Clothes may make the man, but in Diablo 3 and Reaper of Souls, gear makes the Nephalem, and one of the most important affixes is the Wizard’s Arcane Power on Crit, or “APoC.” That affix does what it sounds like; instantly generating X amount of the Wizard’s resource each time a crit is proc‘ed. Of course it’s more complicated than that, with different skills carrying different (hidden) proc rates, which are adjusted to match the power and amount of proc chances for each skill. A free skill that hits lots of enemies per cast has to have a low proc coefficient or else it would instantly refill via APoC. Expensive skills that only hit a few enemies can have a higher proc chance. And so on.
That skill-balancing complexity was largely worked out over the course of Diablo 3, but it’s become an issue again in Reaper of Souls thanks to numerous game changes and new types of Wizard gear. Wyatt Cheng provided some insight into the mathematical and other issues the developers are grappling with in a series of posts on Reddit, and you can read them below:
At the moment the default AP regen is 10/s. A standard AP-On-Crit loadout increases this easily to 22 or more and it was definitely heading into the territory of spamming expensive spenders non-stop right out of the gate. Reaching that point should be a state you aspire to and work towards (usually with some legendary items or a sweet set bonus) – not something that’s handed out virtually for free.
It completely negates the use of Signature spells, and we do strive for a rhythm of alternating between your left and right mouse buttons, as well as all 6 skills on your bar. It’s okay if the most dedicated players reach a point where they no longer need a signature spell, but this should not be the case for people who just put on 2 easily-acquired items and a particular skill loadout.
Yes – we are pushing for people to do more than spam one skill. Running out of resource should be a very real concern – that’s the point of having a resource system at all. You decide when to spend your resource. Sometimes you save it up, sometimes you spend it liberally. If you never run out of resource, what is the difference between a cheap spell on an expensive spell? In theory, a high resource cost spell should do more damage but you can’t use it as much. A low resource cost spell you can use more often but does less damage. What we were seeing on both the Witch Doctor and the Wizard is that people simply spam their most expensive spell possible and hope the monsters are dead before you run out of resource.
The Barbarian had reduction to resource regeneration from passives as well. We’ll continue to monitor the Witch Doctor, the Wizard and the Barbarian. We’re concerned both with making sure all 6 classes have good power levels relative to one another, and we’re also concerned that all 6 classes have a variety of fun playstyles available to them.
Quick follow-up for clarity. We intentionally did not modify any legacy items. We have a core value on the team to take great lengths not to modify items on the live game that people own. The reduction of AP-on-Crit applies to items dropped in Loot 2.0 only.
Click through for followups, where Wyatt answers questions about whether some legacy Wizard gear with higher APoC values might be better than new-found Loot 2.0 gear with lower ApoC values, and how the devs have worked to keep affixes from becoming too complicated in function or description.
For Hats you generally have 4 stats you care about:
So the choice comes down to:
vs Loot 2.0
I admit that a lot of the new Legendary effects are tantalizing, but gaining 2.5x more AP on crit and the ability to use your most damaging skills non-stop is hard to pass up. I’d honestly prefer a lifesteal-like reduction at level 70 just so I don’t feel compelled to use legacy items (although I’m sure this isn’t going to be too popular an opinion with some people).
It’s okay if the most dedicated players reach a point where they no longer need a signature spell, but this should not be the case for people who just put on 2 easily-acquired items and a particular skill loadout.
In that case, have you considered making the rarer legendaries roll higher AP-on-crit values, approaching the legacy values? Or adding AP-on-crit to non-Source/Hat legendaries? Is there a Wizard set (torment-only?) that has AP-on-crit as its powerful set bonus?
Yes, you bring up a great point.
We discussed this internally and to be honest, we’re “worried” that legacy AP-on-crit sources/hats might still be preferable. I put “worried” in quotes because we ran some analysis and looked at the new legendaries are we’re fairly confident people will find Loot 2.0 items to be better but we’ve been wrong before and I’m willing to be wrong again.
As you mention – the new Legendary effects is one of the things we’re banking on. We’re hoping they are tantalizing enough to pull people forward.
We considered a lot of things. We considered a Lifesteal-like reduction going into 70. We considered nerfing the live items directly (but as previously mentioned, we’re very loathe to do that). We considered significantly increasing AP costs on the most egregious spells. We considered removing AP-on-crit completely from orbs but significantly adding +flat damage so the likelihood people “graduated” out of them was extremely high (but this would cause a significant ripple effect of balance across all items in the game and all classes). We considered massive nerfs to proc scalars on the Wizard but that seemed unfairly across-the-board punishing just to fix the AP problem.
It’s really hard to say for sure how things will play out, but it’s an issue we’ll be monitoring closely.
Regarding AP-on-crit for rarer legendaries. I think part of this nerf definitely leaves room for Wizard-set bonuses in the future. Maybe it will be AP-on-crit as a set bonus – though more likely we’ll use some of the space we’ve freed up to do something more interesting like getting a freecast of Arcane Orb when you meet certain criteria or letting one of the cheaper spells become a generator in a manner similar to the Demon Hunter’s Kridershot.
Wyatt – Should have bit the bullet a year ago in RoS development when you toyed with stat caps and converted to a Ratings System a-la WoW. APonCrit isn’t quite %-age based, but could have been done. Just raise the rating required for 10 AP @ 70 higher. Problems across all legacy itemization obsoletetion solved! (No need to block legacy rerolling either)
ie – 10 AP on Crit converted exactly to 100 AP-on-Crit Rating.
@ 60 need 10 rating per 1 AP on Crit.
@ 70 need 25 rating per 1 AP on Crit.
So your legacy item with 100 AP-on-Crit-Rating still returns 10 @ level 60…. but only 4 @ 70. Item stays the same, new ilvl 70 items also have the same 100 rating cap, but with shiny new legendary affixes.
Was there a reason the Dev team did not want to convert to that type of system? It worked for Blizzard across many many expansions, and I know D3 folks really hate hearing about WoW… but c’mon.. Thunderfury? XD
We’re trying to avoid ratings for the forseeable future on Diablo III. Believe me – it’s tempting. Ratings have the advantage that you can continually scale up the stats by making them worth less per point when you move to higher levels. We could have done crit rating, crit damage rating, move speed rating, ApocRating, etc.
The downside of ratings is that it becomes extremely difficult to gauge your gear. Gear is already hard enough to evaluate – but we always have it in the back of our minds to try and make things better over time, not worse. Going with ratings is basically saying “No human being could hope to evaluate this gear, just go to d3up” (not that I have anything against d3up, but I think the game should be playable by savvy players without needing it). This is why the new properties we’ve added are things like cooldown reduction, resource cost reduction, and area damage. We’re trying to provide stats that people can evaluate without needing a spreadsheet.
Internally we had people suggest things like “super-crit” or “crit-from-strength” as new stats to introduce in RoS. I’ll let you imagine what these do. The problem is – these stats added complexity without making choices more interesting -and worse – they further made evaluating gear into a giant math problem.
TL;DR: Ratings let us scale gear but make gear harder to evaluate at a glance. The design team has decided to accept a lot of pain in order to keep the gear legible to humans.
It’s ironic this is an issue for the Wizard, since as we’ve discussed several times on the podcast during the RoS beta; the Wizard is awesome now. Buffs all around have made many more skills and builds viable, and the general combat changes in RoS have decreased the importance (necessity) of using multiple defensive or crowd control skills. This gives the Wizard (and Demon Hunter) a lot more survivability and opens up many more builds and play styles. I just hope the APoC issue doesn’t derail that in some way for the Wizard, though it sounds like Wyatt’s got it under control.