[Restoration] - Are We Overselling Mastery?

[robodude666 2]

Hi All!

I'm starting to think we're overselling mastery as our #1 secondary stat.

I'm currently sitting at ~71% Mastery, however I've noticed that on average health stays at about ~70%. This means that Mastery is only providing a ~21% increase to healing.

If my mastery is dropped to ~60%, I'd be able to gain ~3% crit. My contribution from Mastery would be only ~18% but my gain from crit (not including multistrikes which can also crit) would be 6%... for a total of ~24% on average increase, plus 3% increase in mana regen from resurgence.

Looking at Ask Mr. Robot:

- BIS Mastery-build would provide you with ~96% mastery + ~10% crit.

- BIS Crit-build would provide you with ~63% mastery + ~22% crit.

How's that Mastery bonus better than the crit? Based on the above, the crit-build would theoretically double your bonus throughput (~20% --> 40%); not taking into account the Tidal Wave bonus to Healing Surge.

Is my math correct, or am I missing something?

-robodude666

[Karanir 15]

Comparing the bis stats, you would lose 33% mastery for 12% crit. Assuming an average health of 70%, that works out to a loss in mastery throughput of 9.9% to gain 12% crit.

Honestly, I don't have a detailed mathematical response or analysis for you. Only subjective personal opinion,I'm afraid. Personally, I would still choose mastery over crit. I can share the reasons. Firstly, This is because of the strength and reliability of mastery in low health situations. You can't guarantee a crit when you need it but mastery always works . secondly, our throughout cool downs like HTT are usually used at low health, and the mastery would benefit them a lot more than crit. Thirdly, crit has a much larger chance of overhealing compared to mastery especially for spells like healing wave.

On a related note, can someone share whether it's possible to pull out the data from Warcraft logs as an excel table? Or how can we quantify the effects of mastery in our healing?

[Stoove 347]

I spent a lot of time looking at the relative benefits of different stats during the MoP -> WoD transition. I definitely have ideas on this topic, and if you want to know where I'm getting my numbers from you can find it well documented here. What it boils down to is this; are the relative stat weights for Crit and Mastery reflective of what we care about?

 

Well, that very strongly depends upon the target's health. Most stat weights for Mastery (certainly the recommendations which we make in the IV RSham guide) are based on the assumption that the target health is 50%. At this level, Mastery outweights Crit by a long way.

 

However, your question is looking at the numbers for when target health is at 70%. What happens then? I can run the numbers with my little program and tell you that at 70% health, Mastery is worse than Crit by about 7% for a general rotation in terms of raw throughput. So yes, you are technically accurate when you say that if your target is at 70% health you'd be better off having Crit.

 

However

 

If a target is at 70% health, do you really care about getting 7% more out of your stats? I would say that in that case I don't, because in that range my Druid friend is going to stick a Rejuv on him and I can move onto a higher priority target. My concern is for those people at 50% health, which I tend to see often. In that case, I care very much about getting extra healing out (and Mastery is massively better).

 

Where do they cross? At about the point where I start caring about the target's HP; roughly 65%. That's the real difficulty with the discussion around Mastery - you have to make a value judgement about when you want to start caring about your overall throughput.

 

But then again, if you're worrying about your throughput at 70%, why are you stacking Crit? Mastery Multistrike [typo] is significantly stronger than Crit at all item levels and is free and has a slightly lower chance of overhealing. To me, that would make it a more attractive stat. To still further question; if you're that worried about your throughput near 70% HP, why not go for Haste? Its throughput is close to that of Mastery even when the target's HP is at 50% and at 70% it massively outshines all other stats. Let's face it, if your raid is at 70% for most healing then you can afford to adapt your playstyle around Haste to try to snipe the HoTs.

 

So that's my perspective. You may disagree there, but my value judgement (and that of other theorycrafters) is that we care much more about middle-low HP players than we do about high HP ones, and until this point the WoD healing model has broadly supported that.  :-)

[Cheze 6]

There's an addon called restoshamanstats that tracks the health of heal targets; unfortunately it only represents this as an average percentage (i.e. average health of your heal targets over the course of an encounter was 70%), so it isn't the most effective measure.  It also presents a valuation of your mastery relative to other stats based on that, so you can compare from attempt to attempt how useful it was.  You could also look at the health of the raid over time on warcraftlogs and determine how effective your mastery is at a given point in an encounter (although imo this doesn't really add a lot of info that you wouldn't already learn just by seeing the fight a few times.)

 

I don't think it's worth prioritizing crit over mastery for the reason stoove lays out: at the health% where crit actually becomes better, its additional throughput is likely to generate a lot of overheal anyway.  This is true to a lesser extent with MS.

 

I do think it's worth prioritizing haste over mastery though, especially in larger raid situations.  It's throughput is quite competitive with mastery even in scenarios where mastery is strong, and helps us deal with the raid situation we struggle with most (i.e. spread out/movement) by providing big buffs for HST/HTT/riptide.

 

In terms of the broader meta-argument, I've never been entirely comfortable with the idea that we just shouldn't care that much about healing targets that are at relatively high health.  There are lots of situations where the raid is being knocked from 100-70% (or 110-80% with a filthy disc priest present), and healing that deficit quickly/efficiently has value since there are usually more hits coming.  We could maximize (theoretical) efficiency by not healing anybody until they reached ~50% health, but even if we had a hypothetical raid healed by 3+ resto shamans that actually committed to that model it probably wouldn't produce good results in terms of keeping the raid alive (everything we continuously hear about a 'triage' model nonwithstanding.)

[Pandacho 286]

In terms of the broader meta-argument, I've never been entirely comfortable with the idea that we just shouldn't care that much about healing targets that are at relatively high health.  There are lots of situations where the raid is being knocked from 100-70% (or 110-80% with a filthy disc priest present), and healing that deficit quickly/efficiently has value since there are usually more hits coming.  We could maximize (theoretical) efficiency by not healing anybody until they reached ~50% health, but even if we had a hypothetical raid healed by 3+ resto shamans that actually committed to that model it probably wouldn't produce good results in terms of keeping the raid alive (everything we continuously hear about a 'triage' model nonwithstanding.)

To give an example of what happens in the raid where's too many HoT healers. My healing team currently consists of 3 resto druids, resto shaman (myself) and disc priest. In this situation (Normal and start of HC BRF) to wait for the raid to hit 50% of health to max the efficiency of my mastery... well, I can wait for it until the end of the fight. 

I think that in this situation (all the healers except of the shaman are druids and discs) we should try full haste build with AS and CBT in talents. I tried to play around with all the combinations on Beastlord HC tries: CBT vs HT, AS vs EotE, with glyph of Riptide and w/o, and have to say that CBT gives the most steady output with minimum overheal. HT can roughly compete in terms of throughput, but drains your mana insane, forcing to use CH too much. It may be still a good idea for the short fights like Gruul, but in the long ones and w/o set bonuses I would try Haste and CBT. 

[Stoove 347]

Updated a typo in my original post, so the penultimate paragraph makes a lot more sense:

 

But then again, if you're worrying about your throughput at 70%, why are you stacking Crit? Mastery Multistrike [typo] is significantly stronger than Crit at all item levels and is free and has a slightly lower chance of overhealing.

 

 

In terms of the broader meta-argument, I've never been entirely comfortable with the idea that we just shouldn't care that much about healing targets that are at relatively high health.  There are lots of situations where the raid is being knocked from 100-70% (or 110-80% with a filthy disc priest present), and healing that deficit quickly/efficiently has value since there are usually more hits coming.  We could maximize (theoretical) efficiency by not healing anybody until they reached ~50% health, but even if we had a hypothetical raid healed by 3+ resto shamans that actually committed to that model it probably wouldn't produce good results in terms of keeping the raid alive (everything we continuously hear about a 'triage' model nonwithstanding.)

 

I think that the final part of that quote (bold is mine for emphasis) is the real discussion point here. When a lot of the theorycraft around this was being done, we had only initial estimates of the healing model and a lot of hype from Blizzard about triage healing. The most relevant question to ask now is whether we are actually in the type of situation where not immediately topping someone from 70% will get them killed. That's perhaps a better way of phrasing "at 70% HP, do we really care about our throughput?".

 

Up until this point, the answer has been broadly "no". I can point to a lot of Highmaul fights and show you situations where players in my raid were at 50% or lower for an extended period of time and survived, but I'm starting to think that that's no longer the case in BRF. The more I play of it, the more it looks like we're drifting back towards the "a player not topped is a player about to die" mentality.

 

Perhaps we're not quite in that position yet, but I am really starting to question whether next tier we will be looking back at Highmaul and saying "Hey, remember triage healing? That was fun." And at that point, Mastery is going to be a very niche stat indeed.

 

So, right now I think that Mastery is still just about the best, but the healing environment feels like it's drifting rapidly towards the point where Mastery will stop being de facto the best stat.

[Cheze 6]

Blizzard are on their third expansion running talking about how they're going to implement a 'triage' model and how awesome it's going to be; at this point I'm pretty cynical about those statements.

 

The bottom dollar is that triage healing is easy; every healer knows what their most efficient heals are and regen is almost entirely passive.  So for a triage model to work, incoming damage needs to be either infrequent or relatively low and predictable.  It's hard to design difficult encounters that meet that criteria though, so every expansion we wind up back at the 'top off or die' status quo.

 

The thing I like about their more recent encounter design is that damage is a lot more granular; rather than 'avoid or die' or 'interrupt or die' the raid is rewarded for doing mechanics better by taking less damage (ex: aoes that do damage based on how far away you are.)  But that creates strong incentive to top people off quickly too.

[robodude666 2]

Sorry for not responding for so long! I completely forgot about the thread and received no email notifications .

 

If a target is at 70% health, do you really care about getting 7% more out of your stats? I would say that in that case I don't, because in that range my Druid friend is going to stick a Rejuv on him and I can move onto a higher priority target. My concern is for those people at 50% health, which I tend to see often. In that case, I care very much about getting extra healing out (and Mastery is massively better).

 

Where do they cross? At about the point where I start caring about the target's HP; roughly 65%. That's the real difficulty with the discussion around Mastery - you have to make a value judgement about when you want to start caring about your overall throughput.

 

But then again, if you're worrying about your throughput at 70%, why are you stacking Crit? Mastery Multistrike [typo] is significantly stronger than Crit at all item levels and is free and has a slightly lower chance of overhealing. To me, that would make it a more attractive stat. To still further question; if you're that worried about your throughput near 70% HP, why not go for Haste? Its throughput is close to that of Mastery even when the target's HP is at 50% and at 70% it massively outshines all other stats. Let's face it, if your raid is at 70% for most healing then you can afford to adapt your playstyle around Haste to try to snipe the HoTs.

 

So that's my perspective. You may disagree there, but my value judgement (and that of other theorycrafters) is that we care much more about middle-low HP players than we do about high HP ones, and until this point the WoD healing model has broadly supported that.  :-)

I came up with the ~70% figure as an example. Like Cheze, I too use RestoShamanStats. One of the metrics it records is the average health of a player throughout a fight. While it is true that we'll be more useful to the raid when people are low, if the average fight is around ~70% should we really be gearing ourselves for that rare case everyone is sitting at 10%? Whether we stack mastery or not, I haven't experienced any issues popping Healing Tide Totem and Ascendance and bringing the entire raid to full within seconds. Instead, I've found my role to mostly be keeping people topped off as much as possible in case a mechanic aren't followed exactly right or in general preparing for large bursts of raid damage.

While I do like Multistrike, Crit seems to be a more natural alternative as it'll also benefit our Mana via Resurgence. You can get a very nice benefit from Multistrike in this regard, as Multistrikes can crit and cause up to two Resurgences. However, you need a healthy amount of crit for this to be possible (around ~20-25% from spreadsheets I made). With AG copying Multistrikes in 6.1 it'll certainly be more appealing.

Likewise, Haste is also an attractive stat to me, as about ~60-70% of my healing usually comes from HoT-based heals (HST, HTT, HR, RT, etc.), however Crit is still more appealing due to the mana return it can provide -- that is, until Mana no longer becomes an issue. 

 

I don't think it's worth prioritizing crit over mastery for the reason stoove lays out: at the health% where crit actually becomes better, its additional throughput is likely to generate a lot of overheal anyway.  This is true to a lesser extent with MS.

 

I do think it's worth prioritizing haste over mastery though, especially in larger raid situations.  It's throughput is quite competitive with mastery even in scenarios where mastery is strong, and helps us deal with the raid situation we struggle with most (i.e. spread out/movement) by providing big buffs for HST/HTT/riptide.

 

In terms of the broader meta-argument, I've never been entirely comfortable with the idea that we just shouldn't care that much about healing targets that are at relatively high health.  There are lots of situations where the raid is being knocked from 100-70% (or 110-80% with a filthy disc priest present), and healing that deficit quickly/efficiently has value since there are usually more hits coming.  We could maximize (theoretical) efficiency by not healing anybody until they reached ~50% health, but even if we had a hypothetical raid healed by 3+ resto shamans that actually committed to that model it probably wouldn't produce good results in terms of keeping the raid alive (everything we continuously hear about a 'triage' model nonwithstanding.)

If you're raiding with the Methods and Paragons of the world, keeping people topped off to prevent accidental damage might be less of a priority. However, in less environmentally-aware or mechanically-challenged raids you'll come across people dying due to taking accidental damage followed by heavy swings of expected damage that require the raid to be topped off (e.g. inferno slice on Gruul if the person just stood in something else). In these situations, having some extra Crit/Haste will be of more use than Mastery as they will more than likely still be above ~70% HP but you need to bring them up quickly due to fast cast times of boss abilities.

 

 

That's perhaps a better way of phrasing "at 70% HP, do we really care about our throughput?".

 

Up until this point, the answer has been broadly "no". I can point to a lot of Highmaul fights and show you situations where players in my raid were at 50% or lower for an extended period of time and survived, but I'm starting to think that that's no longer the case in BRF. The more I play of it, the more it looks like we're drifting back towards the "a player not topped is a player about to die" mentality.

 

Perhaps we're not quite in that position yet, but I am really starting to question whether next tier we will be looking back at Highmaul and saying "Hey, remember triage healing? That was fun." And at that point, Mastery is going to be a very niche stat indeed.

 

So, right now I think that Mastery is still just about the best, but the healing environment feels like it's drifting rapidly towards the point where Mastery will stop being de facto the best stat.

Generally speaking, I would agree. We shouldn't really care about our throughput at 70%. However, if we're seeing people get to 70% frequently enough that if they're not healing appropriately they could reach more fatal levels -- then it becomes an issue. Unless I want to go mana dead and spam Healing Surge to get my ~1.2s cast times, the cast time of Healing Wave makes it too slow to respond to situations like this frequently. Chain Heal can help prevent the deaths of multiple people, but it's mana cost prohibits its use freely until Tier is obtained.