Growl Testing Method

You may remember from my scattered and annoyed comments that I was not satisfied with the data I was getting last week testing Growl on the Public Test Realm. My method of testing was indeed faulty and was causing me to gather inconsistent and unreliable data, which was driving me nuts.

But now I seem to have settled on a method of Growl testing that works much better, and I wanted to share it with you. (You know, just in case you have a sudden urge to investigate pet threat or something.)

Here’s my new method:

  1. Put your pet on Passive and tell it to Stay.
  2. Cast Eyes of the Beast to control your pet.
  3. Run your pet up to your target and manually Growl once (or more, depending on your test).
  4. Cancel Eyes of the Beast. Your pet will Stay by the target, keeping your target busy and getting hit.
  5. After your pet has taken some damage, cast Mend Pet. Be very careful not to heal your pet completely or overheal, since that doesn’t count towards threat and will make your numbers too complicated to record.
  6. Keep healing your pet with Mend Pet until the target turns to attack you.
  7. Record the total amount of healing you had done just before you pulled aggro, as well as the additional amount you did that caused you to pull aggro.

Instead of recording the healing numbers themselves, I actually set myself up a little table in Excel with the heal-per-tick for each level of Mend Pet — all of which are on my hotbar for testing purposes right now. As I test, I write down how many ticks of each level of Mend Pet went off without pulling aggro, and then how much the final tick did that did pull aggro. This gives me a min and max range of healing to pull aggro.

For example, in one trial with Growl (Rank 8) on my level 70 cat, I cast Mend Pet (Rank 5) and let it tick 5 times (for 1000 healing), then Mend Pet (Rank 4) and let it tick 5 times (for another 700 healing), and then Mend Pet (Rank 1). The first tick of Mend Pet 1 was fine, but the second pulled aggro. So somewhere between 1725 and 1750 healing, I pulled aggro.

You can use this range as a flat comparison between trials (for instance, to look for Growl scaling as you raise your RAP), or you can plug in the ‘known’ values for parts of the equation and get an actual number for each Growl. If you want actual threat numbers, remember that each point healed counts as only half a point of threat and that you need to overcome more than 100% of threat to pull aggro — 110% if you are in melee range and 130% if you are further away than that.

In other words, take your healing range, divide the min and max values by 2, and then divide them again by either 1.1 or 1.3 depending — and you’ll end up with a range for the threat caused by Growl. For most of my tests, the average of the min and max threat ended up as almost exactly 664, which is the base threat of Growl (Rank 8) and indeed the number I expected for a low-RAP trial.

Getting the base numbers you expect is a good sign when you are trying out a new testing method. *grin*

If you saw the method I was using before, you’ll note that I have eliminated the Freeze Trap (since I don’t want to deal with any threat that might cause) and the Distracting Shots (since apparently they scale with RAP). Using healing instead of damage theoretically means that I can keep my pet alive as long as I need to in each trial — although in fact since I want to finish the trial using Mend Pet (Rank 1) if possible in order to narrow down the range of threat as fr as possible, I can’t actually keep the pet alive indefinitely. The damage eventually adds up. But it’s easy enough to run a couple of calibration trials, dropping down to a smaller final Mend Pet each time for more accuracy.

Oh! One more important warning: The cleansing effect of Improved Mend Pet seems to cause threat as well. I respec’d to avoid that talent when I ran these tests today.

Now that I have a method I am happy with, I intend to investigate the between-rank scaling of Growl as well as keeping an eye on PTR scaling changes. Not a lot of research has been done with lower levels of Growl so perhaps I’ll find something interesting.

Incidentally, I am now completely convinced that Growl still scales with pet AP (and not hunter RAP directly) on the PTR. In one trial I boosted my RAP to 1935 (but carefully kept my pet’s AP at 711) and saw no scaling; in another trial I left my RAP at 1682 but boosted my pet’s AP with TSA to 799 — and Growl doubled its threat.

But perhaps the next PTR build will bring the change.

33 thoughts on “Growl Testing Method

  1. Nimizar

    This is very similar to the testing methodology I ended up using (the only difference is that I still ran the pet back to where I had originally told it to stay, because I encountered a couple of problems with it deciding to run back to that point anyway after it lost aggro).

    It will be interesting to see if Hortus or Neth respond on the forums to the fact that things are quite clearly not as we have been told. There is no scaling at all with respect to hunter RAP, and linear scaling with respect to pet AP above 735, just like live.

  2. Leviath

    Just to clarify:
    “The cleansing effect of Improved Mend Pet seems to cause threat as well.”

    …causes the Hunter to gain threat, or your pet? (I presume the Hunter, but wanted to check)

  3. Seidouyumi

    I’ll go ahead and test at the lower levels.

    I have noticed some funky things with the damage that Hunters are dealing too. I am starting to think they altered the formula for crits and how RAP affects our damage. I still can’t figure out why we’re moving much faster, sometimes .5 seconds faster, than our weapons.

  4. Ravesh

    Heya!

    I am a bit confused about your testing (I am also new to all this hunter buisness). How can a (rank 1) Mend Pet cause more aggro than (rank 8) ?

    Sorry for this stupid question, but again, I have no experience with hunters what so ever, but wants to learn :)

  5. Seidouyumi

    Ravesh-

    For some reason, Healing does a lot of threat in and of itself. I have a lvl 70 Paladin, and even a lower rank of a healing spell can generate a lot of threat. Sometimes even more than the threat being generated by the Hunter himself. That is why a lot of healing trees have threat supression talents. Hunters are the only class with a healing spell that does not have any supression to it.

  6. Znodis

    That looks pretty good. I’ve been trying to figure out a good way to test threat with only white damage, but hadn’t thought of healing. I have Leggings of Beast Mastery on my Test char, some kibler’s bits, and IHM is a pet AP buff too. As a suggestion, you could do this on any level mob, say a level 50 or 60, so that your pet dying is less of an issue and you have more control over the healing/damage ratio.

    Wife is working tonight so I’ll have some time and post results :)

    @4 The threat from mend 1 was just added to the previous healing totals. She was using it to get more finely tuned results.

  7. Znodis

    Brief notes about threat:

    Most white damage is 1 threat for each 1 damage.
    Most healing is 1 threat for each 2 health healed.
    To pull agro (have the mob turn to attack you) you need 10% more threat than the mob’s current target if you are in melee range, or 30% more threat if you are at ranged (at one point this was 20yards, but I’m not sure anymore).

    Advanced info: http://www.wowwiki.com/Threat

    I also realized I can now test screech’s threat, huzzah!

  8. Sarayana

    Znodis, report on the screech please! I’m very curious about that since I just recently got a carrion bird and I find she pulls from the other pets in my group sometimes. Dies faster :o(

    I have also found lately that the growl dance followed by my throwing in an intimidation doesn’t work… I was doing the pre-kara quests and the 70 mobs were unaffected by intimidation most of the time. I’m currently level 68 and trying to get my ravager to catch up, he just hit 67, so perhaps his lower level explains it, but I haven’t had that problem more than an occasional mob before.

  9. Bobo & SgtPork

    Know what? If Growl is still scaling with Pet AP, and Hortus says it’s as fixed as it’s gonna get, I can live with that, as long as it goes off first after Charge ;)

  10. Ryai

    Sarayana: I’ve been having intimidate fail, seriously, I had a trick that usually worked, after 2.4 it didn’t, and became static in 2.4.1, it was basically pull with concussive shot, sic pet, let pet growl+attack+intimidate, FD, wait 5/7 seconds, pop up with an arcane/steady shot and if it crit, kill command..

    And had a mob in my face. Nearly all the god damn time. And intimidate realy pulls a mob off me now, like it did before, even if I’ve barely damaged it sometimes :/

    >> makes me realise Blizzard has yet to confirm and fix the FD bug that’s been going around, but anyways, I have to say that I think Screech does more threat, atleast atm, and for me, then growl, I switched to my carrion bird again, ditched my poor Dragonhawk and released my cat, and she holds aggro fairly well aslong as I let her cause 5/10% dmg, I think it’s because screech can constantly be used, and requires less focus than claw, even if it does less dmg than claw. It can also hit multiple targets at once.

    That and the debuff helps survivability pve wise.

    Also back to intimidate, before 2.4, unless there was a DRASTIC level difference which would cause intimidate to be resisted. If it hit it normally, for me, would act the same as if a lv 70 pet pulled it off. When I was leveling that aqua tiger, it wasn’t working untill lv 61/63, on the withered giants/bog lords in Zangamarsh. So it should, theoretically [did I spell that right?] work if your pet is within three levels, atleast, of the target mob.

    [but I think it was a bit busted via the whole growl fiasco or something]

  11. Znodis

    Screech – My preliminary testing gives me an guestimate of 35-48 threat per debuff. This was just messing around with some data, not accurate testing.

    Intimidate – Yes, you pet being low level is the issue. Intimidate is an affect added to your pet’s next attack. No hit, no intimidate. I’ve noticed that if you miss it seems to disappear now rather than stay until the pets next hit. The lower level means it’s more likely to resist. To your 67 pet, that level 70 mob acts like a raid boss would.

    Conc + pet growl+attack+intimidate + FD – If I’m reading this right, this is a grinding strategy? 1) You should be using your pet to pull most of the time which if you keep it at ranged, gives you the extra 30% buffer that I mentioned above. 2) The only agro you’re dumping with FD is from concussive, so once you do the above you should only need to use it when the mob then agros on you.

  12. Mania Post author

    Leviath: The cleansing effect of Improved Mend Pet was adding threat to my hunter (not the pet). Sorry for not specifying.

    Seidouyumi: The frustrating thing about narrowly targeted testing is that it often doesn’t tell you what is really going on in the big picture. But at least I can now study my tiny little picture without pulling my hair out. :>

    Ravesh: As others have clarified, Mend Pet (Rank 8) does cause more healing and thereore more threat than Mend Pet (Rank 1). I start with a high rank (like 5 or 8) to add a solid base of healing (threat) and then add smaller and smaller amounts with lower ranks (often 3 and then 1) as I know I am getting close to the threshhold where I pull aggro. This allows me be get a narrower range of threat (say 1725-1750) than if I just used high ranks (which might leave me with a range like 1600-1800).

    Znodis: If I drop too many levels of mob, they don’t do enough damage for me to heal my pet and it takes forevere. *grin* But thanks for the WoW Wiki link — I forgot to include that — and the quick test on Screech!

    Bobo & SgtPork: I believe Nimizar found they’d already hacked Charge AP to not work on the PTR. Nimizar, can you confirm?

  13. Seidouyumi

    Mania-

    I certainly understand about the hair pulling out. I actually checked my damage on the PTR vs my damage on the normal server and found that there was a difference.

    I also checked this on Growl 3. It took two hits of Mend Pet, one rank one and one rank two, and it the mob attacked my hunter at the very end of the second Mend Pet. I don’t know if that helps, but it adds up to 375 healing.

    I don’t think I can help any more than that. I think I’ll just keep gathering data on the Dragonhawks and Serpents.

  14. Mania Post author

    Seidouyumi: I’m not sure what I said came out right. What I *meant* was: It makes sense that there are other things going on with hunter aggro in general. This Growl testing doesn’t illuminate that so much, but I’m still happy I can at least test Growl. And I’m pleased that you are looking at other aspects of the situation. :> (And thanks for the data you sent me in mail. I appreciate it, even if I haven’t replied yet!)

  15. Seidouyumi

    Mania-

    You’re welcome. And yes, this doesn’t really illuminate a lot of the other issues that might be there. I’ll keep watching to see if there’s something going on. However, Something odd has to be happening for Growl to break like this.

    I don’t know if this makes sense, but I’ve always looked at situations like the old style Chinese resteraunts. Three from column A, two from column B, four from column C. I probably just showed how old I am. So, Growl might be ‘working as intended’, except that this other issue isn’t working as intended and that also be an issue.

    Still, Growl’s got a problem. The higher I’ve been getting, the more it’s been breaking.

    And you’re most welcome for the data. I’m keeping a log as I level.

  16. Palladiamors

    *Groans* Now I have to hop on and test Illi. I’ll make a note to PTR copy him and get to it in the next day or so.

    Also, apologies for my lack of posting, things have been….. rough.

    Mania, your testing seems to indicate that each level of growl rank eight does roughly 850 worth of threat. ((If my math there is backwards, I am so sorry!)) 850….. seems like an awfully low number for level seventy. I understand that our pets will be wailing on the target in the mean time, but….. I dunno. I sort of rely on growl heavily to keep threat where it needs to be, so mobs don’t try to have me for lunch. 850 is less then two auto shots, with no other flavor tossed in. Huh. Blaugh.

  17. Znodis

    Base threat of growl 8 is 664. Before BC, I don’t believe growl even scaled at all (when your pet was a level than ended in 9 it was horrible). If that seems low check out that wowwiki link I posted and consider that warrior’s ‘high threat’ attack sunder armor only generates 301 threat. Their 51 point talent devastate generates only 176 threat max. Shield slam is their other top producer with 301.

    And now you know why warriors want us to turn growl off in a group. And why 3k threat with a single boar charge was considered imbalanced.

  18. Flyv

    Nice work! The method you’re using is very similar to what the Omen and KTM authors do for testing their threat calculations. You might find it easier to test with how much white damage it takes to pull aggro; heals can be a bit strange, although 2 heal = 1 threat is the general rule.

    Znodis, if you get accurate threat numbers for Screech it’d be great if you contributed them to the Omen authors. Omen still has no estimate for the threat Screech generates. There’s two components to Screech; the damage it does and the debuff it applies. Not sure if it’s possible to separate those two. I rely on Screech to AoE aggro groups of mobs, it’d be helpful to be able to compare the threat Screech generates to the threat Mend Pet generates.

  19. Palladiamors

    Touche, Znodis, and very good point. I actually never knew that sunder was such a low aggro ability. Sadly, going with your numbers, its no wonder paladins are becoming the top tankers. It also explains why my warlock, even with an imp out and salvation up, still manages to sometimes pull aggro. Just blaugh on important numbers being so low.

  20. Nimizar

    As far as I could tell in my PTR testing, the Charge+Growl combo has been nerfed as Blizzard stated it would be (I certainly wasn’t able to get it to exhibit anything even remotely resembling the 3-4k threat that combination generates for me on live). I *think* Blizz did something to ensure a melee attack consumes the charge buff immediately (before Growl gets a chance to benefit from it), but they may also just be explicitly ignoring that particular buff when calculating the threat from Growl.

  21. Ryai

    No, this was my first strategy to try and keep mobs from eating my face, before I knew exactly wtf was going on with pet aggro and why I was pulling aggro off a pet with two auto shots at points, with no crit on my part, yet my pet should have been higher on the threat list then I was.

    And even if FD just drops my threat off after concussive, my pet should still be the highest on the list when I start to attack, with growl/attack/inimidate and hitting the mob repeatedly for the 5-7 seconds, one crit should not then pull the mob onto me, this never happened before, and it really made me freak out.

    Also I didn’t let the mob get fully into melee range of myself, I was pulling them away from pats and etc, if you’ve been to the isle lately, or atleast the first few weeks, you’d know how insane the respawns are, and how tightly packed they can be. You still can’t deny the fact, that something is wrong, when a hunter pulls a mob off the pet, after the pet has done EVERY ATTACK IT CAN and the pet should be more on the I will destroy you! list of the mob.

    Anyways current grinding is pretty much, sic Alice on target a: wait five seconds, sic Alice on target b: wait five seconds, sic Alice on final target, wait five more seconds, cast mend pet, cast ice trap, cast multi shot, start auto shooting target to death with ocassional multi shot + mend pet. And Alice keeps aggro pretty decently now, with just Screech spam :s

    Feels alot like when I was leveling my warlock tho, VW+wand = boring

  22. Seidouyumi

    Palla-

    What it looks like is that Rank 3 does around 138 threat. If 850 threat is done by Rank 8, then there’s a massive scew there.

    If my observations are right about damage being up too, then I hate to say it, but there’s a major problem with Growl.

    In other news, I finally saw Quel’Danas for the first time. I took my Paladin there today.

  23. Nimizar

    @Flyv: it should be possible to get an estimate for the Screech AoE threat by using EotB on an owl or bat, flying up to a mob, Growling at the mob, flying over near another mob, using Screech twice (still with the first mob targeted), then dropping EotB and casting a rank 1 Mend Pet or two. The point where the second mob breaks for the hunter will indicate the amount of threat generated by the debuff (Growling on the first mob just avoids any distractions due to both mobs breaking for you at around the same time)

  24. Seidouyumi

    Ryai-

    I’m MM spec. That means my first shot takes three seconds to cast and does a lot of damage. I had to switch out my shot set up so that my second shot was a sting insteand of Arcane Shot because Aimed shot and Arcane shot were doing enough damage to pull a mob off my pet.

    I’ve only been to Quel’Danas on my Paladin. Now, let me tell you, those guys are more resistant to damage and doing more damage. Some of them are level 71 to boot. I’m not sure that I would claim that the mobs there are indicative of the issues with Growl given that they are a little tougher than average. If she hadn’t been a Holy Paladin, Joushigun would probably have died several times over out there.

    And I hate to say it, but if all you were doing was wanding on your Warlock, you really didn’t know how to mix it up right. Opening with DoTs that don’t have any front end damage to them helps keep the mob at bay a lot longer on any pet. Until I got to level 50, I ran with my Succubus and hitting Improved Affliction and Curse of Agony first helped a lot and let the mob burn down a bit before I hit Immolation and THEN I hit something else, usually shadowbolt. Until the mobs became more resistant to my shadow and fire damage, I could burn a mob all the way down before it broke to attack me.

  25. Drakkena

    I think the best thing that would make things easier in this case would be Omen’s Threat Meter ( http://www.curse.com/downloads/details/9101/ ) that would also all for you to see just when you break and pull the threat from this. It helps wonders (Though for some odd reason I cannot get the thing to install for me XD)

    But yes I agree that Growl is not scaling with the pets attack power. I was doing a few tests with the Dire Raven I had tamed, and I was seeing just which part pulled more aggro. Naturally it was Screech (With growl turned off) because of the little extra damage it does but when I did Growl I had to let the Raven pummel the mob for several moments before casting that first shot because if I didn’t I’d end up pulling aggro. And I never had to do that before.

  26. Tempric

    This is great work guys. I would really like to know more about the screech numbers, as I use it a great deal.

  27. Znodis

    Still working on Screech, but I definitely will do that.

    I did some live testing last night, and basically found what you’ve already posted Mania. The one exception is that I have shown IHM does not affect growl. I’m curious if what Hortus is actually saying is that things like Kibler’s Bits and Leggings of Beast Mastery will not have an affect on Growl. The leggings at least did for me.

    http://themystichunter.wordpress.com/2008/05/06/growl-testing/

  28. Campfire

    Hi!
    I’m a survivial hunter, and I am so tired of this whole pet aggro thing. I’m at about 32% crit @~2000AP selfbuffed right now, and I’m able to pull aggro from my pet after misdirected to it with Aimed – Steady – Arcane just with Auto Shots before the mob dies…
    By now, I’m mostly doing SSO dailies and I figured out, that the best way to kill mobs is to burst them down:
    1. Send the cat to the mob (with sprint if it’s not on CD)
    2. Immediately open with Steady as soon after the pet’s departure (’cause it won’t keep aggro anyway)
    3. Concussive, Steady (for extra dazed damage), Arcane

    With a bit of critting, the mob is dead or nearly dead. So

    4. Raptor Strike or
    4. Scatter Shot, Steady
    5. Loot

    That method is very far away from levelling my hunter three years ago. But right now, it is the only viable way for me to get things done without getting mad. Or wating several seconds (5-10) at each mob for my pet to get a sufficient amount of aggro.

  29. Pingback: Mania’s Arcania » PTR 2.4.2: The Growl Saga Continues

  30. Pingback: Mania’s Arcania » Patch 2.4.2: No Changes to Growl Scaling

  31. Pingback: The Hunting Lodge » 2.4.2. Growl Issues

  32. Pingback: Mania’s Arcania » PTR 2.4.3: Growl Changes Confirmed

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>