MoPping up Mastery, Part I

In the latest “The Light and How to Swing It,” Matt Walsh (a.k.a. Rhidach from Righteous Defense) made some pretty sharp criticisms of our mastery design in Cataclysm.  I think he’s got some valid points, but I don’t agree with all of them.

For example, he derides the stylistic choice of block chance, and calls it a “returned and renamed stat which was stamped as fit for deleting.”  I don’t think that’s entirely fair.  Certainly, the old block chance was unattractive in Wrath, but that was primarily because block itself was weak.  However, that’s not the fault of block chance.  If block had been very strong, block chance would have been very attractive.  This is exactly the situation we have in Cataclysm – a very strong block mechanic and a high valuation on mastery.

In that light, I think it’s erroneous to call it a stat that was “fit for deleting.”  Far from it, in fact.  The real reason block chance was removed was not weakness, but redundancy.  Since block chance was only useful to two of the four tanking specs, it was a natural choice for conversion into a mastery stat.  So block chance wasn’t removed, really.  It’s still with us, just under a different name.  And it’s actually our most prized stat.

I also disagree with him on the issue of style.  As paladin tanks, we’ve traditionally been the “block tank.”  Blocking is in our blood, and it’s thematically appropriate for the class to have a block-centric mastery.  Block chance seems like it was a reasonable choice on stylistic grounds, as paladins traditionally blocked more than warriors.

However, Rhidach’s post does hit upon an important point.  Block chance differs somewhat from most other stats, in that it can be capped and that cap is fairly readily achievable.  Crit and haste can be capped, but not passively or easily.  Armor can be capped, but again, not with obtainable gear.  Primary stats, health, attack and spell power are all unbounded.  And dodge and parry have diminishing returns (nonlinear scaling) that keep them from reaching an upper limit.  Block chance joins hit and expertise as the only three stats that have this unique combination of linear scaling and an upper bound.

And that’s what makes block mastery broken.  Not in the sense that it’s overpowered (though you could argue that it is, in some cases).  But in the sense that it’s incredibly hard to balance effectively.  It’s either so strong that you want to stack it to cap, or it’s weak enough that you ignore it entirely.  It’s difficult to find a middle ground, where you’d be just as happy with 1 mastery rating as you would with 1 dodge or parry rating.  And it’s especially difficult when only a subset of the tanking classes can do it.

Rhidach is also right that this problem has its roots in the combat system itself.  The single-roll system, where the result is between 0 and 100, creates the boundaries, and the linear scaling of mastery happily expands to fill them.  A bounded system can lead to problems simply because when those bounds are reached, funny things happen.  You can hit discontinuities, get accelerated scaling, and create other unintuitive behavior.

A familiar example to some might be avoidance.  In Burning Crusade, avoidance scaled linearly, so that every X rating gave 1% dodge or parry.  This meant that as you stacked more avoidance, every point of additional rating actually got better.  To see that, consider that adding 1% avoidance when you’re at 50% total avoidance (so, going from 50% to 51%) reduces your damage taken by 1/50 (=dx/(1-x), where x=0.5 is avoidance and dx=1% is the increase), or 2%.  Adding 1% avoidance when you’re at 75% (75%->76%) reduces your damage taken by 1/25, or 4%.  Doing the same at 90% avoidance (90%->91%) reduces your damage taken by 1/10, or 10%.  You can see why this could be a problem for balance – the tank that manages to squeeze out an extra bit of avoidance gets a huge benefit.  So even small discrepancies between classes could lead to large performance gaps.

As of patch 3.0, avoidance sidesteps this behavior with diminishing returns.  No matter how much avoidance you stack, you can’t get 100% of either dodge or parry individually.  And the formula is set up such that it would take a ridiculous amount of raiting to get even 50% of either one.  Even now, in nearly best-in-slot gear in the last patch of the expansion, we can only manage 20-25% of each.  The diminishing returns formula acts as an arbitrary limiter so that dodge and parry rating act like they’re an unbounded system, even though they’re doing so within a bounded structure.  It also makes sure that the impact of avoidance rating gets smaller the more you have of it, so that the 1/(1-x) behavior of Burning Crusade doesn’t happen.

The problem doesn’t generally occur with hit and expertise, despite being a bounded system.  In fact, quite a few DPS classes have a hit cap as their first priority, because it’s their strongest DPS gain.  But damage output works a lot differently from damage taken.  The DPS gains you get in reaching hit cap have a finite upper bound.  You’re increasing the success rate on your way to 100%.  With avoidance, you’re decreasing the boss’s success rate, which gives you the scaling I described above.  That’s why getting 1 more hit out of 100 isn’t game-breaking, but taking one fewer hit out of 100 can be.

So, you may be asking yourself, what can Blizzard do to fix the problem presented by block chance mastery? Or, more generally, what can Blizzard do to redesign our mastery for Mists of Pandaria?  In tomorrow’s post, I’ll be exploring a number of possible ways they can implement mastery for shield tanks and discuss the pros and cons of each.  We will, however, adhere to one constraint: all of the ideas have to involve blocking, somehow.  Sure, they could give us our own version of Savage Defense, but that wouldn’t fit the theme of a shield tank.  There’s no reason to homogenize tank mastery, so we’re going to avoid doing so (apart from keeping both warrior and paladin mastery bonuses shield-themed, of course).

So stay tuned for tomorrow’s installment!

This entry was posted in Tanking, Theck's Pounding Headaches, Theorycrafting and tagged , , , , . Bookmark the permalink.

One Response to MoPping up Mastery, Part I

  1. Pingback: MoPping up Mastery, Part II | Sacred Duty

Leave a Reply