Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

Tankassin, Absorption rate or Defense rate?

STAR WARS: The Old Republic > English > Classes > Shadow / Assassin
Tankassin, Absorption rate or Defense rate?

theonepanda's Avatar


theonepanda
12.25.2011 , 10:44 PM | #21
Quote: Originally Posted by Modyn View Post
Do you have any hard data to back that up? The guys at Sithwarrior have been theorycrafting this since early Beta.

The thread I originally posted was split from the original thread. The first post in the link has the rating graphs and you can find plenty of math to back it up, if you want. Obviously, this is all preliminary until there's a much larger level 50 sample size, but for where we're at now, it's about the most solid info we have.

Basically, Defense is definitely good, but you're not severely gimped if you favor another stat more than others.
I looked at the numbers given by TORHead. They simulate a fresh lvl 50 in greens.

Lets assume that no darkness tank (madness or pure) is going to shock w/o energize. Lets also assume they will not lightning w/o 3 stacks. Generously, you will get 1 useable shock every 6 seconds (thrash thrash saber shock, for ~force neutral combo). That puts getting 3 charges of the lightning thing at about 18 seconds, or 21 seconds for the full combo.

This deals 4k damage in greens, heals you for 9?% of your health. This equates to 0.4% hps. Doesn't seem terribly good, but whatever, we're concerned about damage here.

If we get even ONE Raze proc as madness (saber strike hits 3 times, thrash hits twice, and we have +9% crit from the last energize we used), we will do ~1.25x the damage of lightning, at the cost of 1 gcd as opposed to 2 gcd's spent channeling, where we cannot move, and without a force cost.

Furthermore, if we compare wither and death field, death field easily does twice the single target dps. However, I actually don't think death field has any place in a single target rotation (similar, perhaps slightly more damage than thrash, w/o any of the procs or crit bonuses).

Because there are TWO rolls (first def, then shield), any def you get will diminish shield. And vice versa. Thus spreading your stats results in the LEAST efficient defense.

Contrary to popular belief, stacking def is more likely to have you negate a hit than stacking absorb.

Obviously stacking shield is the greatest chance to negate a hit, but not by as much as def (100%) or by stacking absorb (up to 74%).

So...basically 3 choices here.

1. Max defense. Moderate high chance of negating full hit. Moderate high total mitigation.
2. Max shield. Very high chance of negating full hit. Very low total mitigation.
3. Max absorb, overflow into shield. Low chance of negating full hit. Very high total mitigation.

As you can see, maxing shield and maxing absorb + overflow are complete opposites in terms of how they mitigate. Defense is a safer middle ground, and better than absorb until you can max it (~500+ rating).

Wootle's Avatar


Wootle
12.26.2011 , 06:30 AM | #22
asdf misread.

theonepanda's Avatar


theonepanda
12.26.2011 , 07:18 AM | #23
Man those graphs are pretty psycadelic (also like, whoa man, i had to get an account to get access? way out there duuuude).

Anyway the deciding factor on who's right is...

whether there are diminishing returns on rating to percent or not.

If there ISN'T, which is what I assumed, stacking one rating will always be better.

Why?

Well think about it this way. Going from 0% def to 1% def gives you...1% damage reduction (99%/100%). Going from 99% to 100% def gives you 100% damage reduction (0%/1%). Obviously def is worth more going from 99% to 100% than it is from 0% to 1%.

Any other value of def to def+1% is going to be somewhere between those two extremes, in increasing value per point.

Same is true for shield and absorb.

Furthermore, I simulated all 3 types of absorb.

EDIT:

"I calculate a value I call "Squishiness", which is ratio of damage taken to the damage you would take with no armor, shield or avoidance."

This is an odd metric. Armor we will hold to be constant...it is dependent solely on ilvl (er...item roll). Furthermore, Armorings have only primary stats, whether we use Force Wielder or Resolve is up to you (i pick resolve, END stacking is silly), and has no effect on mitigation.

My metrics were total damage mitigated, and chance to eat full hit.

Also I still don't see how a split would be better in any scenario (until you cap a stat). Not only does what i said before mean stacking a single stat in a vacuum would be better, but also stacking one actively makes the other worse.

Think about this. If we have 50% def, we only use shield on 50% of the hits. If we have 75% def, we only shield on 25% of the hits. This devalues shield, the more def we have.

Vice versa, if we have 50% shield, our def only mitigates 50% of the damage that would've gone though. If we have 75% shield, our def only mitigates 25% of the damage that would've gone though.

The only synergistic ratings are shield and absorb...but because absorb scales nearly twice as fast as shield, we actually have better success with total mitigation by maxing absorb and then stacking shield.

dotNex's Avatar


dotNex
12.26.2011 , 09:14 AM | #24
Quote: Originally Posted by theonepanda View Post
Also I still don't see how a split would be better in any scenario (until you cap a stat). Not only does what i said before mean stacking a single stat in a vacuum would be better, but also stacking one actively makes the other worse.

Think about this. If we have 50% def, we only use shield on 50% of the hits. If we have 75% def, we only shield on 25% of the hits. This devalues shield, the more def we have.

Vice versa, if we have 50% shield, our def only mitigates 50% of the damage that would've gone though. If we have 75% shield, our def only mitigates 25% of the damage that would've gone though.

The only synergistic ratings are shield and absorb...but because absorb scales nearly twice as fast as shield, we actually have better success with total mitigation by maxing absorb and then stacking shield.
Split is always better. Diminishing returns.

The question in our case is, when do we stop stacking attribute A and start with B (e.g. Shield, Defense etc). Maxing one thing would be terrible, you need a combination of all of them. If diminishing returns here are the same as in wow you would get r***d by them before you can hit 60% avoidance. Im pulling numbers out of my a.s.s. right now, I hope I was able to explain what I mean.



Quote:
Think about this. If we have 50% def, we only use shield on 50% of the hits. If we have 75% def, we only shield on 25% of the hits. This devalues shield, the more def we have.
While you are right here, the amount of stats needed to increase your defence from 50% to 75% will be just ridiculous. Instead, you can invest in Shield AND Absorb thus increasing your survivability / total damage mitigation alot.

I am not sure if you are forgetting about diminishing returns, but it felt like it when I was reading your post.



Regards

Pijinz's Avatar


Pijinz
12.26.2011 , 09:41 AM | #25
Quote: Originally Posted by theonepanda View Post
I looked at the numbers given by TORHead.
The Torhead numbers aren't accurate right now. They just need a bit of time to get the numbers right I think (it's a new game after all). But until then, if you're wondering, I've got a few of the formulas saved to a text file:

Saber Strike = 99% Weapon + 99% MBonus + 0 (3 attacks) (Weapon)
Thrash = 99% Weapon + 148% MBonus + 119.14 (2 attacks) (Weapon)
Shock = 185% Fbonus + 297.85 (Energy)
Force Lightning = 316% FBonus + 508.76 (3s, 4 ticks) (Energy)
Discharge (Dark) = 74% FBonus + 119.14 (AOE) (Internal)
Wither = 94% FBonus + 151.34 (AOE) (Kinetic)
Lacerate = 48% Weapon + 71% MBonus + 114.31 (AOE) (Weapon)
Assassinate = 206% Weapon + 309% MBonus + 497.49 (Weapon)
Maul = 158% Weapon + 237% MBonus + 379.96 (Weapon)

The other formulas are somewhere around the Sithwarrior forums. I've been mainly looking at the deep Darkness build, so I haven't looked much at Death Field or Raze yet. Raze will be a pain to model due to possible DoT clipping.

theonepanda's Avatar


theonepanda
12.26.2011 , 10:06 AM | #26
You don't have the numbers on Crushing Darkness, we'd need that to make a suitable comparison.

Also my point exactly was that we have no idea how the diminishing returns scale, or at what point they kick in.

Diminishing returns on WoW also just meant that since mastery did not have diminishing returns, it was the stat to stack.

Does anyone know what the diminishing returns are? Has anyone stacked something and found if the diminishing returns exist?

The only thing I read stated that it took 27.5 points to raise def 1%, 16 for shield, and 9 for absorb. It also stated the hardcaps.

If what you say is true, that there are diminishing returns, what are they? What is the formula for them? Math guestimation gets us nowhere.

The Raze proc lasts longer than 6 seconds, you never need to clip it. Of course, you run the risk of wasting a proc then. We need a good simulator to determine good times to use Energize and Raze procs, etc, and what to prioritize. My current rotation involves never using either unless i am about to melee, if I decide I need to melee, I will use Raze and then Energize. Obviously I will also use them if the corresponding buff is about to fall.

The reason being that they are not wasted if I'm using force attacks (mostly discharge), and no procs can occur during that time as well.

EDIT:

Hey look, it turns out there are diminishing returns. A spread is better then in this case. The formulas I was using are outdated.

...

Even with diminishing returns, my simulation shows stacking one stat to be the highest mitigation.

Though, the stack is turning into shield chance rather than def or absorb. Weird.

dotNex's Avatar


dotNex
12.26.2011 , 10:33 AM | #27
I think that was the point of the graphs from sithwarrior.com, but I have no idea what the exact formulas are

Slightly offtopic, have anyone figured out how is the Discharge (with Dark Charge active) aoe range calculated? Some times it feels like its just a tad larger than melee "aoe" range, other times .. I had some lightning-across-the-screen situations at least several times this week, I really hate breaking CC and this makes me slightly mad . I suspect that the center point of the AoE is actually your target, but even then it doesnt feel right.

theonepanda's Avatar


theonepanda
12.26.2011 , 10:53 AM | #28
Unless I am implementing them wrong, or these formulas are also outdated...

It is best to stack shield until something like 1000 total rating.

The formulas I found were

Defense Chance % = 5 + 30 * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( DefenseRating / max(Level,20) ) / 0.55 ) )
Shield/Glance Chance % = BonusOnSheildGenerator + 50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( GlanceRating / max(Level,20) ) / 0.32 ) )
Shield/Glance Absorption % = BonusOnShieldGenerator + 50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( AbsorbtionRating / max(Level,20) ) / 0.18 ) )

Which at 50, for an assassin tank, are

Defense Chance % = 16 + 30*(1-0.966666667^(DEF/27.5))
Shield Chance = 35 + 50*(1-0.98^(SHIELD/16))
Absorb Rating = 24 + 50*(1-0.98^(ABSORB/9))

(Hey look, thats where my numbers came from).

The diminishing returns are so minor, but the ratio on Def kills it.

EDIT: I WAS SO WRONG! IGNORE THIS POST!

Daemonlaud's Avatar


Daemonlaud
12.27.2011 , 05:54 PM | #29
Interesting discussion...but where does this leave a pvp 'tank' player?

See, you will be specced into Dark Ward, meaning that you will have around ~40% shield chance in PvP. This is pretty huge, so is it better to improve defense, which is the best defensive stat on paper, for an increase from 16% without gear up to, probably no more than 22-25% with?

Or is it better to stack Shield, since we start a huge amount of % chance, so the reward per point of stat due to increasing returns is very high?

Or is it better to stick with 40% and get absorb up from 24% to ~40%?

40% shield 40% absorb seems quite achievable without giving up any expertise or anything, what do you think?


Keep in mind that shield is slightly more attractive than in PvE due to the fact that shields seem to work while stunned or mezzed.

Kirnan's Avatar


Kirnan
12.27.2011 , 06:48 PM | #30
Probably defence for pvp.

Even if shield stacking turns out to be the most optimum damage reduction overall, you can't shield a crit. I'm not sure what kind of crit rate mobs have so it may be irrelevant there but with some decent gear and buffs it is going to be pretty high in pvp 30%+ and likely higher with trinkets/procs. At level 45 I'm already at 29% with a proc that's pretty much always up.