Ion alluded to some of this in his post last week, but essentially the goal here is to address what we believe to be the core of the issue with the 7.0 final traits: the large gap in player power it created between players who spent a lot of time farming AP and those who spent their time on other endeavors. To that end, there's four key changes:
#1: The individual ranks are less impactful. This was honestly one of the biggest issues with the 7.0 design. Grinding out a couple million Artifact Power for a 0.5% raw damage increase was just too lucrative compared to other methods of endgame progression - even eclipsing gear for some players. The goal for the new 7.2 design is that the next rank is still an increase, and you won't turn it down, but it's not your primary focus.
#2: The rate at which the cost for the next rank increases is higher. In the 7.0 design, someone who farmed twice as much AP as you had roughly twice as many ranks as you. While rewarding the extra effort isn't a bad thing, it doesn't need to be nearly that rewarding. By making each rank's cost increase exponentially, we can help ensure that you're never too far behind even if you aren't spending as much time farming AP. It also means that, as Artifact Knowledge increases, it'll be easier for alts or newer players to catch up.
As a quick aside, to put some extra context on both of those changes: we always want Artifact Power to be of at least some value to you. It's fine to reach a point where you're not going out of your way to earn it, but it's purpose is to be a fairly reliable form of progression. If you spend an evening raiding, or run a few dungeons, or do some PvP, but don't get any gear upgrades, you should still be able to say "at least I earned some Artifact Power" with some level of satisfaction.
Anyway, key change #3: The new trait gives a primary stat bonus instead of a percentage-based increase. With the 7.0 design, as your gear improved, so did the total benefit you were getting from your final trait. Changing to a primary stat bonus means it's giving roughly the same benefit to someone at ilevel 900 as it does to someone at 850. Again, the goal here is to reduce the overall power gap.
And finally, #4: It's a proc. I know anything that involves RNG is often controversial, but this is, in my opinion, a great example of where it's extremely useful. This is for two reasons.
First, it kind of muddies the waters a bit. When you wipe on a boss at 1%, or just barely miss a kill window, it can be easy to say "if Todd was doing 2% more damage we'd have won." But when it's a proc, you can't actually be that sure. Maybe Todd needs more AP, or maybe he just got unlucky with procs. Maybe the wipe wasn't Todd's fault at all. Maybe you should be a little nicer to Todd.
Second (and more importantly), it allows for player skill to play more of a factor. If you're the sort of player who can pay attention to procs and adjust your rotation on the fly (say, a healer who chooses to use cheaper spells while it's active, or a damage-dealer who saves a charge of their hardest-hitting ability), you're going to get more value out of the new trait than someone who ignores it.
I've seen some initial feedback that indicates some specs are likely to benefit from these more than others - to some degree that's expected, but specific feedback on which specs those are and why is very helpful. We're still actively tuning and tweaking things, so please keep that feedback coming.
Artifact Power has been a hot topic lately, both around the community and within the development team. With Patch 7.2 on the horizon, introducing both new artifact traits and additional Knowledge levels, we have been reflecting on the way the system has unfolded during the first months of Legion, and evaluating changes based on the lessons we have learned thus far.
First off, a look back at where we started.
From the outset, Artifact Power was intended to serve two intertwined purposes: First, it offered max-level progression that was not entirely item-driven, along with choices and elements of character customization as players traversed their trait trees; second, it was meant to serve as a universally desired, consistent reward from all types of content.
In crafting the systems that delivered Artifact Power, we weighed the merits of hard caps versus a smoother system of diminishing returns. We had extensive experience with hard caps, through multiple past iterations of currencies like Valor Points and Conquest Points, and wanted to avoid several of the downsides of that approach. For example, a cap inherently feels like more of an expected quota, where missing a week or falling short of the cap puts you clearly, and potentially permanently, behind the curve.
Instead, as everyone knows, we settled on an open-ended system of diminishing returns. Without any hard caps on how quickly players could earn AP, it was essential to have some sort of limiting mechanism on the gap in power between players of different playstyles, and different levels of time investment. We accepted the admittedly complex design of Artifact Knowledge because it solved this problem, effectively reining in the size of this power gap. Players trying to progress past the expected artifact level for their Knowledge would run into those rapidly diminishing returns, while those who played less than that would have Knowledge as an accelerator to help them catch up to the cutting edge. When Emerald Nightmare was new content, while the average raider was at 20 or 21 points, the most dedicated might have been at 24 or 25 – a relatively modest gap.
Now, where things went wrong…
We feel that we made two major missteps with the Artifact Power system that increasingly manifested themselves as we got deeper into Patch 7.1 and 7.1.5. And both of them served to undermine that core goal of ensuring that the gap between players with different levels of time invested into the system could not grow too large.
First, the cost of ranks in the 20-point final trait remained relatively flat, as opposed to the rapid exponential scaling up to that point. This meant that someone who spent twice as much time gathering AP as I did would have roughly twice as many ranks as me. Instead of the 24 vs. 21 gaps we saw in Nightmare, a number of hardcore raiders entered Nighthold with 54 points, while others were just beginning that final progression and found themselves with nearly 10% less health and damage, equivalent to being almost a full tier of gear behind. Players who switched specs or characters along the way found themselves in a similar position. The power gap was larger than ever before, which created a sense of obligation and a number of negative social pressures that the system had previously tried to minimize. In short: We’re not at all happy with how this worked out.
A common suggestion is to simply reduce the amount of Artifact Power required to fully unlock the artifact in 7.2. This would not solve the underlying problem, but would rather reduce its duration while heightening its intensity, as competitive players sprinted to finish their Artifacts in order to be “ready.” But then we would inevitably tune around that completed power level, and other players would simply be playing catch-up the entire time. And in the long run, Artifact Power would not be serving its intended purpose of ongoing parallel progression. A capped-artifact player who goes a week without getting any item upgrades ends the week literally no stronger than before. Part of the value of the artifact, both for personal progression and guild progression, lies in ensuring that everyone is at least a bit stronger next week than they are right now, and a bit closer to overcoming whatever obstacle stands in their path. Our goal is for Artifact Power to always be of some interest as a reward, whether from a World Quest, or as a consolation prize when failing a bonus roll.
Instead, we are focusing on fixing the mistake of flat cost scaling at the end of the progression, and instead keeping the increases exponential throughout, while also strengthening Artifact Knowledge as a core pacing and catch-up mechanism. These changes should be visible in an upcoming PTR build.
This is done with the primary goal of reducing the power gap based on time investment, while preserving Artifact Power as an endgame reward that everyone values. If the leaders in Artifact Power were only a few points ahead of a more typical player, rather than crossing the finish line when most were just leaving the starting blocks, players with less time to commit would not be as disadvantaged in competitive activities. If a Warlock were choosing between having 48 points in a single spec or 44 points in all three specs if they’d split their efforts evenly, the barrier to playing multiple specs would be significantly reduced. We are still tuning the curve for 7.2 trait costs, but we’re currently targeting scaling such that someone who earns twice as much AP as me will have an artifact that’s only ~1.5% stronger; someone who earns four times as much AP as me should only be 3% stronger. On the whole, this should be a massive reduction in the power gaps we see in the live game today.
The second problem with our initial implementation was that repeatable sources of Artifact Power (Mythic Keystone dungeons in particular) dominated time-limited sources such as Emissary caches and raid bosses. The fact that a large portion of the community evaluates their Artifact Power needs using “Maw runs” as the unit of measurement is ample evidence of this failure. We very recently deployed a hotfix to increase AP earned from Nighthold in order to make raiding, with a weekly-lockout, better compare in efficiency to repeated Mythic Keystone runs. And in 7.2, we’re more thoroughly addressing this issue by adding a significant amount of AP to the weekly Mythic Keystone cache, while somewhat reducing (and normalizing based on instance length) the AP awarded by repeated runs. These changes are being made to narrow the gap in AP earning, and thus power, based on time investment.
All of the above changes are aimed at allowing players the freedom and flexibility to decide how they want to spend their time, and which goals they wish to pursue, while limiting the difference in power between players who arrive at different answers to those questions.