Diablo III Blog: Balancing Class & Set Design
The balance and set design blog that was promised to go live before the patch 2.6.8 PTR is now available for your reading pleasure! This blog provides incredible insights into how the Diablo III team determines balance, and gives players a better idea of what to expect from from future balance changes.
Originally Posted by The Diablo III Team (Official Post)
Developer Insights is our blog where we reveal some of the processes and philosophies behind our updates and designs. We have a lot going on in the world of Sanctuary these days, so please note that the following entry comes from the Diablo III team. Enjoy!
We’ve been working on multiple content patches for Diablo III, including Patch 2.6.7. which we released last November. This included two new class sets, one each for Crusader and Monk, and heavily revamped Whirlwind gameplay for the Barbarian. The result was shattering to the endgame meta—a result we were both ambitiously hoping for and pleased to see.
However, our work is far from done! We’re now developing Patch 2.6.8 and want to provide context on the approach we’re taking to address class set balance in our future patches before unleashing you onto the PTR. Let’s dive right in!
Table of Contents
- Paragon and Greater Rift Levels
- Solo vs. Group Design
- ”Creative” Game Mechanics and Animations
We look at a ton of data when it comes to balancing Diablo III. Data comes in many forms, including build guides, gameplay from your favorite streamers, and the leaderboards. Most importantly, we also track our own data internally, which allows us to see a lot more than just the top 1000 player clears for each category.
None of that data should be taken in a vacuum; it all comes together to paint a much bigger, constantly shifting picture. Here are a couple things we consider:
- Individual performers and their Paragon Level
- Paragon level directly impacts a player’s long-term progress; it’s important that this is accounted for when comparing different points of data
- The intended design and in-practice functionality of each class set
- What does this set do best? Does it clear a lot of smaller monsters, or is it a good Rift Guardian killer? Does it serve a support role? Is it better in group play over solo? Does it need to be doing something different?
And here are a couple things we specifically avoid when considering overall class balance:
- Seasonal buffs and their impact on overall power
- We want to design fun, engaging Seasonal buffs without worrying about how they might impact class balance
- Data from non-Seasonal players is especially helpful here
- Some buffs may be better or worse for different classes, but since Season effects are temporary, we’re okay with this
- Direct 1:1 class comparison
- While we want each class to perform similarly, it is still important that classes retain their unique class fantasy
- Each class should be able to accomplish similar goals, just in different ways
When balancing, we need a point of reference to work around. The “ideal” class set performance for Diablo III is approximately Greater Rift 130, solo, and assumes the character has 5000 Paragon levels. That might sound high to some of you and low to others. If so, that’s good—it means we’ve landed in a middle ground that’s beneficial to the most players!
Assuming a player is at 5000 Paragon, here’s where all the classes landed, on average, a few weeks into Patch 2.6.7 for non-Seasonal play, aggregated across player leaderboard data world-wide:
From the data above we have a pretty good idea which classes are under or over performing.
Just for fun, here’s a similar comparison during the same timeframe upscaled in Season (where many players hadn’t reached 5000+ paragon):
It’s interesting to see which classes most likely benefit most from the Seasonal buff—it’s also evident which classes are probably not being played as much or pushed as hard as others.
While this is our goal, we also recognize we aren’t always going to hit it perfectly. Like many games, Diablo III has a lot of mechanical details. A single change can ripple through many other parts of the game, so it’s important we’re mindful of what each change can affect. We also realize that, even with special care, it may take us a few tries to achieve our intended result. To account for this, we have a scale for error, based on how a class is performing above or below our guideline:
- +/- 1-2 Greater Rift Levels: Very close. Probably fine, when accounting for random elements (the perfect “fish”) or high player skill cap (excellent play and timing).
- +/- 3-4 Greater Rift Levels: The warning zone. We need to watch for buffs/nerfs in this area, but action may not yet be necessary. Time to keep an eye on it!
- +/- 5 or more Greater Rift Levels: Warrants significant change. At this range, it’s clear that something is over (or under) performing and needs to be addressed.
Bear in mind this is aggregate data; the above tables only cover overall class performance rather than individual class set performance. Better performing builds may be equalizing out weaker ones when it comes to the broader picture. Because we make changes at the set and item level, we must additionally parse out data by build performance (whether it’s a class set or Legacy of Nightmares-based). The above method is meant to serve as an example of our general approach and informs us which classes most likely need more attention first.
An important note, especially for those of you who have been looking for this answer: we have no plans to increase the Greater Rift level cap beyond 150. As to why, the short answer is because it causes more problems than it solves.
The long answer is that continuing to expand the end game through additional Greater Rift levels hasn’t been the healthiest approach for Diablo III. At this late stage of game development, we’d prefer to focus on making the current game the best, and most varied, experience it can possibly be. We hope to accomplish this by (1) continuing to add new builds and (2) improving existing builds that have fallen behind. Maintaining a cap, and even backing away from it a little, will allow us to focus on greater gameplay variety.
A common misconception is that we balance solely around 4-man groups. While we certainly take it into account, group play is not our only focus because not all players enjoy playing in groups. We want to make sure that the content we’re designing can be enjoyed by the most players, so our design decisions should take both styles into account. If we were to design solely around one style of play, the other would be severely impacted (and likely a lot less fun).
There is one notable exception to this. zDPS, or “zero Damage Per Second,” builds tend to only thrive in group play. The team debated for a long time as to whether this was a style we should encourage or actively avoid (or possibly even remove). Any Diablo game is, at its heart, about killing monsters and getting loot, so we considered heavily whether this gameplay fit the spirit of the game we had made.
zDPS is a style of gameplay that some players like, but not all do. We ultimately decided that it’s good that there are unique ways to play the game, and we don’t want to take that fun away from those who enjoy it. However, we’re also not actively trying to create more zDPS builds. Our intent is to design new item powers and sets that either facilitate entirely new builds that can be used solo or in groups, or to give more item support to some of the most requested class skills. Inevitably, regardless of what we add, we know the community will find combinations we never anticipated—and we can’t wait to see how those play!
There have been a handful of problematic gameplay styles that have emerged over time, largely due to creative use of snapshotting or taking advantage of the ability to cancel skill animations. This impedes our efforts at overall game balance, and it makes it very difficult for us to change skills or items without overly punishing the whole class when these tactics aren’t used. (We’re looking at you, Bazooka Wizard and Stutter Steppers.)
We want to be able to address the balance issues these builds pose, but we also don’t want these classes to fall behind as a result of these changes. So how are we handling this? It’s admittedly going to be a game of whack-a-mole as we make fixes and other issues arise through player experimentation. Our first area of concern will be disabling animation cancellation on certain skills. The Crusader class is the most subject to this (though not alone) and we’ll ensure no class falls behind as a result of losing this “trick.” After we tackle this specific issue, we’ll continue to review and revisit areas of the game where similar mechanics are having a negative impact and act accordingly.
Thank You for Reading; We’re Reading Too!
If you’ve read everything, thank you for your attention and taking the time to look through our lens! There’s a ton of info here, and you might not agree with it all; that’s okay. We love to hear what you have to say, so please make sure you’re sharing your feedback in your Class forum, the General forum, or the PTR Feedback forum during our next testing phase. We love Diablo, and whether we’re reading global chat, hanging out on our favorite subreddits, or bantering with you through social media, we’re glad we get to share this journey with you.
We hope this lends an inkling into where we’re going while our other game teams are hard at work on the future of Hell. Thank you for playing—we’re grateful to have you with us along the way!
-The Diablo III Team-