Beta Class Balance Analysis Part III

Blizzard is asking for feedback on class balance in the Mists of Pandaria Beta. This article continues our coverage of hunter-related discussions in the thread (see Part I here and Part II here).

– – –

Ghostcrawler wrote, “Barrage vs Powershot: We’re currently planning to reduce Barrage’s damage a bit, so that it ends up being on par with the others on the row in terms of overall DPS.” [link]

Jigsawe replied, “What? It’s already not worth using either single target or in an AoE situation. Also, hunter shots are currently partially blockable as are pet attacks. Pet attacks are also suffering from glancing blows. Is this intended?” [link]

Ghostcrawler replied, “All of the hunter level 90 talents are getting an across the board damage buff (they’re intended to be worth using single-target, and currently they’re not). I just meant that Barrage’s damage will be lowered a bit relative than Powershot. Hunter ranged attacks should not be blockable, and they currently are (known issue, will be fixed). Pet attacks should be blockable and able to glance. [link]

– – –

Gobuchul wrote, “A couple more questions from the SimulationCraft team: […] We’re also really wondering about the various “ignite-like” mechanics in the game (ignite, deep wounds, piercing shots, unholy blight, blackout kick, I may be forgetting others). In the past we’ve really struggled to properly simulate Ignite: A whole theory developed in the community with terms like “munching” and “rolling” being used to explain its inconsistent behavior, and we tried to simulate what the theory described as faithfully as possible, but at some point between the end of WotLK and now something seems to have changed and our simulation is off again, so we’re back to square one in trying to analyze its behavior. We’ve also faced similar issues with Deep Wounds in particular, which we had to treat as a separate problem because we couldn’t trust that the two mechanics worked the same way.

Specific questions: Do these mechanics have the same basic implementation internally? In other words, can we trust that there is a single set of assumptions we can make about how the damage calculation and propagation works that will allow us to simulate all these mechanics correctly? And are there any details you can give us that would reduce the number of assumptions we need to make?

Personally I suspect the inconsistent behavior we see from these abilities are at least partly a result of server-side processing delays, probably related to multithreading, in which case I guess there’s a good chance the resultant behavior is not quite deterministic, but any details you could share would still be helpful.

[link]

Ghostcrawler replied, “Ahhh, Ignite mechanics. What a storied history they have. :)”

While there have been several versions of Ignite and similar effects, historically, the general goal for quite a while now has been for this sort of effect to be simply a way to move part of the damage from an attack into a DoT, to reduce burst, and add flavor. And since these attacks were likely to happen again before the DoT finished, each one needed to “roll” the damage left over from the previous DoT into the new one. However, that damage rolling processing has always been somewhat delayed. That leads to the “munching” problem (two hits could land at nearly the same time, and the first one wouldn’t have been processed yet when the second one lands, leading to the first one’s DoT getting overwritten by the second one’s).

However, we got some code changes done in 5.0 that handles rolling periodics better. If Blackout Kick does 20% of its damage as a DoT, then at the end of the fight, Blackout Kick’s DoT should have done exactly 20% of the damage of Blackout Kick (barring any rounding errors or overkill of course). The actual logic involved is rather simple: When reapplying this DoT, take the damage left in the previous DoT and distribute it evenly across the new DoT’s ticks. Do remember that refreshing a DoT refreshes it to the full duration, plus the next tick (so that you can clip DoTs by 1 tick and have no loss).

For example, suppose Blackout Kick does 20% additional damage as a DoT that ticks every 2 sec, for 4sec.

0sec: Blackout Kick lands for 100 damage. DoT applied, with 4sec remaining, doing 10 damage per tick.
2sec: Blackout Kick DoT ticks for 10 damage. 2sec remaining, still 10 damage per tick.
3sec: Another Blackout Kick lands, and this time it crits, for 200 damage. The DoT is refreshed and extended to 4sec past the next tick; 5sec remaining, or 3 ticks. It needs to do a total of 10 damage still from the previous DoT, and 40 more for this one, for a total of 50 damage over 3 ticks, so ticks for ~17 damage, 3 more times.
4sec: Blackout Kick DoT ticks for 17 damage. 4sec remaining, still 17 damage per tick.
6sec: Blackout Kick DoT ticks for 17 damage. 2sec remaining, still 17 damage per tick.
8sec: Blackout Kick DoT ticks for 17 damage, and expires.

Hope that helps.

Spells that use this mechanic include: Ignite, Blackout Kick, Echo of Light, Devouring Plague, Touch of Karma, Hand of Gul’dan, Bloodbath, Temporal Shield, Piercing Shots, Explosive Shot, and several set bonuses.

[link]

Zeherah asked, “I’m not sure if I read your description of how this works properly, but does this mean Explosive Shot would always do the full damage regardless of whether you refresh when there is more than one tick left? Please say yes, I would love to never see an argument again over the proper timing of LnL shots.” [link]

Ghostcrawler replied, “One additional addendum on the Ignite things: Explosive Shot does not do this in your current build, but will in the next. The others should all work that way in your current build.” [link]