Fixing Shadow Tank Spikiness

[ScatteredAshes]

Normally, I would be pulling out an L2P as well, and I actually *did* when we were testing this stuff on the PTS, but it's become apparent to me the more I play my Shadow tank that it *is* a problem. If a Guardian or VG in my level of gear runs into an FP, there isn't a single time where they actually *need* to burn a CD in order to survive (Guardians will often burn Saber Reflect but that's more for the threat generation than for the actual mitigation). There is at least 1 pull in *every* Hm FP (and often 2-3+) where, if I don't start off or use Deflection or Battle Readiness almost immediately after the fight starts, I stand a *very* good chance of dying. I dramatically overgear the content and yet, because of the spikiness, I can still die. That's a problem.

 

I'll state up front that I don't actually disagree with you about the whole issue of Shadows' spiky mitigation, but this a pretty silly example to use in support of your point. The HM FP dog pack pulls are hard for Shadows because of the mechanics of Kinectic Ward: a bunch of hard-hitting mobs burn off all the charges before Kinetic Ward's CD resets. A Shadow tank therefore spends some fraction of 15 secs being extra squishy. Working as intended, imo. If we're the "skill" tank, we need to be skillful enough to be aware of our Kinetic Ward charge count. It's sort of a class design irony that adds flavor to the game. Shadow tanks are great at pulling and maintaining more AoE aggro than Kinetic Ward can mitigate.

 

The huge hits from Terminate seems like a much worse problem. Any sort of Op boss mechanic that reduces Shadows' viability relative to other tank classes is VERY BAD NEWS for Shadows tanks as a class. Shadow tanks will be excluded from content if we're a liability on even one Op boss.

[Kitru]

A Shadow tank therefore spends some fraction of 15 secs being extra squishy. Working as intended, imo.

 

Except that, you know, the survivability calculations for Shadows assume 100% uptime on Kinetic Ward. As soon as KW drops off, Shadows because *noticeably* squishier *and* spikier compared to the other tanks.

 

The huge hits from Terminate seems like a much worse problem. Any sort of Op boss mechanic that reduces Shadows' viability relative to other tank classes is VERY BAD NEWS for Shadows tanks as a class. Shadow tanks will be excluded from content if we're a liability on even one Op boss.

 

Half of the problem is that it's an F/T attack. Shadows have a lot of Defense, which means it takes up a lot of our total effective mitigation. Since Defense is applied only to M/R attacks, even with our higher Shield/Abs, we still take more damage from F/T attacks than the other tanks do:

 

A BiS Shadow tank manages a survivability profile of 35/22/57/52 (factoring in KW and average contribution of KB). Against an F/T K/E attack, a Shadow manages 54.26% average mitigation. Shielded, the mitigation is 68.8%. Unshielded, it's 35%.

 

A BiS Guardian tank manages a survivability profile of 50/26/41/35 (factoring in Guardian Slash and Blade Barricade). Against an F/T K/E attack, a Guardian manages 57.175% average mitigation. Shielded, the mitigation is 67.5%. Unshielded, it's 50%.

 

A BiS VG tank manage a survivability profile of 51/19/42/58 (factoring in average contribution of Power Screen and Energy Blast). Against an F/T K/E attack, a VG manages 62.9% average mitigation. Shielded, the mitigation is 79.4%. Unshielded, it's 51%.

 

As such, a Shadow, thanks almost *entirely* to terrible K/E DR, manages *lower* DR against the F/T attacks, largely a tie with Guardians when they shield and substantially worse than a VG, while also being *absolutely* worse when it's unshielded. It's like they didn't even *look* at the math for F/T + K/E attacks when they were working on the tanks in 2.0 and only made it worse when they designed the huge individual hits to be K/E instead of I/E (which, honestly, they should be since all of the tanks mitigate those pretty much identically).

Edited May 23, 2013 by Kitru

[ScatteredAshes]

Except that, you know, the survivability calculations for Shadows assume 100% uptime on Kinetic Ward. As soon as KW drops off, Shadows because *noticeably* squishier *and* spikier compared to the other tanks.

 

YOUR survivability calculations may depend on that. I'm not at all sure the designers' do.

 

My point was that problems with HM FP dog packs aren't really indicative of Shadow tanks' Op viability and therefore aren't a serious problem. In any situation where a Shadow has to has to take damage from many hard-hitting mobs simultaneously, Kinetic Ward's charge count is going to be a survivability issue. That's sort of intrinsic to the Shadow tank class as we know it, and isn't the root the spiky damage problem that's actually a danger to our Ops-worthiness.

 

I mean... I'd love for Kinetic Ward to be a perma-buff, too, but... not going to happen. Until it does, our solution is to pop a CD on the dog pulls in HM Mando Raiders and complain in Group Chat about how Saber Reflect is OP.

 

It's like they didn't even *look* at the math for F/T + K/E attacks when they were working on the tanks in 2.0 and only made it worse when they designed the huge individual hits to be K/E instead of I/E (which, honestly, they should be since all of the tanks mitigate those pretty much identically).

 

This is the real problem, and it is a big deal. Shadow tanks will get left behind because of this. Though... I suspect that they'll fix it by changing the damage type of Terminate rather embarking on any sort of overhaul of the way Shadows mitigate damage.

[Kitru]

YOUR survivability calculations may depend on that. I'm not at all sure the designers' do.

 

I'm pretty sure the developer's survivability calcs do, mainly because they've explicitly *said* that Shadows are the most survivable tanks (actually said to me during the M&G in first person by Jesse Sky). Without the permanent uptime on KW, Shadow tank survivability drops *dramatically*. Unless you assume a near permanent uptime (re: 80-90% uptime) in virtually all cases, Shadows aren't the most survivable which contradicts what I've been specifically told.

 

My point was that problems with HM FP dog packs aren't really indicative of Shadow tanks' Op viability and therefore aren't a serious problem. In any situation where a Shadow has to has to take damage from many hard-hitting mobs simultaneously, Kinetic Ward's charge count is going to be a survivability issue. That's sort of intrinsic to the Shadow tank class as we know it, and isn't the root the spiky damage problem that's actually a danger to our Ops-worthiness.

 

You're misinterpreting my line about the multiple dogs as a problem with KW charges dropping off. The problem is actually in alpha strikes on Shadow tanks themselves. Because KW now packs 15 charges, you're pretty much *never* going to lose all of your KW stacks from an alpha strike (it would require ~38 attacks in the alpha strike to strip all of your charges immediately). Alpha strikes specifically refer to the massive cluster of attacks that occur almost simultaneously when you aggro all of the enemies in a single pack. Normally, attacks are distributed evenly over time, allowing for heals and whatnot. With the alpha strike, all of the damage arrives at one time, which, with a spiky incoming damage profile, means that you can be killed pretty much off the bat by a big enough one (which, yes, does happen with the dog packs). Oftentimes, even if you survive the alpha, a Shadow tank will be reduced to low enough hp that the high incoming damage (since damage is highest at the start of a trash pack before any of em die) isn't able to be overcome by the healer's healing. Right at the start of a fight, healing is almost always below the actual incoming DPS (this is why using bad target priority sucks hard for the tank and healer since you're extending this time of maximum incoming damage) so the natural atrophy from the high damage phase can kill you and the fundamental cause is still the alpha strike (since said atrophy wouldn't have killed you without the sudden drop in hp caused by the alpha strike).

 

This is the real problem, and it is a big deal. Shadow tanks will get left behind because of this. Though... I suspect that they'll fix it by changing the damage type of Terminate rather embarking on any sort of overhaul of the way Shadows mitigate damage.

 

I actually asked Jesse Sky about the damage type on Terminate and explained why it should be I/E instead of K/E. He said they thought about it when they were developing it and said he was quite happy with how it had turned out regardless. It set one of my eyes to twitching in perplexity.

[Panzerfire]

Here's where we're going to disagree. It's not a concern that prevents me from doing content, so it might not be game breaking in *that* sense. However, it's a drastic disadvantage that isn't really present in any of the other tanks, and the disadvantage doesn't even offer a commensurate gain elsewhere.

 

The gain is in reduced overall damage taken and more cooldowns resulting in a longer uptime on cooldowns. Right now, the healing output of healers vastly outnumbers the average damage taken and wipes can only occur due to spike damage, which automatically results in shadows being the *worst* tanks since they simply are the only tanks capable of dying in a group doing a fight properly with end game gear. I'll say that shadows will pick up on nightmare modes if the total damage taken, raidwise, gets closer or goes over the optimal healing output. Kephess in EC comes to mind. If we have to, we'll be offtanks only to be used with cooldowns up in order to free up healing for other raid members. The fact remains that shadows are amazingly good if you tailor your cooldowns around boss abilities and they get better with an increase in damage output of bosses. Except for the self-healing component, which was and is a problem, since 2.0 acts only as a band-aid for self heals falling behind.

 

It's not a question of Shadows being *incapable* of doing content. It's a question of it being such a major disadvantage when *doing* such content, such that, even when you factor in the advantages that Shadows *do* get (short CDs, Resilience, self healing, better average survivability), Shadows are *still* at a disadvantage from a functional standpoint.

 

The functional disadvantage of shadows is debatable per ops group. If most of the overhealing can be avoided then shadows perform a lot better, as this is the main counterargument to shadows being good. How that happens is up to you, whether by healing assignments or cooldown usage, we're rotating 3 assassins and 1 jugg for 16 man and the only problem was the first kill on HM Thrasher, which even now remains the only *fun* fight, healers being used to assassins actually prefer to heal us over other tanks. And I know that we're far from being perfect so it can get even better. This is part of the *skill* requirement, no? As long as shadows are not consistenly dying due to bad luck and they get an advantage over the other tanks they remain the *skill* tanks.

 

I will agree that such an advantage is hard to measure and quantify. And juggs/PT's will just give you an easier time. An interesting thing would be to see how much mitigation we can sacrifice for more endurance in order to reduce overhealing, as there is less incencitive to quickly heal someone if they cannot be 1-shotted at 70-80% of their health. I'll try this out someday by switching from crafted implants/ear to UW once I have them. And maybe enhancements, although I don't think the trade on high vs low endurance 72 enhancements is worth it.

Edited May 23, 2013 by Panzerfire

[ScatteredAshes]

I'm pretty sure the developer's survivability calcs do, mainly because they've explicitly *said* that Shadows are the most survivable tanks (actually said to me during the M&G in first person by Jesse Sky). Without the permanent uptime on KW, Shadow tank survivability drops *dramatically*. Unless you assume a near permanent uptime (re: 80-90% uptime) in virtually all cases, Shadows aren't the most survivable which contradicts what I've been specifically told.

 

Jesse Sky may well have told you that, but I suspect the encounter developers took Shadows' cooldowns into consideration when copy/pasting hard-hitting space dogs all over the new HM FPs. Those encounters are far from impossible for us. We can pop CDs, CC them, kite them, or whatever. My point remains that those gnarly dog packs aren't a general indicator of Shadow tank inferiority vs. Guardians and Vanguards, certainly not of the same severity as Shadows getting 1-shot by Terminate.

 

(very interesting stuff about Alpha Strikes elided in the name of brevity)

 

Excellent explanation. Thanks!

 

I actually asked Jesse Sky about the damage type on Terminate and explained why it should be I/E instead of K/E. He said they thought about it when they were developing it and said he was quite happy with how it had turned out regardless.

 

Yeah, that is perplexing. Hopefully, they come around. And I still think that even if they admit there's a problem, the fix will be that they retune the encounter. I don't think they'll change Shadows' mitigation fundamentals. The TOR devs seem a lot more reluctant to fiddle with class balance and mechanics than the WoW devs are.

 

It set one of my eyes to twitching in perplexity.

 

Your AV is a Miraluka. How does that even work?

Edited May 23, 2013 by ScatteredAshes

[Kitru]

Jesse Sky may well have told you that, but I suspect the encounter developers took Shadows' cooldowns into consideration when copy/pasting hard-hitting space dogs all over the new HM FPs.

 

Oh, he also said that they never tested *and* of the current content with a Shadow tank. At all. He literally had no idea that any of the problems we're experiencing with Shadow tanks would occur because they, quite literally, *they did no testing to see if there would be any*.

 

When he asked about what trash packs were the biggest problems, I told him about the big dog packs, and he mentioned that they were meant to be hard packs but had no clue that they affected Shadows so much more than VGs or Guards.

 

Your AV is a Miraluka. How does that even work?

 

I'm not actually a Miraluka in real life; I just dig the veil and wear it wherever I go. /girlydance

[Altheran]

Okay, I'll give my idea that I kept in mind for a bit now. At first I wanted to write a post about it in suggestions... but well I'll write it here.

 

What if Endurance would have an other effect about incoming damage, in addition to increasing HP pool ? And I mean a kind of anti-main-stat effect.

I don't know if you know how Warhammer Online handled Endurance but that's the main idea : having Endurance negates a part of ennemy's Aim/Strength/Cunning/Willpower.

How I imagined it was that each point of Endurance over the amount of mainstat you have will have this negating effect. That means that you have this effect only when you have more Endurance than mainstat. DPS will not be able to benefit from it as they're supposed to have more mainstat.

I thought of it a like of 1 extra Endurance equals 1 less mainstat from the ennemy but I realise this is too much. A lower ratio is probably better (2:1 or 3:1).

It would help all the tanks, but more Shadows and Assassins than others because they have extra Endurance others doesn't have (Shadow training, +3% in the tree), and they are the only one to really care about it (it will work on top of self-heals), and finally they're the only ones without mainstat talents (according to my idea, less mainstat equals more effect from Endurance).

 

With this, all tanks will now care about Endurance (which is at the moment a filler stats for other tanks : it doesn't help taking less damage, it doesn't help striking harder, it just delay death), and maybe Shadows will use some Endurance augments... at the moment, nobody use them (or at least nobody I know).

In PvP, nobody cares about Endurance. It may change with it. And it allows tanks a way to counter a bit the overflowing amount of mainstat and power DPS have : Tanks have Defense opposed to Accuracy, Shield to Crit, Absorb to Surge, but they can't do anything against Power or Alacrity but trying stacking more defensive stats even with DR, or trying to go for the offense.

Edited May 23, 2013 by Altheran

[dipstik]

this is easy

 

undo 1.3 changes:

 

Combat Technique now generates 100% additional threat while active. The healing generated by this ability has been reduced by approximately 50%. Its armor bonus has been reduced to 115% (down from 150%).

 

also: get rid of kinetic ward... no other tank has to use a non-attack ability to maintain optimum play besides us, and we can still lose the benifits of it. we already get 5% more defense to our base, why not 15% to our shield... and I'll take 8% absorb too while they are at it.

 

maybe link bulwark to out TTK, so we only lose our absorb buff (which still procs off shielded attacks) when we heal ourselves up (giving us 12% instead of 8% would be nice too... since that was another 1.3 blunder)

Edited May 23, 2013 by dipstik

[Kitru]

this is easy

 

undo 1.3 changes:

 

Combat Technique now generates 100% additional threat while active. The healing generated by this ability has been reduced by approximately 50%. Its armor bonus has been reduced to 115% (down from 150%).

 

The problem is that a 45% increase in armor rating will change the fundamental average survivability of Shadows. The nerfs we've consistently received are derived explicitly from having "too high" of an average survivability. Anything that *increases* our average mitigation is pretty much going to be outright refused, because, according to the devs, our average survivability is high enough.

 

Any ideas that stand a chance of being implemented by the developers have to decrease spikiness while having a negligible impact on average survivability.

[dipstik]

what are you using for spikiness? some kind of weighted standard deviation?

 

so we need to change the standard deviation without changing the mean? or change the weights?

[Vaidinah]

Panzerfire:

 

You don't need to look for the Energize proc at all since Project/Shock is a skill you use on cooldown. The only time you don't is when you have 3 HS stacks or your Accuracy debuff has worn off.

 

Kitru:

 

Based on what you've said about Guardians, you definitely fall into the "knows little about Guardian play" category of people I mentioned previously. Your point that Guardian cooldowns are "equally effective" regardless of how they get used proves this perfectly. Based on your argument, a Guardian who blows all their defensive cooldowns immediately upon starting a fight is "equally effective" as someone who uses them at the appropriate time.

 

You do seem to understand what complexity at least (even if you apply it incorrectly) and that is why Guardians have a higher skill ceiling. As I said, not only do Guardians have more skills to use, but those skills in general have more depth in how to use them optimally. They are based more heavily on what other skills you have used and what the situation is in the fight. Guardians have more thoughtful and fluid decision-making to optimally play both offensively and defensively.

 

Knowing what kind of attacks a boss uses is very easy and just requires the ability to read a combat log and/or random tests to see what attacks Resilience works against and doesn't since it's all-or-nothing. Shadows are based mostly on easily memorizable facts both offensively and defensively. They have relatively few decisions that they can use to play optimally. The difference between effective and optimal Shadow tanks has little to do with anything Shadows have on their own and more about other skills that are related to tanking as a whole.

[Kitru]

what are you using for spikiness? some kind of weighted standard deviation?

 

Spikiness, as I'm using it here, is less a specific value and more of a generic concept. In practice, I'm putting together a simulator that's going to create a construct to analyze spikiness by looking at deaths and lowest hp over many interations (my test number has been 100 iterations) of generic (for right now) fight constructs.

 

Right now, the basic stuff I'm putting is in BiS tank stats copied from your and my AMR profiles and throwing it into a generic construct that imitates the fight with Operations Chief, assuming no CD use with both spike and standard healing set ups used at variable overheal points.

 

The results (from what I've done) have been less than favorable: in a fight with 5k pre-mit M/R + K/E attacks every GCD,, one 50k pre-mit F/T + K/E attack every 40 GCDs, basic heals of 4k per GCD, and spike healing of 8k every 6 GCDs, out of 100 iterations of 400 GCD length, a BiS Shadow tank ends up dying more than 50% of the time and, of those that survive, their lowest hp spike is ~800 (out of 36k total hp). A BiS Guardian survives all 100 and only spikes down to an average of ~8k. If you tweak the spike to 10k every 8 GCDs (so it's the same damage), both tanks survive all of the fights, the Shadow's lowest spike is ~25k and the Guardian's is ~27k.

 

I'll be releasing the simulator when I get it into a bit more of a presentable place.

 

so we need to change the standard deviation without changing the mean? or change the weights?

 

If you want to turn it into a calculation, spikiness is standard deviation in incoming damage. To affect spikiness without changing the total mitigation, you have to change the standard deviation of incoming damage while reducing the mean.

 

I'm also working on creating a non-simulation algebraic formula to provide a "spikiness" factor for a class by checking damage reduction in each attack resolution case and comparing the weighted variation (Defense/Resist is extreme DR but it's weighted less because it happens much less than Shield/Def).

 

It's almost disturbing how productive I've been the last couple of days...

[Kitru]

Based on your argument, a Guardian who blows all their defensive cooldowns immediately upon starting a fight is "equally effective" as someone who uses them at the appropriate time.

 

You're creating a laughably bad strawman variation of my argument here. I'm not referring to *timing* of use, but *generality* of use. A situation that Saber Ward is useful to apply to is pretty much the same situation that Endure Pain or Warding Call will apply to. All of the Guardian CDs are more general than the Shadow ones. *That* is the complexity since a Shadow has to consider what type of damage is coming when using whereas a Guardian just has to click one of their multiple CDs, and it doesn't really matter which. The only that Guardians have is *more* and *more* is not "complexity".

 

You do seem to understand what complexity at least (even if you apply it incorrectly) and that is why Guardians have a higher skill ceiling. As I said, not only do Guardians have more skills to use, but those skills in general have more depth in how to use them optimally. They are based more heavily on what other skills you have used and what the situation is in the fight. Guardians have more thoughtful and fluid decision-making to optimally play both offensively and defensively.

 

More skills is *not* the same as more complexity. I'm kind of curious as to how you've managed to find depth of use for your basic priority is. Seriously. The *only* variable is Courage stacks and, even then, it's a variable that can be ignored. There is, quite literally, nothing else that has to be considered. You just use everything on the 12 second cycle. Sure, you've got some 1 minute CDs, but, once again, like your surv CDs, it's really just a choice of a number of incredibly similar choices. It's like a VG being "complex" because they have Mortar Volley. It's a 1 min CD that they have to weigh against using other stuff, but it's not really a hard decision. The only difference is that Guardians just have more options that fulfill the same exact objective. You could click any of them and accomplish what you want.

 

The difference between effective and optimal Shadow tanks has little to do with anything Shadows have on their own and more about other skills that are related to tanking as a whole.

 

And the difference between a good Guardian and a bad Guardian is... not much? The ability to use Guardian Leap effectively is a factor but the sheer rarity of times that Guardian Leaping is going to have any real impact is going to limit how important *that* is. The same applies to the ability to use Focused Defense while remaining some level of effectiveness as a tank: sure, it'll have some kind of effect, but the situations where you *need* that skillset are so *absolutely* rare, that it's not actually going to impact anything.

 

Even if the complexity inherent in a Shadow is something that *all* tanks can learn as good tanks, it doesn't change the fact that Shadows, to be played well, *require* those skills to properly utilize the class. Guardians can be played perfectly well without *ever* grasping the "deep" complexity of Guardian tanking and, if/when they do, it's going to have an almost imperceptible impact on their performance. When someone learns how to use what tools they have a Shadow, on the other hand, the impact is *massive*, which is why people talk about specific Shadows being amazing and just talk about Guardians and VGs in general terms referring to their class. No one but *you* is gonna notice your mastery of Guardian Leap and Focused Defense; my healers are *definitely* going to notice how effectively I abuse the living hell out of Resilience, maximize my self healing, and use Force Speed to my advantage.

 

Let me put it this way:

 

Bad Shadow tank = 50%

Decent Shadow tank = 90%

Good Shadow tank = 110%

 

Bad Guardian = 100%

Decent Guardian = 109%

Good Guardian = 110%

[Vaidinah]

You're creating a laughably bad strawman variation of my argument here. I'm not referring to *timing* of use, but *generality* of use. A situation that Saber Ward is useful to apply to is pretty much the same situation that Endure Pain or Warding Call will apply to. All of the Guardian CDs are more general than the Shadow ones. *That* is the complexity since a Shadow has to consider what type of damage is coming when using whereas a Guardian just has to click one of their multiple CDs, and it doesn't really matter which. The only that Guardians have is *more* and *more* is not "complexity".

 

Generality of use is not something that should matter much as timing is far more important. It doesn't take much time to learn what skills can be resisted with Force Shroud even it is something only Shadows *have* to learn to be effective that other tanks don't. While any of the main 3 Juggernaut defensive cooldowns can be useful to some extent in most situations, some are definitely than others based on the situation. For example, Saber Ward/Warding Call is much more beneficial than Invincible when you don't want to just reduce the damage you take, but also have a good chance at negating effects that come with getting hit (like stacking debuffs). The 2-3 seconds of Blade Turning (100% deflect) makes a massive impact with Saber Ward/Warding Call as well and helps relieve pressure off healers more than almost anything else. Endure Pain should be used earlier in fights since it has a lower cool down and is more effective against more common burst situations.

 

More skills is *not* the same as more complexity. I'm kind of curious as to how you've managed to find depth of use for your basic priority is. Seriously. The *only* variable is Courage stacks and, even then, it's a variable that can be ignored. There is, quite literally, nothing else that has to be considered. You just use everything on the 12 second cycle. Sure, you've got some 1 minute CDs, but, once again, like your surv CDs, it's really just a choice of a number of incredibly similar choices. It's like a VG being "complex" because they have Mortar Volley. It's a 1 min CD that they have to weigh against using other stuff, but it's not really a hard decision. The only difference is that Guardians just have more options that fulfill the same exact objective. You could click any of them and accomplish what you want.

 

I disagree. While the initial rotation is mostly the same for each Juggernaut tank, those 12 second cooldown skills can vary greatly during the fight depending on what you need. Sure, you can hit them in consecutive order every time to do okay, but there are important considerations if you want to play better. You can vary when you use Sundering Assault a great deal. The debuff last 45 seconds so after that, you only need it for Focus/Rage. Depending on how much rage I've had, there's been several times where I haven't used it for a while because I simply didn't need the rage. Another example is Force Sweep/Smash as the Accuracy debuff lasts 20 seconds while it has a 12 second cooldown. Thus, it does not need to be used on cooldown if other skills are up or if you want to spread it out for Courage/Revenge stacks.

 

Doing this gives you more overall rage, which can be used for other attacks like Vicious Slash for extra damage. There are more choices as well like the under 30% phase when any extra rage you have increases your damage considerably. For Shadows, the difference in resource management under 30% is trivial, but it makes a big difference to Guardians if they want to use their execute on cooldown.

 

And the difference between a good Guardian and a bad Guardian is... not much? The ability to use Guardian Leap effectively is a factor but the sheer rarity of times that Guardian Leaping is going to have any real impact is going to limit how important *that* is. The same applies to the ability to use Focused Defense while remaining some level of effectiveness as a tank: sure, it'll have some kind of effect, but the situations where you *need* that skillset are so *absolutely* rare, that it's not actually going to impact anything.

 

Even if the complexity inherent in a Shadow is something that *all* tanks can learn as good tanks, it doesn't change the fact that Shadows, to be played well, *require* those skills to properly utilize the class. Guardians can be played perfectly well without *ever* grasping the "deep" complexity of Guardian tanking and, if/when they do, it's going to have an almost imperceptible impact on their performance. When someone learns how to use what tools they have a Shadow, on the other hand, the impact is *massive*, which is why people talk about specific Shadows being amazing and just talk about Guardians and VGs in general terms referring to their class. No one but *you* is gonna notice your mastery of Guardian Leap and Focused Defense; my healers are *definitely* going to notice how effectively I abuse the living hell out of Resilience, maximize my self healing, and use Force Speed to my advantage.

 

I've already stated that I believe Shadows take the most skill of any tank to play at an effective level (at least in 2.0 due to how Guardians were simplified). However, there are a ton of things that Guardians can do to stand out and I and other people I've met (especially my healers) definitely notice a massive difference between a normal tank and a good one. This was especially true pre-2.0 since there weren't a lot of people who could really play the hybrid Juggernaut tank well enough to be worth it. I say that as someone who plays every role and has dealt with a large variety of tanks in guild groups and pugs.

[JouerTue]

just yesterday i was healing cademimu..shadow tank. good...he would have handled those packs easily.

nope..he was going down like a stone from the stratosphere.

then inspected him and found he was in tank gear, i had seen him pull an enemy so tank spec..

the gear wasn't completely at the rating, but i started doing 55hm more less 2 seconds after reaching the cap on my guardian and on a couple of healers and i had always good performance..maybe i had to wait a bit to clear hammerstation bonus or athis...but shadow tank is definetly too much spiky.

i like the suggestion about a cd (like kolto overload of PT tanks) but proccing damage instead of healing..or an increased armor or self heals..would be nice

[Leafy_Bug]

Eh, Brutus, some of us have been saying this for a while and you pulled the l2p argument out of your trumpet giving us some nice numbers saying how awesome shadow tanks are. I see you open this thread and all I can say is L2P, shadow tanks are fine, what you told quite a few of us when we complained about the spikyness.

[Kitru]

Eh, Brutus, some of us have been saying this for a while and you pulled the l2p argument out of your trumpet giving us some nice numbers saying how awesome shadow tanks are. I see you open this thread and all I can say is L2P, shadow tanks are fine, what you told quite a few of us when we complained about the spikyness.

 

When *everyone* was complaining about the difficulty of the HM FPs (which is what was happening on the PTS), it isn't a problem. When *only Shadows* are disadvantaged in content because of the massive spikiness, it becomes one.

 

You'll also note that I specifically mention that it's not as if you can't mitigate the spikiness through playstyle. Most of us have already been doing that, recognizing that incoming damage is ridiculously spiky. The issue is that the only ways to *do* this reliably involve using CDs that no other class has to and, even then, in some situations, you'll *still* get wiped out thanks to spikes, and no other tank has to do this. Shadow CDs are now consumed primarily to mitigate spikiness rather than to mitigate high incoming damage phases. They've had to become a fundamental part of the basic Shadow playstyle rather than being held in reserve.

 

Also, I find it important to note, especially thanks to your little bit of self-righteous pseudo-vindicated petulant commentary at the beginning, that you're conflating the concerns on the PTS with the concerns I'm bringing up here, *and* you'll also note that, concerning *this* issue, I actually admit that I got it wrong. On the PTS, the cries were about Shadow average incoming damage being too high. Not spikiness, incoming damage. The *incoming damage*, which can be determined pretty simply by some basic mitigation math, was actually slightly better than what Guardians or VGs had. From *that* standpoint, the complaint doesn't hold ground. It was only *later* that spikiness began to take the forefront, when the topic was buried under piles of average mitigation stuff (mostly about complaints that the loss of 20% armor from Stasis was going to *completely ruin* Shadows, even though the 1-2% additional DR would actually have a negligible impact on spikiness as a whole).

 

I've never claimed that I am infallible (I've never claimed such, I only claim that I'm *less fallible* ), and I admit that, upon first experiencing RotHC, I expected the spikiness issue to be something of an l2p and followed suit with that (since that's generally been the explanation of Shadow spikiness before). *However* with the changes to Guardians and the change in content design to apply levels of spikiness that are only really survivable by what was previous the domain of only VGs (and is now the domain of *both* VGs and Guardians), the tables have turned and spikiness is no longer a *play* concern but a fundamental *balance* concern.

 

Spikiness is *normally* an l2p issue, and, right now, it's possible to treat it as such sometimes, though it will still bite you in the *** remarkably often (bad RNG on Thrasher multiple times in the same fight will chew through your CDs and then outright kill you when they're down; bad luck with the Resilience bug will cause you to make a coin flip for life or death with Operations Chief). When content isn't designed in such a way that specific situations only have the options "instant death" and "didn't even notice it", spikiness is a tolerable issue and is only a problem when you don't know how to play (generally by not doing something on a skill tank that you should be doing to maintain the expected mitigation). It's only when the *content* is designed as with massive spikes like they currently have that it becomes a major concern. Since I doubt the developers are going to change the content (having actually talk to Jesse Sky about it and seen his nonchalance about the entire topic), the only recourse is to find some mechanism by which Shadows can become *less* spiky so as to ameliorate the issues that we *do* have.

 

In short, yes, when we were discussing this 2 months back on the PTS, I was telling people they were doing it wrong. Now that I've got more data and experience, I'm changing that. It's part of my prerogative as a human being, so cram it up wherever your species traditionally crams things. I don't have a problem admitting I was wrong when I actually *was* wrong. It doesn't change the validity of what I'm saying *now*.

 

P.S. The line is "Et tu, Brute". If you're going to quote Shakespeare, do it right. Though I'm not exactly sure how I'm backstabbing you when I'm championing the cause you supposedly believe in.

Edited May 24, 2013 by Kitru

[dipstik]

i like your simulation approach to the problem, but your statement that we need to change the standard deviation while reducing the mean seems to go along with my simple solution of giving us back our pre 1.3 armor rating buff. this will reduce the mean. to change the standard deviation we just need to change the mitigation numbers until we find a sweet spot.

 

in terms of an equation for spikiness:

 

standard deviation will be based on the square of the difference between each attack state and the mean

let the square of the difference between the mean and the mitigation for an attack state i be called sig_i

each attack state will be given a probability a_i, and will be coupeled with a mitigation state probability mit_i

spikiness= sum_i(p_i*mit_i*sig_i)

 

=w*(0.5*(d+0.1)*(0-mean)^2+0.5*d*(0-mean)^2+s*((1-a)*(1-arm)-mean)^2+(1-d)*(1-s)*((1-arm)-mean)^2)+f_k*((1-s)*(1-r)*((1-arm)-mean)^2+r*(0-mean)^2+s*((1-a)*(1-arm)-mean)^2)+f_i*(r*(0-mean)^2+(1-r)*((1-ir)-mean)^2)

 

w is probability of a melee/ranged kinetic/energy attack

d is defense chance

mean is post armor squishiness

s is shield chance

a is absorb chance

arm is armor kinetic/energy damage reduction

f_k is proability of force/tech kinetic/energy damage

r is resistance

f_i is probability of force/tech internal/elemental damage

ir is internal/elemental damage reduction

 

 

i got the following weighted standard debviations and coefficents of variation (stdev/mean squish) and post amor+heals/absorb damage taken per second assumeing 4500 dps with SnV hM dmagae weights:

 

shadow: 0.05437 , 0.171138397, 1111.76

jugg: 0.04197, 0.130544346, 1330.13

Vanguard: 0.03504, 0.126945887, 1242.12 (shoulder cannon and adrenaline rush not included)

 

setting shadow armor to 150% gives:

0.04886

0.162665032

1033.85

 

so we still have a high coefficent of variation. i would liek to get the coefficent of variation down to .13 while still gertting 1100 damage taken per second

 

setting armor buff to 150% and getting rid of 5% defense buff (the one that we have that no other tank has) gives:

0.04598

0.145510263

1104.20

 

i think this is a REALLY VIABLE solution. nerf 5% defense for 35% armor. if you can plug this into your simulator, let me know if it is effective/reasonable.

Edited May 24, 2013 by dipstik

[Kitru]

i like your simulation approach to the problem

 

I've already tossed the basic simulator to the devs with some of the first run information so, hopefully, they'll at least check out the information. It's a very basic simulator atm, but I'm planning on throwing in a fair deal more complexity (healing and tanking are *really* simple; plan is to make them a bit more complex and generate different profiles for each boss/healer). The goal, in its current form, was just to demonstrate the massive vacillation in survivability and variability in damage taken incurred between the 3 tank classes, which it does admirably.

 

your statement that we need to change the standard deviation while reducing the mean seems to go along with my simple solution of giving us back our pre 1.3 armor rating buff.

 

The point was that increasing our armor would increase the mean as well as reduce the standard deviation, which isn't likely to be something the developers want to implement.

 

so we still have a high coefficent of variation. i would liek to get the coefficent of variation down to .13 while still gertting 1100 damage taken per second

 

Since Juggs are already sitting pretty at .13, I think it would be perfectly fine if we could bring our spikiness factor down to .14 or even .15. .13 might be a bit *too* not spiky.

 

The problem, as I indicated in the OP, the only purely stat based way this is going to happen is if we reduce Shadow Def/Shield/Abs to increase damage reduction, which probably won't be a very popular move amongst us Shadows. We likes our Def/Shield/Abs. Even with your own numbers, we take the least damage over time (by about 15%), but we're ~50% spikier. The disconnect between those 2 numbers just isn't tolerable, and it susses out in the simulations I've done. I kind of wish we could post polls in these forums just to determine exactly what stats the Shadow tank community would be willing to sacrifice in exchange for higher DR.

[Kitru]

setting armor buff to 150% and getting rid of 5% defense buff (the one that we have that no other tank has) gives:

0.04598

0.145510263

1104.20

 

i think this is a REALLY VIABLE solution. nerf 5% defense for 35% armor. if you can plug this into your simulator, let me know if it is effective/reasonable.

 

Keep in mind that the 5% increase in Defense that we get is a baseline quality of being a Consular. Removing that 5% defense would necessitate either removing it from all Consulars (doubtful, since that 5% higher natural Defense chance is supposed to be part of the fundamental balance calculation to account for light armor) or having the tank stance apply a debuff to Defense chance. Conversely, returning to my original idea, doing the math for armor rating, putting CT up to 150% from 115% amounts to an increase of roughly 4% K/E DR, which is what I recommended to make up for the removal of Force Breach's 5% accuracy debuff. So, the simplest solution might just be to get rid of the accuracy debuff on FB and replace it with 4-5% DR.

 

I plugged the removal of 5% defense and 4% increased DR to the simulator and ran it through the same parameters as before (100 iterations, 400 GCDs, 5000 basic M/R+K/E per GCD, 50000 F/T+K/E spike per 40 GCDs, 4000 heal/GCD, 8000 spike heal w/ 6 GCD downtime) and it turned out a lot better. The average lowest hp was still *really* low (1204.0), but there weren't any deaths. For comparison's sake, Guardian numbers were 0 deaths with an average lowest of 7.6k.

[grallmate]

I'll just chuck out a couple of things regarding the whole debate but let me start by saying, I've noticed the massive spikiness difference between my Guardian and my Shadow (and my co-tank who is a Shadow).

 

With Terminate, it being Force/Tech or Melee/Ranged makes no difference to spikiness. In fact, it being Force/Tech is BETTER since you can (somewhat unreliably) use Resilience to completely mitigate it compared to only having a chance to mitigate it, even with Deflection active. The issue isn't the attack type, its the fact that its a massive discrete hit, and probably more importantly, that Resilience is currently somewhat buggy.

 

If you want to look at bad encounter design, look at Xeno. A Shadow can solo tank HM by using Force Cloak to break the channel on Thermal Tolerance while the other tanks have to eat a full channel (or preemptively move to LOS). That fight is DRASTICALLY easier with a Shadow than either of the other 2 tanks. I took my Shadow in and we 1 shot our first attempt, when we came back with a Guardian and a Vanguard (I was DPS for that) we wiped a couple of times adjusting to that mechanic that we'd ignored the first time. The difference is, rather than being more difficult with 1 type of tank, its stupidly easy with one type. Sometimes it works both ways and it isn't enough to make it unclearable with one type but its enough to really highlight the differences between the tanks.

 

With attempts to reduce spikiness, you're only option is really to increase K/E DR rather than DR from chance based mitigation or self healing, or that overhealing barrier thing Kitru mentioned somewhere. If that was done Shadows would have to lose a similar degree of survivability from one of their other mechanisms, either spec bonuses to mitigation, KW or the self heal. Alternatively, you could remember that tank gearing involves taking enough HP to stop yourself getting 2 shot and not just stacking mitigation.

 

As the spikiest tank, increased HP gives your healers more reaction time and has the added bonus of increasing your self heal. I do think a degree of the problem is in your self admitted stubborn and crotchety nature Kitru. That isn't to say I wouldn't mind spikiness being reduced nor am I saying that stacking HP is better than mitigation. All I am saying is get back to the root of tank gearing: get enough HP to not be 1 (or 2 or 3) shot, enough mitigation to bring your incoming DTPS to a healable level and what you do after that is up to you... or your healer. Trying to minimize damage taken is nice but not at the expense of those 2 goals.

 

Complexity between tanks I find to be largely moot. Shadow 'rotation' isn't difficult and unless you're standing still doing nothing (or using the wrong rotation) the only consideration compared to the other tanks is: can I get off a full TkT before I have to move. It's only 'complicated' in the sense that if you aren't actually following your priorities you end up with lower self healing and if you do nothing you'll take more damage than a Guardian or Vanguard who also does nothing. Yes the Shadow CDs are somewhat mutually exclusive but the Guardian ones are better when used for a particular attack/damage type, the difference is 'all or nothing' compared to 'great or just good'. Saber Reflect does skew that a bit since its an 'all or nothing' CD and can make an amazing difference to the Guardian's damage taken. However the number of times its actually useful in an Op is somewhat limited thanks to the single target restriction. Off the top of my head here are times that Resilience works and Reflect doesn't in current content:

- Kephess' Dread Bomb

- Huge Grenade

- Isotope Release

Comparitively, Saber Reflect works and Resilience doesn't on:

- The TFB's Scream

 

If you look back to NiM EC as well you can add Stormcaller's Cleave and Kephess' DoT to the first list while adding nothing appreciable to the second (Pulsar Droid Railshot.... wooo....).

[Kitru]

If you want to look at bad encounter design, look at Xeno. A Shadow can solo tank HM by using Force Cloak to break the channel on Thermal Tolerance while the other tanks have to eat a full channel (or preemptively move to LOS).

 

I know I've seen more than a few groups who use a Sage to Rescue the TT tank to a pillar to break LoS, as well as others that have just had an rDPS stand by the pillar to let a Guardian leap to them. Cheesing the TT isn't particularly *hard*, it's just easier for a Shadow.

 

On that same topic, I know TT is a single target effect since it only affects the one target (you can move around and stand on other people and it doesn't bounce to them). Are you sure that Saber Reflect doesn't work on it at all? If it does (and I would be surprised if it doesn't), then the only disadvantaged tanks are VGs (which, once again, doesn't really surprise me since VGs have kind of been the red headed stepchild tank as far as survivability has gone for a long while).

 

Alternatively, you could remember that tank gearing involves taking enough HP to stop yourself getting 2 shot and not just stacking mitigation.

 

The problem with this is that the only way you're going to get enough hit points to actually avoid the quick 2 shot after a spike is to have in excess of 40k hp atm. In *getting* that much hp, you're going to drop your survivability *substantially* so, in order to survive the spikiness of a Shadow tank, you're both making it more common and undoing the entire reason *why* we tolerate having a spiky incoming damage profile (lower average damage taken).

 

The only time that I sacrifice Endurance without getting something survivability based in return is on Armorings. If *that* difference is making a substantial impact on my survivability, I'll undo it because that's my natural craziness impacting my survivability. However, if all of the other tanks are perfectly capable of absolutely maximizing their mitigation when Shadows are forced to reduce their mitigation and stack hp (and, for 40k hp, that's a *lot* of lost mitigation), I just can't agree with that design.

 

As the spikiest tank, increased HP gives your healers more reaction time and has the added bonus of increasing your self heal. I do think a degree of the problem is in your self admitted stubborn and crotchety nature Kitru. That isn't to say I wouldn't mind spikiness being reduced nor am I saying that stacking HP is better than mitigation. All I am saying is get back to the root of tank gearing: get enough HP to not be 1 (or 2 or 3) shot, enough mitigation to bring your incoming DTPS to a healable level and what you do after that is up to you... or your healer. Trying to minimize damage taken is nice but not at the expense of those 2 goals.

 

Concerning the attacks/content design, I have to say that the only reason this is being brought up because of a confluence of 2 design constructs: spikiness *alone* isn't a bad thing in class design (as evidenced by there being largely the same level of Shadow spikiness from a raw design standpoint pre-2.0 as post-2.0 and no one really had a problem with Shadows in pre-2.0), nor is hard spike incoming damage in content design (as evidenced by how Guardians and VGs react to it). The problem only arises when *both* of those design constructs occur simultaneously. Since I doubt the developers are going to change the content design (the "easy" solution would be to turn Terminate into a 28K I/E damage attack, but Jesse Sky already pretty much dismissed that when I brought it up), the easiest solution is the change the class design.

 

Hopefully, the changes recommended here (either new CDs to provide some degree of burst avoidance or a change to the average mitigation profile to normalize incoming damage to some extent) will be implemented since, honestly, they could be implemented without much problem and would solve the problem *very* well.

[grallmate]

I know I've seen more than a few groups who use a Sage to Rescue the TT tank to a pillar to break LoS, as well as others that have just had an rDPS stand by the pillar to let a Guardian leap to them. Cheesing the TT isn't particularly *hard*, it's just easier for a Shadow.

 

On that same topic, I know TT is a single target effect since it only affects the one target (you can move around and stand on other people and it doesn't bounce to them). Are you sure that Saber Reflect doesn't work on it at all? If it does (and I would be surprised if it doesn't), then the only disadvantaged tanks are VGs (which, once again, doesn't really surprise me since VGs have kind of been the red headed stepchild tank as far as survivability has gone for a long while).

 

Yeah it can be Reflected and Resilienced which isn't too much of an issue. The game changer is using Force Cloak during it to end the channel at 1 stack and while the others can cheese the mechanic if they're good (I preempt it on my Guardian and move to the pillar using taunt to keep aggro and then LOS when it starts) it is way easier on my Shadow. The extent of my pre-thought then is "don't use Force Cloak for something else".

 

The problem with this is that the only way you're going to get enough hit points to actually avoid the quick 2 shot after a spike is to have in excess of 40k hp atm. In *getting* that much hp, you're going to drop your survivability *substantially* so, in order to survive the spikiness of a Shadow tank, you're both making it more common and undoing the entire reason *why* we tolerate having a spiky incoming damage profile (lower average damage taken).

 

The only time that I sacrifice Endurance without getting something survivability based in return is on Armorings. If *that* difference is making a substantial impact on my survivability, I'll undo it because that's my natural craziness impacting my survivability. However, if all of the other tanks are perfectly capable of absolutely maximizing their mitigation when Shadows are forced to reduce their mitigation and stack hp (and, for 40k hp, that's a *lot* of lost mitigation), I just can't agree with that design.

 

While the other tanks may be able to eat it to the face it may require a tank swap for Shadows. Shadows have been able to ignore tank swaps on other fights (Incinerate Armor comes to mind) so I don't find it too bad that they may require a swap where the others don't. In exchange you get the stupid easy time with Sunder, what I wouldn't give for a short CD root breaker on my Guardian just so I can spend 4 minutes running in circles bored >.< Oh how I envy you Shadows and Vanguards who get to tank him

 

You used to talk about using Power Crystals instead of Endurance. Have you gone back to Endurance ones now? Anyway, the encounter needs to be tuned for Arkanian rather than Underworld, so even if the idea is to have some degree of HP to survive it as a Shadow tank, the damage needs to be reflective of feasible HP levels at that gear tier.

 

Concerning the attacks/content design, I have to say that the only reason this is being brought up because of a confluence of 2 design constructs: spikiness *alone* isn't a bad thing in class design (as evidenced by there being largely the same level of Shadow spikiness from a raw design standpoint pre-2.0 as post-2.0 and no one really had a problem with Shadows in pre-2.0), nor is hard spike incoming damage in content design (as evidenced by how Guardians and VGs react to it). The problem only arises when *both* of those design constructs occur simultaneously. Since I doubt the developers are going to change the content design (the "easy" solution would be to turn Terminate into a 28K I/E damage attack, but Jesse Sky already pretty much dismissed that when I brought it up), the easiest solution is the change the class design.

 

To be honest, changing it to an I/E attack just makes it much worse for Vanguards since Guardian and Shadow can (bugs permitting) ignore the attack anyway. I think a more effective option would be to turn it into a Series of Shots style attack with 4 hits (maybe with increased total damage) rather than 1 big hit but then you have to rethink the knockback, which is really just an annoyance rather than of actual benefit to the encounter IMO.

 

Hopefully, the changes recommended here (either new CDs to provide some degree of burst avoidance or a change to the average mitigation profile to normalize incoming damage to some extent) will be implemented since, honestly, they could be implemented without much problem and would solve the problem *very* well.

 

I just had a thought, Kinetic Bulwark could be redesigned to affect K/E DR instead of Absorb. The effect would have to be lowered but it would suit the initial goal of Bulwark better: to reduce the effect on survivability when KW charges run out. Maybe 0.5% K/E DR per stack.

[dipstik]

Keep in mind that the 5% increase in Defense that we get is a baseline quality of being a Consular. Removing that 5% defense would necessitate either removing it from all Consulars (doubtful, since that 5% higher natural Defense chance is supposed to be part of the fundamental balance calculation to account for light armor) or having the tank stance apply a debuff to Defense chance. Conversely, returning to my original idea, doing the math for armor rating, putting CT up to 150% from 115% amounts to an increase of roughly 4% K/E DR, which is what I recommended to make up for the removal of Force Breach's 5% accuracy debuff. So, the simplest solution might just be to get rid of the accuracy debuff on FB and replace it with 4-5% DR.

 

I plugged the removal of 5% defense and 4% increased DR to the simulator and ran it through the same parameters as before (100 iterations, 400 GCDs, 5000 basic M/R+K/E per GCD, 50000 F/T+K/E spike per 40 GCDs, 4000 heal/GCD, 8000 spike heal w/ 6 GCD downtime) and it turned out a lot better. The average lowest hp was still *really* low (1204.0), but there weren't any deaths. For comparison's sake, Guardian numbers were 0 deaths with an average lowest of 7.6k.

 

okay so 5% damage reduction instead of accuracy debuff (that trades def for dr as we wanted)...

 

still need a solution for going down so low in health though. maybe a free off GCD self heal like the sages have... or something that absorbs when a single ability's damage spikes over a certain amount... we are pretty close but going from 35k to 1k where a jugg goes from 32k to 5k is not balanced.

 

that damage reduction bulwark sounds pretty interesting. need to see how broken that is, and family is in town. ttyl

Edited May 25, 2013 by dipstik