In your personal opinion...

Most of the analysis here is just comparing the primary powersets, which revolves around things like defense, resistance, regen, and hitpoints. However, things not taken into account are things like end drains, slows, and -dam (those 3 things Ice armor is capable of, too).

I love my Icer and accept it's shortcomings in some content (pure psionics, toxic, and buffed AVs). Could anyone weigh in on the effects at least Chilling Embrace brings to the table? 32% recharge and 14% damage debuffs (to those that can't resist it) seems pretty significant to the dps that can be sustained by the set. It would be pretty awesome if the -damage portion of CE was unresistable. If I could change 2 things with the set, it'd be making CE unresistable and removing that blasted nophase from Hibernate.

Speaking of which, why was the no phase added to Hibernate? I want to think it was something that had to do with pvp, but I can't remember.

No, I totally agree that Stoney's are better. In fact, I put in my stoney's numbers and they beat the Dark, if I factor in my aid-self numbers (44% every 5.8 seconds), despite the fact that he isn't optimized for defense, but more for speed. If I put aid-self on the dark, he loses on the defense cap which will make aid-self much harder to use, considering it's interruptible nature.

So what I am saying is that this would clearly put Dark as the *next* most survivable tanker to Stoney? KB prot can be built in easily, ddr would still be an issue - but other than DDR, there doesn't seem to be anything that would prevent us from placing dark as clearly the second most survivable tank?

I'd say WP is more survivable than Dark(and Inv more survivable than WP), looking at the max HP and HP/Sec numbers. But it's pretty close.

In terms of offense there's no clear victor either. A damage aura is great, but dark has some endurance issues. WP has no damage aura, but it has good end management in Quick Recovery, and relies on passive mitigation to stay alive(meaning more time to attack). I still think that the damage aura trumps over anything WP offers offensively.

I plugged the numbers in, Dark is coming out to be way ahead than WP or Invuln, any of the builds.

You're looking at the wrong builds then. Try Iggy's.

Maybe you're looking at the wrong build?

Iggy's invuln: Composite score - 1629
Iggy's WP: Composite score - 1745
My Dark build: Composite score - 3118


What other build are you talking about?

I think at this point in the discussion, the definition of the threat being survived will make the biggest difference between these primaries. Dark's greatest strength in survivability hinges on the amazing heal, which is pretty vulnerable to recharge debuffs, to which Inv is less vulnerable and WP hardly at all. Dark wil require several (let's say, three) -knb IOs to be regarded as roughly secure from knockback; you can cram one into a travel power, but the others are likely to cost you two slots. How much impact does that slot cost have?

Dark looks very good with solid resists, some controls, and the fantastic heal; defense and recharge debuffs and the slot cost of -knb offset some of that. How often do you feel like you encounter slows and debuffs? I think that might be where we are differing. Willpower suffers from defense debuffs less than dark but it still suffers; Inv suffers measurably less but still suffers some unless it consumes slotting with otherwise-not-useful excess defense. Slows hurt Dark pretty harshly, Inv a little, but Willpower essentially not at all.

Good point, clicky mitigation consumes both animation time and operator attention (decision cycles), at least to some degree.

Some very excellent points, Its definitely is important to remember that these comparsions do not take in the whole picture that could be going on while trying to tank/survive.

Control abilites are harder to factor in to surviable, and can easyly push biulds over other biulds in the real game.

Lost count the number of times I have seen high end biulds faceplant to a hidden Sapper Ambush.

Yeah, those were really dangerous to my Inv back pre-IO and especially in issue 3-4. I took CMA from 1-40 in issue 3 tanking for large teams most of the way and never faceplanted... until the first time I faced Malta and the Sappers drained me and the rest planted me in an instant. It was a somewhat unpleasant wake up call.

Now with 45% E/N defense Sappers are far less dangerous; they have to hit me in order to sap me after all. They're still a priority mob to kill first of course and it's a good idea to find the Sapper before you jump into the group.

1. Composite score is flawed as I stated. Just because content exists doesn't mean it's popular.

2. Did you take Aid Self and SoW into the calculations?

3. How many targets are surrounding, and for how long?

I find the chart Dechs has more telling.

I think you might be a bit disconnected from the rest of the thread here. Those questions have been clarified many times over in many different posts.

I would suggest reading through the thread again to get familiar with the discussion before getting into it.

I'm not sure I understand what you mean by that. Although the damage ratios used in the composite score are admittedly guestimates, you seem to be suggesting they are wildly incorrect. I would assume, if you're going to make that assertion, you have an actual counter-estimate for the relative ratios of damage a tanker is likely to see in-game without cherry picking content.


Dechs spreadsheets are actually using simpler versions of the same calculations mine uses, focusing solely on long-term survivability ("sustainable" on mine). Its also, as mine is, build-sensitive. Factoring those two out of the equation, I don't see a discrepancy between Dechs spreadsheets and mine. They sort of can't, since they are basically using the same equations (at least for the sustainable part of mine).


Edit: one difference between Dechs Dark Armor build and the slainsteel build I entered into my spreadsheet is that Dechs build seems to be slotted low for heal (Dark Regen) - which is reasonable for Dark tankers that assume they will normally be hitting multiple targets with it. But that also means on single target comparisons, it will be lower than it can be. Slainsteel's posted build slots Dark Regen for almost 70% heal on a single target, almost twice Dechs build. My spreadsheets have been assuming single targets throughout; scaled up Willpower would quickly become stronger, while most other things wouldn't. Invuln normally gets stronger, but the build we've been looking at was already soft capped.

I don't think I'd ever make that claim.

I created my spreadsheet without the knowledge of yours, Arcana. I intended to take the project a great deal farther, and would eventually have come up with the same thing you have. Realizing that, of course, I said "to hell with it," and stopped where I was at.

I can't exactly remember my slotting there. I think I prioritized endurance, recharge, accuracy, and heal in that order.

One thing I want to make sure to point out: Stainsteel may be "overhealing" to get his numbers. With Dark Armor calculations, it's important to make sure you don't heal for any more than my maximum hit points.

If you pick a specific circumstance, like say tanking the STF or tanking the ITF, you can generally narrow down the scope of the kind of threat you'd be facing. But I think its appropriate when talking generically about "strong" tanker builds that you assume the tanker won't always be able to pick the threat, and needs to be strong against all of the normal threats in standard content, and even (these days) on both the red side and the blue side. You can say cold damage is the least common (and it probably is) but does that mean you're disqualifying your tanker from tanking the Winter Lord?

At least I think its fair to assume a "strong" tanker should be able to tank for, lets say all the high level task forces that have come up as WSTs: ITF, STF, LGTF, LRSF, Apex, Tin Mage, Kahn, Barracuda. You do get a mix of damage in that content collectively.

Under those circumstances, you also get a lot of debuffs. But debuffs are a bit trickier to incorporate into a comparison, and I'm still working on that. The "DDR" row in my spreadsheet is part of those experiments to make debuff comparisons that don't make the spreadsheet too unwieldy (i.e. incredibly long).

Mine assumes only one target, so I don't have that problem specifically although its a good point: I've seen erroneous Dark Armor calculations that assume you can get 250% heal per cycle, which would be a neat trick (you can get more than 100% heal as Mids specifies it, if you have higher than normal health, since the heal is calculated as a percentage of base health).

All heals also have another problematic issue that you cannot really easily hedge with calculations: you just have to know its out there. The *maximum* effectiveness of a heal like Dark Regen can sometimes assume a 100% heal, or a heal from zero health. But you can't really do that in-game because even with perfect reflexes, if you get too close to zero health a momentary burst of damage will kill you.

Back in I5-I7ish I experimented with, but ultimately discarded, a concept called the "run line." The run line was a reserve amount of health that you couldn't touch on a player in calculations, because at that level of health you're "effectively dead" meaning while you're still alive, statistically speaking at that point the only thing keeping you alive is not defense or resistance or regen, but luck. One damage tick goes the wrong way, and you're dead. But it was likely to be at least a little arbitrary and difficult to calibrate with calculations. The *concept* is correct, though: real health staircases downward and bursts upward with heals: but when you hit zero the game is over. You can't average out death (it was called the run line because that's the point in health where, if you wanted a high percentage chance of surviving the fight, you had to temporarily run away and disengage).

The run line is an interesting concept because it wouldn't be the same for everyone: in particular it would be lower for characters with resistances. In effect, it would be trying (among other things) to account for the difference in burst vulnerability between defense and resistance.


What makes this really complex is that heals have a counter-advantage that, say, regeneration doesn't have. Heals can be "banked." In other words, if you have reconstruction on a 30 second recharge, how many times can you use it in a 5 second fight? The answer is 1/6th of the time on average, but in practical terms the answer is one, because you can always start a fight with recon recharged. That's what I mean by "banking" a heal: waiting until its recharged before diving into the next fight. You cannot bank regen or for that matter defense and resistance. But you can bank heals, which is an advantage unique to heals. It doesn't matter on average, because to bank a heal you essentially have to slow down, and slowing down reduces the threat you face on average also: the average calculations still work. But in practical terms, heals have a situational disadvantage and a situational advantage that the players can leverage, which is *incredibly* difficult to account for in calculations. Simulations can to a degree, but even there only to a certain point.

Yes and no. One complication to defense debuffs is that most of them are not autohit. So the presumed accuracy of the critters is important to determining how often they hit, and what they average effect is as a consequence. Also, one -30% debuff every minute is different from dozens of -5% debuffs every second. In particular, the latter requires more careful consideration of cascade failure, because each debuff is making the previous one easier to land. Defense debuffs weaken most characters with defense, but cascade amplification can make those not only cause more damage, but also self-amplify their own debuffing ability (by making it more likely for the debuffs themselves to land and take effect).

Still working on a model for that. The other three effects I'm looking at don't have that problem and will be ultimately easier to incorporate: -recharge, -regen, and -res. And most of those have to hit as well, making -DEF the keystone of the entire thing, since so many super-strong builds rely on soft-cap defense.

Unfortunately, while that would be an interesting learning experience in terms of observing the relative strengths of things like debuffs and foe debuffing effects, there's too many random elements to make that sort of test useful for comparisons between builds. Each run-through would be different.

Wouldn't a large enough sample of runs provide a meaningful average?

I would suggest you answer instead of telling me to read the entire thread again, as your numbers don't come close to mine. Dechs might be right about the Dark Regeneration bit, how did you calculate it?

I'm not saying they are wildly incorrect. I just don't think these particular guesstimates are a good idea for these sort of comparisons. It's pretty agreed upon that Negative Energy isn't as common as Energy, and Cold isn't as common as Fire. Also consider that TF's like Apex and Tin Mage are largely S/L/E based, and those force you into facing level 54 mobs which do more damage. Add to that the fact that defense debuffing mobs are largely S/L/E based, and that the game's hardest hitters are primarily S/L(Lord Recluse, Battle Maiden, Venged Bobcat, Goliath War Walkers).

Basically I think the 60% S/L and 40% everything else numbers not very agreeable. I personally find S/L more important than that, and the remaining damage types not even in importance. Another person might disagree with me based on his personal experiences.

Sure, but who would take dozens of builds through hundreds of runs of dozens of scenarios?

Why hundreds of runs? Because the AE itself is going to randomize the challenges, and then each challenge will have random jitter in its run, so to get a meaningful average you'd likely have to run at least a hundred runs just to begin to average them all out.

I should point out that a long time ago someone tried to estimate the strength of an SR toggle by counting hits and thought 2000 was a lot of swings. They came up with 14.6% as the value. As SR toggles are actually 13.875%, that was slightly off. It always takes more samples to generate a valid measurement than people think. I did a 50,000 swing run once and came up with 13.91%.

I could specifically make a challenge mission designed not to kill but to test survivability. But the (legitimate) quibble would be over what kinds of attacks and damage to put into it. It would have coarse resolution, meaning we would be able to say two different tankers survived a particular level of challenge but not which one did so "better" most of the time. And one tanker could be stronger than another, and yet both fail a particular challenge and both pass a different challenge just because there isn't a specific challenge that exists in between them.

I have been thinking about some standard scores, though. I think there are six specific situations that are important when considering tanker strength.

1. Average spawn size for team size four
2. Average spawn size for team size eight
3. Aggro cap of typical spawn proportions
4. Aggro cap of bosses
5. Eight AVs.
6. Aggro cap of AVs.

What are these six data points? The first one is the data point that the devs have said in the past represents what a higher level tanker ought to be able to tank on their own without significant outside assistance. If you can't tank that, something's wrong. The second one is the obvious "solo a mission scaled for full team" data point. The third is possibly identical to or related to the second one, but I threw it out there just in case research determined it was slightly different. The fourth is the obvious maximum aggro you're likely to be able to draw in a mission if you are tanking and multiple spawns are overlapping and your team kills all the minions and LTs fast. All that';s left are the bosses, and you could theoretically be facing all bosses at one some point. The fifth is the likely limit the devs would throw at a player normally - this is comparable to tanking the entire AV fight in the LRSF solo (its a little higher than that). And of course realistically, number six is as high as you can go without resorting to higher level foes. I don't think there are many tanker builds that can do that anyway so there's no need to talk about +4 x 17 AVs.

Each of those scenarios can be represented with an average incoming damage score in theory. I'm trying to figure out what those are.