Builds aren't irrelevant to the discussion, but shouldn't become the main focus.

It's kinda hard to make actual statements on each class without first looking at build/item/ect that would make them that popular or unpopular to begin with, along with why in each league.

For example marauder has always been popular in HC (the mode where you do not want to ever die) due to the pure fact they are suppose to be the "tanky" class. It would be hard to change that unless you suddenly want to remove the "tanky class" concept from the game. Essentially homogenizing the game for the sake of balance.

Even if you do that to the marauder for HC, it may imbalance the league further to ranged classes and the most tanky ranged class would take the marauder placement. Thus by fixing a "class imbalance" you just replaced it with something else really.

Edited OP for a more clear, less slapdash opinion section. Data is the same.

To reiterate: buff Templar, nerf Scion/jewels, rest is really nitpicky.

balance based on looking at statistics, I think rory and qarl are capable of fucking up the game via this sort of detached, clueless method of evaluation on their own tbh.

I think none of these changes should really be made in all honesty, if GGG want to listen to feedback on class balance they should listen to highly experienced players who make a lot of builds and know why you would chose one class over another based purely on knowledge without looking at statistics at all. Nerfing duelist in an unspecific way because hes being used for unknown reasons is just flat out ridiculous. Making ranger more tanky when the reason people chose ranger is because its more tanky than the other options around there... yeah.

I really do feel like you should play this game more and comment less on it tbh.

This post illustrates the achilles heel of abstract statistical analysis - too much data, not enough context. As posters above pointed out, it's obtuse to generalize about class balance without reference to the builds most strongly associated with each class. To understand the Duelist's dominance in Warbands Class Rankings, you have to evaluate the builds that put him on top. Statistical data is nothing more than evidence, and the skill tree is just the starting point. Insight comes from connecting the dots, not simply tabulating them.

"

Snorkle_uk wrote:

Nerfing duelist in an unspecific way because hes being used for unknown reasons is just flat out ridiculous.

Well no shit. Obviously one would need to actually solve the mystery of it all first. Maybe no action is required, idk. But it would be nice to know why, and I'm advocating looking into some potentially OP shit.