Limit Player Paragon and Gem Level in PTRs

General Discussion
1

As you know, D3 playerbase ranges from low level players to high level players.

On top, there are players with 10000 paragons, 150 calsessans and 150 gems. These players play the game at a very high level (1% of the playerbase).

On the other hand, average player base 80% of all players, compete below 120 GR at most.

It should also be remembered that power curve starts changing significantly after a certain paragon and at high GR levels.

The problem is that PTRs are about testing the new versions of the game for bugs and balance changes. However, this high level playerbase can unintentionally cause wrong observations about the power potential. Considering the power efficiency curve at average playerbase and top players are so different, what is valid at highest GRs might not be good for average players.

As PTR by name is a testing realm, like in any experimental work, variables must be well defined and controlled. High level playerbase skews the outcomes of PTRs in this regard.

Therefore, in PTRs I suggest:

  • Limit the top paragon level to 800 for all players

  • Limit the legendary gem levels to 100

This should eliminate any bias caused by very high level players and should give more logical testing grounds for everyone’s benefit as what is a well working buff for average players will be better for top players.

Regards

Prokahn

Appendix

A comment below calls this out being emotional so I would like to clear this doubt by putting real time example from Era 11 solo leaderboards.

  • I carried out this statistical observations in another thread https://us.forums.blizzard.com/en/d3/t/analysis-balance-why-wwrend-nerf/5298?u=prokahn-1764

  • Here I would like to take out two example classes: Wizards and Barbarians. Below you can see the distribution of the GR clears by these two classes

https://imgur.com/dSRoK0h

  • If let’s say, the developers decide to balance according to the top clears of these two classes, which are 131 for barbarians while 141 for wizards. Let’s assume that they buffed barbarian resulting in +10GRs. The resulting distribution of the GR clears may look like below (not certain as in reality we may get real balance and the following distribution may be different as my efficiency assessment shows. For this we need new data, but let’s make a crude assumption here). If you look at this new clear distribution, the top clears match perfectly, which constitutes less than 1% of the whole playerbase. In other words, balance will be felt only at the very high levels. However, if you go towards the median, which now lies at 120 GR, Wizards are still 4 GR more powerful. Note that this GR level constitutes about 80% of the whole player base. Therefore, the game is still unbalanced for most of the average players.

Imgur

  • Now let’s assume that the developers decided to stop looking at the very top clears and decided to look for the median clears. As a result, instead of buffing barbarians +10GRs, they buffed them +14GRs. The resulting distribution may look like below. As you see in this figure, not only the most of the playerbase clear distribution across most of the median GRs the same, but also the top clear lines are almost the same. However, there will always be either wizard or a barbarian pushing for the highest possible rift. There will be a real competition in the leaderboards. Moreover, more than 80% of the playerbase will also feel the balance with the top class (Wizards in this case). In this example, the barbarian and wizard distribution overlaps very well thanks to the past balancing efforts on barbarians, but this exact overlap will not be observed in every class. As long as the median values are matched, the feeling of balance should be similar.

Imgur

I hope this clarifies that the intend to balance the game considering the average playerbase is actually a good thing for everybody playing this game and the PTR adjustments must be made accordingly. This will only be possible if paragon is limited to a typical 800 value and the legendary gems are limited to 100, which is pretty achievable by most players (solo or group).

Cheers

22 Likes
2

Yup. Dopiest garbage I ever saw was them letting someone with 10,500 paragon and 150 gems dictate how a class is balanced. Totally moronic.

9 Likes
3

Agree. Tired of having to throw “seasonal buff” and “10k paragon” in posts when someone mentioned the barb on ptr. Then intentionally left that part out or there’s that many people who do not pay attention to these details.

2 Likes
4

I don’t realistically see this happening… Your post seems too emotional and not based on facts.

1 Like
5

In all honesty if we have to ask for this we need to ask for a new development team.

1st year college or university will teach why we don’t balance anything around outliers. Common sense could be used here as well. If this team has neither we need a new team, and I say that in light of people playing other classes which are clearing 146/147 with 3k less paragon. Those classes are taking a 2-3 gr Nerf while barb takes a 7 gr Nerf from 140 on ptr.

If balance was the goal this team missed the mark totally by nerfing Lamentation. If the goal was to wreck season 19 for barbs and others who may have wanted to try ww, force another season of wall charge play and please trolls the team succeeded brilliantly.

Revert the lamentation Nerf Devs.

4 Likes
6

FYI: Just added this explanation as appendix.

1 Like
7

I like this idea.

Your gr efficiency thread is much, much more relevant. I still have problems with your graphs as you are analyzing a population’s tail of unknown length where a normal distribution statistic is not applicable.

8

Well just an example man. If balancing needs to be done really, more in depth look is needed. I am just trying to convey information to people to let them see the power differential across GR levels. +10 GR may actually work as well depending on the class. For example, for barbarian, it does, gives similar efficiency curves but distribution may be different. However it won’t be off by a large margin. I am trying to explain the concept here for the average mind.

Concept is limiting paragon and gem level is good for balancing and uniformity of the recorded data.

9

I realize that, but I strongly suspect that Blizzard can pull all players who have a range of paragons and make pairwise or more sophisticated comparison even within the context of noisy PTR data relative to the live era. When I looked at the PTR #40 barb vs. America region #40 in non-seasons. I think I know what the developers might have been thinking. I strongly suspect if you compare the rank #40 barb if the buff goes live as is, #40 will clear a 133. This is actually powerful in comparison to other classes (not named wizards that are being nerfed.)

10

Absolutely. If you don’t normalize the initial conditions the results aren’t even worth looking at. You can’t compare data sets with multiple large variables unaccounted for and draw meaningful conclusions.

I strongly suspect Blizzard can’t examine data sets without a written manual and a Ouija board.

2 Likes
11

A side effect to this might be many people would start playing mainly the PTRs, because these would introduce more even playing field. Which in turn might make the developers to consider making such change official for Seasonal play and bring many players back to the game. This sounds amazing!

1 Like
12

I agree. but can we also toss in a vendor and open all the roll slots at myriam. so that we can test with your 800 paragons and perfect rolls. so that there is no question about the power.

1 Like
13

Yes maybe item drops are all set to primals etc. Or all existing items are change with primal counterparts etc. Good idea.

14

I’ll start by saying I agree on the paragon/gem limit cap for some PTR testing. However, it should be noted, that some builds only become viable at high paragon due to the asymmetric scaling of GRs and damage/toughness from primary attributes. So in capping the PTR, these builds would go unnoticed and could be problematic in higher GRs. For an in depth example of this asymmetric GR scaling, see my post on Dexterity scaling vs GR progression.

Next, as a mathematician, I always appreciate players using math to support their decisions and opinions. However, there are a number of flaws in your appendix:

  • Is the data truly Gaussian? (I think actual player data might be skewed, with difference variances for each class)
  • Does the data consider “Barbarian” and “Wizard” with a single curve for every possible build?
  • How is paragon incorporated into these graphs? If a class is weaker, they generally obtain less paragon per season than other classes unless they have multiple characters.

These simplifications make it almost impossible to tell the whole story with just a few plots. Do you have access to the recent (current era) and exact data (specific distribution by class, build, paragon, etc.)?

As someone who has analyzed D3 game balance from a PVP perspective, I am confident it is not possible to balance the game with flat damage and/or toughness flat percentage buffs/nerfs.

1 Like
15

Simply put, he is ignoring population size. For example, imagine two populations where a characteristic has identical values (e.g. average value of 200 with a standard deviation of 30). Lets assume that the two populations differ in size (e.g. 5,000 vs 10,000). He is taking the top 1,000 from each population, that is 20% and 10% respectively, and analyzing the tail of the population using a normal distribution which is completely invalid in my mind. In his analysis, the population where the top 10% were analyzed would had a higher characteristic value than the population where the top 20% were analyzed.

He and I have both analyzed the era 11 data from Blizzard’s API. Unfortunately, the data is limited. For the top 1000 on the leaderboard, it is easy to get GR clear and its time and paragon level of player. The information on builds needs to be done manually in game and is impractical time-wise.

His better analysis is present here. You can check out the fist few posts.

16

Yeah, that’s what I feared. I’m not paid to analyze D3 and just do this as a hobby! I analyze components used in particle accelerators instead for money!

17

His greater rift efficiency is imperfect but it is informative if you keep in mind the limitations of the dataset. I was analyzing D3 manually for awhile. Blizzard’s API makes life much easier in that regard.

Nice career choice.

18

I explained it couple of times. Distribution is flawed. Micro was doing the averaging so here is a complete distribution. Buff numbers should not be decided here. My real analyses were mentioned many times earlier. They are crude approximationsm And I am an engineer not a scientist. For me, as long as the numbers get the job done,I dont dig too much details. :slight_smile:

1 Like
19

There is another solution for this:

On PTR, disregard player paragon lv. Put a paragon goblin in town and gives players 3 paragon bags of choice 800, 2000 and 5000. This will give players the ability to test the game on early stage (800), mid game (2k) and end game (5k).

1 Like
20

While this would be what a QA team could use, the vast majority of players would take the 5k bag and just laugh at the other options.

1 Like