Legion Talent Review

July 3, 2016

Hello Guardians!

For a long time now there hasn’t been many details to talk about that weren’t totally subject to change, or simply easier to talk about on TankCast. However we’ve entered the tuning phase of talents now, and since most of them are mechanically fine it makes sense to finally examine them and see where they are in terms of effectiveness. Obviously some talents are more of a feel or playstyle choice, but there are definitely still informed decisions to be made. So let’s get started.

Level 15 Talents

We have one talent with a very defined offensive and defensive benefit in Brambles, and two talents that generate more Rage which can be used for both offense and defense in Blood Frenzy and Bristling Fur. The quesiton is how good are they at each? Since Brambles is obvious, we need to first figure out how much damage Rage contributes. Since the only way to convert Rage into damage is with Maul, we just determine how much damage one Rage is worth. To do that we follow some simple steps to find the damage-per-Rage (DPR) of Maul:

  • First we define how to calculate the desired value: DPR = (Melee * Multiplier * Versatility * Artifact) * (1 + (Crit+0.1))*(1 – Armor DR)/Rage Cost
  • Then we remove terms that duplicate for Brambles: DPR = (Melee*Multiplier)*(1+(Crit+0.1))*(1 – Armor DR) / Rage Cost
  • Since we know that the biggest difference the Maul Crit trait will cause is 9.1%, we can replace it with an additional constant: DPR = (Melee*Multiplier)*1.091*(1 – Armor DR) / Rage Cost
  • Fill in known constants: DPR = (Weapon DPS + (AP/3.5))*1.5*1.091*0.68/20
  • Simplify: DPR = 0.055641(Weapon DPS + (0.28571*AP))

So it really comes down to how much weapon DPS you have which determines your total Maul damage. The overall contribution from AP is only ~1.5897% per Rage, but how does that stack up for each of the talents? Blood Frenzy is 0.66 RPS per target (or 1 RPS per 1.5s melee), and since Brambles also scales with the number of targets attacking you that is never going to be enough Rage income to overtake Brambles. However Bristling Fur scales with incoming damage, so generally it will end up generating more Rage over time except in sustained AoE situations. Is it enough? The easy answer to that question is unsurprisingly, math.

  • Determine Rage needed to equal Brambles AP coefficient: 0.01589719011X = 0.24 = 15
  • Determine Rage needed per Bristling Fur use: 40*15/1.5 = 400

That tells us the absolute most you would need is to take 400% of your HP in damage when Bristling Fur is active and then spend all of that Rage on Maul. Reality is probably somewhere closer to half that value, which is ~200% of your hp in damage (to get 200 Rage) which is close to 40% per melee swing, which we’ve been told is not going to happen. Therefore we can draw the conclusion that neither Bristling Fur nor Blood Frenzy will ever be equal to Brambles for pure damage. But what about survival?

Let’s start at the other end. Blood Frenzy gives one extra use of either Ironfur or Mark of Ursol every 67.5s per target. The easiest way to look at it is in terms of EHP which simplified, is (HP / (1-%DR)). Using that we can determine how much of an increase in EHP one MoU gives:

  • %EHP Increase = (HP / (1-%DR2)) / (HP / (1-%DR1))
  • Add known Magic DR: %EHP Increase = (HP / 0.567) / (HP / 0.81)
  • Simplify: %EHP Increase = 0.81/0.567 = 1.4285714285714285714285714285714 or 42.857% more EHP for ~10 seconds or so after traits and talents

Unfortunately Ironfur isn’t so easy. We know that any equal percentage increase in Armor generates the same constant amount of EHP, which means the percentage increase in EHP decreases as Armor increases. Instead we’ll have to think about it slightly in the abstract. Consider that the EHP increase from one Ironfur is some value Y. We know that Brambles is esentially a flat EHP bonus since it applies to every damaging event you receive. Therefore in the case of Blood Frenzy, we’re looking at 45 melee swings at 0.24AP each, for a total of 10.8AP. Let’s use a real-world example to see if they’re comparable using a Beta 110 PvP Premade which has 2,296,297 Health, 5572 Armor, and 17,705 AP.

  • Using the values above we know that the you will gain ~1,202,296 EHP with each Ironfur application.
  • Since we would be adding HP for Brambles, we remove the DR to find that we need a total of 714,549 to match Ironfur.
  • 10.8 * 17,705 = 191214 which means Brambles will never come close to Blood Frenzy, or subsequently Bristling Fur.

Now, these assumptions ignore that Ironfur would only last ~10 seconds. However I believe that limitation is handily countered by the fact you can apply the Ironfur whenever you want within those 67.5 seconds, and that EHP has consistently been demonstrated to only valuable when you need it. We also know that HP scales a lot faster than AP making it impossible for Brambles to catch up from starting values this low. Thus while Brambles is clearly far and away the best DPS talent of the three, both Blood Frenzy and Bristling Fur handily defeat Brambles for pure survival.

On the topic of Bristling Fur vs Blood Frenzy, I think Blood Frenzy is a little bit too weak in comparison. There are almost never situations where 3-4 targets are active for long durations. Usually you only end up with 1, or possibly 2. Sometimes you have a large number of targets for 5-10 seconds, but that’s not enough to justify taking Frenzy for difficult content. There should still be a gap, but maybe one not quite so large. I’d suggest nerfing it to 0.75-0.8 Rage per % HP in damage taken.

Level 30 Talents

The only change for movement talents is that Guttural Roars replaces Feline Swiftness (which is now part of Feral Affinity). However there’s a new, additional problem for Displacer Beast. Legion introduces a new “shapeshifting cooldown” which is incurred every time you change forms. That means every time you use Displacer Beast you are prevented from using any damaging abilities for 3 seconds. You can still use cooldowns and trinkets, but that’s not much consolation. Unfortunately this change means that use of Displacer Beast is going to be further restricted to only those instances where a blink is absolutely necessary. Unfortunately I think this punishes Guardians the most out of everyone.

Guttural Roars will end up being the default talent for most raid scenarios since Stampeding Roar can function as both a personal and raidwide movement increase. Wild Charge is clearly great for solo, 5 person, and pvp content. I think Displacer Beast will be the odd one out for this set of talents.

Level 45 Talents

Here we have the new quasi-support “affinity” talents. We’ll set aside Restoration and the passive effects for now, and just look at the pure DPS output of Balance and Feral. To start with, Feral Affinity grants a 35% bonus to energy regenration which is multiplicitive with Haste. However we’ll start with a base value of 135 Energy per second (EPS) for now. Since we don’t have to worry about Primal Fury it’s easy to approximate ability usage.

In order to “accurately” (at least within a reasonable measure of error) we’ll define a cycle as 30 seconds, since that works for both Feral and Balance fairly well. Within that window for Feral we know there will be 1 Thrash, 1 Rip, and some number of Shreds. We also know that we’ll start with 100 Energy and gain 10*1.35*(1+Haste) energy every second. Combine all of that information and we can come up with a formula to approximate Feral Affinity DPS.

  • DPS = ((Melee + Shred)*Armor) + Rip + Thrash
  • Add unique variables: DPS = ((1+Haste)*(Weapon Damage))*0.68+(# Shreds*13.74/30*(Weapon Damage/2)*0.68+(AP*0.6186/30)+(AP*2.19/30)+(80*0.14*AP/30)
  • Determine Shreds: DPS = ((1+Haste)*(Weapon Damage))*0.68+(((20+(30*10*1.35*(1+Haste)))/40)*13.74/30*(Weapon Damage/2)*0.68+(AP*0.6186/30)+(AP*2.19/30)+(80*0.14*AP/30)
  • Simplify: DPS = ((1+Haste)*(Weapon Damage))*0.68+(((20+(405*(1+Haste)))/40)*0.31144*(Weapon Damage/2))+(AP*0.46695333333333333333333333333333)
  • Fill in Sample PvP Data: DPS = ((1+0.1033)*(12750))*0.68+(((20+(405*(1+0.1033)))/40)*0.31144*(6375))+(18450*0.46695333333333333333333333333333)
  • DPS = 9565.611+23171.7795+8615.288 = 41352.6785

It’s worth noting that I’ve excluded all multipliers which are common between both affinities. This includes Versatility, Crit, Mastery, and Artifact Traits. Haste is included because it impacts each of the specs very differently. Additionally I accounted for Cat Form’s doubled AA damage by dividing the Shred multiplier by 2, instead of doubling the AA damage. So, let’s move on to Balance.

Surpriingly Balance is much easier to calculate. All of the spells are purely based on AP, and the only ability that has a varying number of casts in a 30 second window is Solar Wrath. Everything else is fixed. There will be 3 Starsurges, which also means 3 empowered Lunar Strikes and 3 empowered Solar Wraths. There will also be 2 Moonfires and 3 Sunfires. All remaining time is spent on regular Solar Wraths, since its AP/Cast Time value is higher than a regular Lunar Strike. We then combine all of that information to again, come up with a formula to approximate Balance Affinity DPS.

  • DPS = (3*SS+3*ELS+3*ESW+2*MF+3*SF+((30-Total Cast Time)/(1.5/(1+Haste))*SW))/30*NI
  • Add unique variables: DPS = ((3*1.1*4.5*AP)+(1.2*3*3.5*AP)+(1.2*2.2*3*AP)+(2*(0.45+(4*(1+Haste)))*AP)+(3*(0.9+(3*(1+Haste)))*AP)+(((30-(3*2/(1+Haste))-(3*2.5/(1+Haste))-(8*1.5/(1+Haste)))/(1.5/(1+Haste)))*2.2*AP))/30*1.2
  • Simplify: DPS = (35.37AP+(2*(0.45+(4*(1+Haste)))*AP)+(3*(0.9+(3*(1+Haste)))*AP)+(((30-(3*2/(1+Haste))-(3*2.5/(1+Haste))-(8*1.5/(1+Haste)))/(1.5/(1+Haste)))*2.2*AP))/30*1.2
  • Add sample data: DPS = (35.37*(18500)+(2*(0.45+(4*(1+0.1033)))*18500)+(3*(0.9+(3*(1+0.1033)))*18500)+(((30-(3*2/(1+0.1033))-(3*2.5/(1+0.1033))-(8*1.5/(1+0.1033)))/(1.5/(1+0.1033)))*2.2*18500))/30*1.2
  • DPS = (654345 + 179938.4 + 233649.45 + 206186.2)/30*1.2 = 50964.762

So we can clearly see that Balance pulls ahead of Feral pretty substantially. Interestingly enough it appears to be purely because of Nurturing Instinct granting an extra 20% spellpower. If that was reduced back to being equal, Balance and Feral would end up being pretty close together. However I don’t know what the consequences would be for the intended target of Resto, but that would be easy to solve by having Resto grant a healing bonus of some kind. But one remaining question is how does this compare to Guardian DPS?

  • DPS = (4WD*1.2*Mangles*1.15*Armor) + (2.4AP*Thrashes + 0.15AP)*1.15 + 1.2*(0.45AP*Moonfires + ((0.5/2)*Haste)) + (1-Mangles-Thrashes-Moonfires)*2WD*1.2*Armor + Melee*Armor + (Maul*WD*Armor)/3
  • Insert formulae and simplify: DPS = 1.9332857504WD*Haste + (0.4AP*Haste+0.15AP)*1.15 + 0.43AP*Haste + 1.632WD*(1-0.20889*Haste – 0.1666666666666667*Haste – (0.125/Haste))
  • Add sample data: DPS = 1.9332857504*(15937.5)*1.1033 + (0.4*18450*1.1033+0.15*18450)*1.15 + 0.43*18450*1.1033 + 1.632*15937.5*(1-0.20889*1.1033-0.1666667*1.1033 – (0.125/1.1033))
  • DPS = 33994.5945591351 + 12546.3321 + 8753.03055 + 13469.20634270108 = 68763.16355183618

So obviously Guardian DPS (on a single target) approximation beats the pants off of either Feral or Balance Affinity DPS. That’s clearly not the intention because you have no reason to ever shift except for resto, because there’s no benefit. Plus this doesn’t talents or cooldowns. Expect significant buffing of both in order to make them remotely viable beyond passives.

As for, Restoration, it ends up being the only affinity with a distinct survival benefit. This also means it will end up likely being the default selection for most people. I doubt many people will end up using Feral or Balance for actual DPS outside of Mythic raiding or possibly PvP, since those are generally the situations where maximizing DPS is of reasonably high importance. That’s fine, since it harkens back to the Bearcat era anyway.

Level 60 Talents

Here we have the new and consolidated crowd control talents. None of these have changed in function, which means Mass Entanglement is still effectively useless. You’re left with a choice between Mighty Bash and Typhoon, so you’ll end up picking whichever one is most appropriate for the situation. It’s quite a shame that Mighty Bash, out of all of the possible talented stuns for tanks, is by far the worst. Hopefully some tuning ends up being done. Maybe the Guardian version of Bash should be a cleave? I don’t know, but certainly something.

Level 75 Talents

These talents are all about active Rage generation, or generating additional Rage from rotational abilities. At least on paper anyway, in reality you end up with 3 talents that have varying levels of Rage generation and DPS benefit – which is what makes a good set of talents anyway. In any case Soul of the Forest is a good baseline to start with, so let’s look at its total Rage generation. Thankfully the formula remains (almost) unchanged from Warlords, so it’s pretty easy to evaluate.

  • RPS = ManglesPerSecond*Rage
  • Insert formula: RPS = 1/(1/((0.2*(1/(GCD*2))+(0.8*0.2)*(1/(GCD*3))+(0.8-(0.8)*0.2)*(1/(GCD*4)))))*5
  • Convert to Haste: RPS = 1/(1/((0.2*(1/((1.5/H)*2))+(0.8*0.2)*(1/((1.5/H)*3))+(0.8-(0.8)*0.2)*(1/((1.5/H)*4)))))*5
  • Simplify (<3 Wolframalpha): RPS = 1.044445H
  • Sample PvP Data: RPS = 1.044445 * 1.1143 = 1.1638250635

The reason I converted it to a simple equation based off of Haste (or rather, Wolframpalpha did) is because Galactic Guardian also scales from Haste, and I needed a common way to compare the two. In any case, the result is a reasonably simple and accurate equation for determining the tangible value of Soul of the Forest. If we’re to compare it to Galactic Guardian, we need to come up with a way of calculating the number of procs per second.

  • RPS = (Melee+GCD+Thrash DoT)*15
  • Convert to Time: RPS = ((0.1*(1/(2.5/H)))+(0.1*(1/(1.5/H)))+(0.1*1/3))*15
  • Simplify (<3 Wolframalpha): RPS = 1.600005*(H+0.3125)
  • Sample PvP Data: RPS = 1.600005*(1.1143+0.3125) = 2.2828

We can see from the start that Galactic Guardian is nearly twice as good as Soul of the Forest before even considering the fact that GG also scales with the number of targets since it triggers per damaging event. That’s a little unfortunate. GG should be better, but not that much better. But what about Incarnation?

We know that for Rage generation purposes all Incarnation does is remove the CD from Mangle. However that also means you don’t generate any Gore procs because you aren’t pressing buttons other than Mangle (in terms of the best case survival scenario). So at best we’re talking about ((30/(1.5/H))*6)/180 or 2/3H RPS. That’s…..pretty bad in comparison to either of the other two talents. However Incarnation definitely has a redeeming quality in that it allows us to do huge burst DPS for cleave or AoE situations. That fact alone means it will likely see huge amounts of play in Mythic+ dungeon content.

Level 90 Talents

The gut reaction to this set of talents is going to be “why on earth would I ever use Earthwarden“? Well we know that Earthwarden is bad for any situation where incoming damage is split between more than one source, because each charge cannot apply to more than one damaging event whereas GoE would apply to all of them within the 2 second extension. So what we need to is compare the 2 second extension on some number of Ironfurs (since if you’re extending Mark of Ursol then GoE automatically wins) to one charge of Earthwarden.

In a 2 second window at 1.5s attack speed we’ll assume that for the best case scenario Ironfur will affect two attacks. Which means GoE has an equivalent of a 15% EH increase spread over two attacks. We know from using our sample data earlier that we’ll end up with an extra 2 seconds of ~1.2mil EHP, while GoE would grant ~400k for the same period of time. However if you’re able to also have one Ironfur up at the same time (which isn’t that unreasonable with full talents) that value jumps to ~675k which turns out slightly better than the additional 2 seconds of Ironfur, and that doesn’t include the starting 3 charges or the fact you will generate them faster than you would a new Ironfur. What does it all mean?

Take GoE for tanking more than one target, dual wielding targets, magic damage, or where self-healing is important. Earthwarden will probably perform better for you against a standard 1.5s (or slower) swing boss. Survival of the Fittest is merely for more cooldowns, similar to where you would have taken the old Bristling Fur. However you might find that it also performs very well in Mythic+ 5 person content.

Level 100 Talents

So that leaves us with the last talent tier, which is really easy to evaluate. We know that Lunar Beam heals for 24*AP total, however that value is also multiplied by Versatility, Crit, and our Mastery bonus. Using the sample data, we can easily generate a total heal of ~691,465. Unfortunately that only happens every 90 seconds. Healing for ~30% of your HP is no small thing, however consider Rend and Tear over the same period. For R&T to equal that, you only have to take (100(691,465))/6 or 11,524,420 over 90 seconds which is ~128,049 DTPS. That’s a pretty trivial number to reach in any kind of content. In all honesty I’d probably cut the cooldown of Lunar Beam in half for it to be a midpoint between R&T and what Pulverize will end up sitting at.

Speaking of Pulverize, it’s a complete waste of a talent at 8%. It needs to be at least 12% for anyone to consider taking it over R&T.


  • Mata says:

    Thanks for all your hard work on this! :)

  • Atnema says:

    My hopes were high for Affinity for a little stance dance, but deep inside I expected this outcome -.-

    But still thx for all your hard work. It helps a lot!

    • Arielle Arielle says:

      I wouldn’t be surprised if damage is untuned still. We’re still a pretty far way off.

  • Atreyou says:

    Hey and thanks for all your work.
    I am wondering what the best stats for guardian druids are.
    I guess mastery is the best but what is with the other stats?

    • Arielle Arielle says:

      Mastery, then probably Haste/Versa followed by Crit because Dodge is unreliable.

      However I haven’t really looked in detail at secondary stats yet as I was waiting until closer to launch (since pre-patch doesn’t matter).

  • hostik says:

    it’s a shame about Pulverize, I really like that button rotationally, keeps a “skill ceiling” a little bit higher. I hope they make it better so it’s not a dead talent.

    • Arielle Arielle says:

      Yeah, the nerf was needed back when it was on the L15 row. It just hasn’t been re-buffed after the talent swap (which was a good swap FWIW).

  • Kioga says:

    Have you had a chance to play around with the Legendary items? I’m guessing we would prioritize the “leather” items over the neck, simply because they have armor and both primary stats. If that’s true, I would also wager legs are the best because of the larger stat budget given to their slot. It’s also the only item that benefits both offense and defense when using the R&T talent. Wrists are completely offensive in nature whereas boots and neck are defensive in nature.

    • Arielle Arielle says:

      They were available for a short period during Alpha, but haven’t been brought back since. I know they’re still undergoing tuning since some of them (most notably the one that increases Thrash stacks, and the one that increases FR healing) are incredibly overpowered compared to the rest. We’ll see where they end up.

  • Kioga says:

    What are your thoughts on the on the tanking benefits of the other two affinities? Despite giving us a heal, Resto wasn’t even on my radar compared to a 15% speed increase or a 5yard attack range increase. I also read that guardians receive a 50% penalty from the heal while everyone else get’s its full effect.

    • Arielle Arielle says:

      The 50% effectiveness is probably still in there from Warlords.

      As for Feral or Balance tanking benefits, I agree that the 15% movement speed is a nice perk – however no encounter will ever be designed in a manner that will “require” it. Simply because Deathknights exist. The extra 5 yards is meaningless for tanking unless you’re kiting something – which case it’ll be really awesome.

  • Kioga says:

    Wohead is reporting “Blood Claws” as our artifacts hidden ability but I cannot find anyone that knows how to unlock it. Have you came across it in your testing?


    • Arielle Arielle says:

      No. I generally don’t play much alpha/beta beyond figuring out how abilities and talents work mechanically, unless there are max level premades to use. This time around the only premades are pvp (sads) and the servers have been notoriously unstable which makes leveling feel unproductive (plus you spoil the story for yourself).

      Based on the description it sounds like the single best trait in the entire artifact though.

  • Kioga says:

    NVM. Got a response from the guide’s author: “Yeah that was one ability I wasn’t quite sure about. I’ll remove it, given it is no longer implemented. Thanks!”

    Agreed. That was the one thing I never liked about betas, spoiling all of the content with an unfinished game.

  • Palewind says:

    Great job, Arielle! I played around on the beta long enough to feel out which talents I liked and didn’t by feel, but always wait for your maths to help me make informed decisions! :)

  • Palewind says:

    Oh hey — one thought regarding Pulverize.

    Even though it feels disappointingly undertuned, I could see it still having a niche in progression AOE situations with casters where you just can’t round them all up, but want to benefit from the damage reduction. Mythic Hellfire Assault progression — when my standard AOE rotation would outstrip my warrior tanking partner’s threat — comes to mind. I can run around Thrashing and Swiping everything to establish threat, but can’t hit them all with 3 full stacks of Thrash in all the chaos to establish the -6% damage benefit. It feels like in some incredibly niche situations like this, Pulverize would be a good option, even as-is.

  • Semor says:

    Thank you for your in depth analysis. In regards to stats, it seems that Mastery>Haste will be our stat priority. Can you comment on that?

  • PotatoChips says:

    What’s the math behind Pulverize needing 12%? as i see it, even at 8% it is better than R&T

    • Arielle Arielle says:

      6% DR and 6% Damage is far better than 8% DR and no DPS.

    • Ahanss says:

      Not sure where the idea that pulverize provides no DPS comes from. It’s true in multi-target, but it reeks of “I’ve never tested this” when people say it in general, my experience on live is that pulv is ~5-6% DPS over no talent on single target.

    • Ahanss says:

      Actually, found it: the pre-patch launch tuning fixes also buffed pulverize by 50%. Already know you missed these for swipe/mangle, so you probably missed this pulv buff too.

  • Navoan says:

    Thanks for all your effort here Arielle, appreciate it! Bit sad that Berserk was removed and Ursoc seems to be a niche thing, I really miss having some burst DPS (BL+pot+CD).

  • Bearlington says:

    What are your thoughts on the Nature’s Guardian nerf? I honestly thought cutting it in half was a little much but at the same time I want to say it was warranted. Do you see this as having a big impact on us going into Emerald Nightmare? I think it will just force Bears into using mitigation a bit more rather than mauls. Also do you see this changing our stat priority? Ive been going with mastery => haste = vers => crit myself since 7.0. They say they “reduced mastery accordingly” so this leads me to believe we are only going to get half the benefit from mastery. Thanks!

    • Arielle Arielle says:

      It needed to happen. Perhaps not to that extent, but it definitely needed nerfing. I probably would have liked to see it end up around 0.6 – 0.66 or so (to make it definitely better than Versatility), but it was going to be absolutely absurd going into Mythic Nighthold and beyond.

      We’re still fine defensively. Mastery being the way it was just made us hilariously overpowered. The only absolute change to the stat priority is having Mastery’s value crash. Haste didn’t really come up as much as they probably wanted, but that’s just a problem with how Haste is designed to work for us.

Leave a Reply

Your email address will not be published. Required fields are marked *