Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

Ideal Tank Stat Distribution in 2.0

STAR WARS: The Old Republic > English > Classes > Roles > Tanking
Ideal Tank Stat Distribution in 2.0

elidion's Avatar


elidion
11.01.2013 , 10:44 AM | #961
Quote: Originally Posted by dipstik View Post
I noticed that the other day, the Assassins and Shadows especially, on your profile linked for the shadow Double Bladed-Saber Defense is only counting as 1% instead of 4%, the assassin equivalent works but on both assassin and shadow the 2% defense from shadowsight doesn't count. They should have 6% from skills and buffs, not 1%.

KeyboardNinja's Avatar


KeyboardNinja
11.01.2013 , 11:16 AM | #962
Quote: Originally Posted by mongoosee View Post
Thx for the info. Here is what I see for the PT tank that you posted.
Defense: 517 = 16%
Shield: 1228 = 43.96%
Absorption: 976 = 45.20%
Total Stat = 2721.

Now compare those numbers with {2700,{defense->1035,shield->1023,absorb->642}}. They are way off.
What I'm confused is that the math says more Def and less shield/absorption. However, most folks are still doing less Defense and more shield/abs.
Powertech percentages computed from the optimum stat ratings at 2700: Defense = 21.04%, Shield = 42.91%, Absorb = 41.39%.

A lot of people do under-stat on defense. I hate to be elitist, but those people are wrong if their intention is to gear for least damage taken. A lot of people (especially vanguards/powertechs) are stuck in a very pre-2.0 mindset where defense was nearly useless for them. That simply isn't the case anymore. The vanguard/powertech tree got significantly redesigned in 2.0, and the shifting of bonuses from shield/absorb to defense was more than enough to push for a more balanced distribution of stats. It was a good change, I think, but some people are slow to adjust.

There is also an additional class of people who under-stat on defense because they think it improves the smoothness of their mitigation. This is sort of true, to a point, but because defense is weighed against absorb (not shield), it means that they're improving their spikiness not by increasing the probability of mitigation but by *decreasing* the probability of mitigation. It's effectively like smoothing out a mountain range not by filling in the valleys but by grinding down the mountains. The net result is better smoothness, but at a rather unfortunate cost.

My sincere advice would be to trust the math. Elidon, dipstik, thok, myself and more have been working on this stuff for quite a while. We've collectively put in a lot more effort into ensuring these numbers are accurate mathematically (and using them in the field!) than anyone else in the game. I'm very confident in their near-optimality.
Computer Programmer. Theory Crafter. Dragonslayer on The Ebon Hawk.
Tam (shadow tank) Tov-ren (commando healer) Aveo (combat sentinel) Nimri (df scoundrel)
Averith (hybrid sniper) Alish (lightning sorcerer) Aresham (jugg tank) Effek (ap powertech)

dipstik's Avatar


dipstik
11.01.2013 , 11:17 AM | #963
Quote: Originally Posted by mongoosee View Post
Thx for the info. Here is what I see for the PT tank that you posted.
Defense: 517 = 16%
Shield: 1228 = 43.96%
Absorption: 976 = 45.20%
Total Stat = 2721.

Now compare those numbers with {2700,{defense->1035,shield->1023,absorb->642}}. They are way off.
What I'm confused is that the math says more Def and less shield/absorption. However, most folks are still doing less Defense and more shield/abs.
i was shooting for soething closer to:

{2700,{defense->793,shield->1150,absorb->757}}

which is for the average. i have:

2700 825 1133 741
and
2700 396 1360 944

so those are the points i was trying to balance between. for 2682 i get:

shadow:
526
940
1216

jugg:
1057
940
685

PT:
613
1236
833

they have been updated

elidion's Avatar


elidion
11.01.2013 , 04:27 PM | #964
I'll have to see what I come up with.

elidion's Avatar


elidion
11.02.2013 , 12:04 AM | #965
Okay KBN, I've been trying to see where some of our differences might be coming from and I see a few things that could be related, although, our calculations are exactly the same for Guardians/Juggernauts and there is still some difference, which probably has to do with the way I generate the numbers.

I first verified my mrke and ftke values matched what you have in your formulas, I actually rounded up, but just to be precise I followed your double digits after decimal. I then verified that we were both calculating mrke and ftke damage correctly, taking into account the 2 roll system ect. And found my numbers matched what your numbers said based on the formula you posted.

Next was the base values, and that is where I see a bit of difference. As I said, the Guardians/Juggs seem to be the same, Shadow/Assassin are close, only difference is your calculation for absorb bonus from dark bulwark is slightly higher than mine, but they were still close. Vanguard/powertech is quite a bit different though and so this is where the big shift changes I think when i've been comparing numbers (since powertech is my main), so I wanted to pick your brain and see why you did what you did. Maybe I'm overthinking it.

Here is the relevant part in your base calculations.
Code:
vanguardBase:={name->"vanguard",defenseAdd->0.06,defenseBonus->0.075,shieldAdd->0.17,shieldBonus->0.0,absorbAdd->0.04,absorbBonus->0.1633,resist->0.02
So looking here I see you are adding 0.075 defense because of Oil Slick (Riot Gas for you pubs), which is cool, it's pretty straight forward, I came to the same conclusion, 30% for 15 seconds every 60 seconds, averages 7.5%. Now where this gets me though, is you also have the 0.02 for resist, which tells me you are assuming they have the 4 piece set bonus. If they have the 4 piece, that means they should have the 2 piece as well, which gives Oil Slick/Riot Gas 2 more seconds, which brings it to 30% for 17 seconds every 60 seconds, average of 8.5%.

Next is the Absorb Bonus, which I assume is from Heat Screen & Heat Blast (Power Screen & Energy Blast for pubs) I did a bit of research on this and saw you and Dipstik had different theories on this...and so well, here's mine. Anyway, the way it works is you generate up to 3 Heat Screens, each of the screens gives 1% absorb for 20 seconds. Once you have 3 stacks, you can spend them to activate heat blast, which grants you 25% absorb for 6 seconds, on a 15 second cooldown, however, each stack of heat screen lowers this cooldown by 1 second, so you can pretty much use it at best about every 12 seconds. So I personally count the heat blast as 12.5% (25% for 6 seconds every 12 seconds) Now the heat screen is tricky since it is consumed when you use heat blast.

Technically you wouldn't get the 3% benefit for long because you should be using it as soon as you get the 3rd stack. I've searched my own logs and see that I average a heat screen every 4 seconds during most boss fights. I average them out to 1.25% bonus absorb. How you get 16.33% has me stumped, even if I did 12.5+3 it would still only be 15.5, to get close to your number I have to average 3 seconds per heat screen with a +3% constant, not accounting for them being consumed.

elidion's Avatar


elidion
11.04.2013 , 11:39 AM | #966
Over the weekend I was testing my heat blast. I ran a few ops and flashpoints. And even went into a 55HM flashpoint with only a healer to let me get a good amount of time on the bosses. I also bound Heat blast to my space bar which allowed me to continuously press it and not even have to think about it. I used rocketpunch(stockstrike) on cooldown since that always triggers 1 stack. I tried to pay attention to the cooldowns, and frequency of gaining stacks.

There were a couple things I noticed. The tooltip on the heat screen talent does say that the effect can only happen once every 1.5 seconds, so at BEST, assuming you are extremely lucky and got it instantly, it would be 6 applications of the screen in 9 seconds. Taking into account shield chance a PT would have, the 15% chance on attacks and the 100% chance on rocketpunch/stockstrike with the 7.5 second cooldown, you have an average of 1 heat screen every 1.9 seconds after the initial 1.5 second cooldown, which averages out 1 screen per 3.4 seconds. But that means you would have got 3 applications of heat screen in 10.2 seconds, which lowered the cooldown to 12 seconds, meaning from a theorycraft perspective, every 12 seconds would be the average (12.5 absorb). When you add the average absorb from screens over 12 seconds (1.25 absorb), you get a theoretical average of 13.75%.

Now that I've said all that, from my observations, I think there might be a bug. The tooltip implies the cooldown keeps going down for each stack or "refresh" of the heat screen. But it seemed to me that once you have the max 3 stacks the cooldown doesn't lower anymore. which means, 12 seconds would be the lowest the cooldown could go to.That brings me to my testing this weekend with both Operations and flashpoints. And my absolute best parse was when I was with a single healer in Athiss 55 HM, on the 2nd boss that spawns the 4 adds every 30 seconds. I figure it was my best parse because with 5 things hitting me, I got a lot more shields in, which procs my heat screens. I can pretty much guarantee I was at 3 stacks of heat screen before the the cooldown of heat blast was up. And with my hammering away at my spacebar which was bound to heatblast (and heat blast being off the GCD) it was fired near instantaneously each time. in the end, the fight was 307 seconds long And I had 25 heat blasts, which 300/25 is 12, so based on my tests and personal experiences over the last few months, I'm going to go on 12 seconds being the absolute minimum on the cooldown.

Marb's Avatar


Marb
11.05.2013 , 06:58 AM | #967
Quote: Originally Posted by KeyboardNinja View Post
I'd love to see a buff to the shield specialization skill. I think this would resolve the problem handily. Removing the threat loss from Enraged Defense would also be fantastic, and would represent a really nice cooldown in general and a neat skill element (sacrifice 6 rage for healing?). Sonic Barrier really should scale with damage, and this would resolve some of the issues on hard-hitting content, but that's a much larger change.
I was playing around with the idea of just removing the damage absorb shield from sonic barrier, replacing it with a temporary buff to shield and absorption.

It removes a bit of the flavor from the class and thematically conflicts with the vanguard, but with the changes coming to shadows, guardians may start resenting the non scaling implementation of sonic barrier.

I know I will.
Zerkhan - Harbinger

MGNMTTRN's Avatar


MGNMTTRN
11.05.2013 , 01:28 PM | #968
RE heat blast: it doesn't cap at a 12 second effective cooldown, you can still get cooldown reductions when you have 3 stacks of heat screen.

Immediately after you fire energy blast you could rebuild your heat screens, so you can get a maximum of 28% absorb from heat effects and +4% absorb passive from your spec. Stats modelers tended to instead calculate abs bonuses as +25 for 6/cooldown and +1/+2/+3 absorb for (cooldown-6)/cooldown time, which was wrong since it's assuming that you could not have both +25% absorb and +1/+2/+3 absorb at the same time.

Trying to statistically model the eblast cooldown becomes an overfit/ornate problem because with effects like 30 second shield/absorb relics, different hit chances on various attacks (basic/'melee attack', other melee attacks, force/tech attacks), and the tendency for each attack to be composed of several smaller attacks, the amount of calculation needed to model the actual cooldown exceeds the amount of work needed to pull the +25% effect from combat logs. The +1/+2/+3 effects are not worthwhile to try to parse or model closely.

There was a good exchange of data in this thread. In that thread the final conclusion was assume 20.5% absorb over time, backed by actual data.

elidion's Avatar


elidion
11.05.2013 , 11:32 PM | #969
Well mine was backed by actual data as well. Aside from my personal observations and logs, it's quite simple, lets take a boss hitting you once per second over 100 seconds, so 100 hits. Let's say I have a 33% chance to shield, that means I get 33 heat screens from that. I also get rocket punch on cooldown every 7.5 seconds, that's another 13 screens. Now add the 15% chance from other attacks (about 66 GCDs minus the 13 from rocket punch in 100 seconds), that gives us 7 more stacks for a total of 53 screens in 100 seconds. That gives almost an average of a screen every 2 seconds, it rounds to 1.9 seconds. Again this is the average, this isn't minimum or maximum. What there is though is a 1.5 second cooldown, which means on "average" you will get a screen every 3.4 seconds. Using that average I can calculate 0 screens for 3.4 seconds, 1 screen from that to 6.8 seconds, 2 screens from that to 10.2 seconds, and 3 for the rest.

I agree with you that everything depends on cooldowns, shield chance, and even how many mobs you have on you. And for the purpose of boss gearing that this thread is all about we're not talking into account said cooldowns, and we certainly don't have 9 mobs beating on us like the the video you linked me to trigger an enormous amount of shields. While it is completely possible to get it as low as 9 seconds we are talking averages here. Also in a fight like Dashroode where there are tank swaps, it's unfair to discount the time it took to build up screens during the swap. it seems like people on the thread started counting at the first blast.

See if I was to go by my 12 second average in 96 seconds, I would blast 8 times. But if I started counting my first blast at 0, that means I would blast 9 times...that gives an average of 10.66, pretty much the 10.7 people were saying that tank was averaging...because they didn't count times he wasn't tanking. That doesn't give you true numbers. We run stuff without tank swaps, so I am getting beat all the time and my personal best is a blast every 12 seconds on average. Even the guy that posted torparse of his HM fights said he had heat blast up 47% of the time...which is pretty close to the 50% that a 6 second effect every 12 seconds would be. I don't see any data that supports 20.5%, in fact I would go as far as saying that 20.5 is completely impossible. Assuming you magically started with the +3% absorb from three stacks of screens (which is impossible on its own) that brings us to 17.5% from the actual blast part. With a 1.5 second cooldown on a screen, you can get 6 stacks in 9 seconds, and 15 second cooldown - 6 stacks = 9 second cooldown...meaning 9 seconds the absolute minimum. at 9 seconds that brings you to 16.67%. Now if that happened, you would be at 0 stacks for 1.5 seconds, 1.5-3.0 you would be at 1 stack, 3.0-4.5 you would be at 2 stacks, 4.5-9 you would be at 3 stacks. Do the math, that averages 2% over the 9 seconds. So at BEST, the maximum absorb you can get from blast/screens is 18.67%(16.67 from blast + 2% average from screens). To get 17.5% you would need to get a blast every 8.5 seconds, which means you would have an uptime of 70.5% of the time as well as circumventing the internal cooldown. I challenge you to find me a log that has a blast uptime over 70%.

MGNMTTRN's Avatar


MGNMTTRN
11.06.2013 , 01:45 PM | #970
Quote: Originally Posted by elidion View Post
Well mine was backed by actual data as well. Aside from my personal observations and logs, it's quite simple, lets take a boss hitting you once per second over 100 seconds, so 100 hits.
All enemies tend to attack once every 1.5 seconds.
Quote: Originally Posted by elidion View Post
Let's say I have a 33% chance to shield, that means I get 33 heat screens from that. I also get rocket punch on cooldown every 7.5 seconds, that's another 13 screens.
If your rocket punch comes off cooldown naturally and then you get a shield proc to reset the cooldown, you can actually rocket punch more than once every 7.5 seconds
Quote: Originally Posted by elidion View Post

Now add the 15% chance from other attacks (about 66 GCDs minus the 13 from rocket punch in 100 seconds),
Only ranged attacks, Rail shot, and Rocket Punch have a chance to proc heat screens and heat blast cooldowns reductions. The only ranged attack that matters to powertechs will be your autoattack, since Unload doesn't get pushback protection.
Quote: Originally Posted by elidion View Post
See if I was to go by my 12 second average in 96 seconds, I would blast 8 times. But if I started counting my first blast at 0, that means I would blast 9 times...that gives an average of 10.66, pretty much the 10.7 people were saying that tank was averaging...because they didn't count times he wasn't tanking. That doesn't give you true numbers. We run stuff without tank swaps, so I am getting beat all the time and my personal best is a blast every 12 seconds on average. Even the guy that posted torparse of his HM fights said he had heat blast up 47% of the time...which is pretty close to the 50% that a 6 second effect every 12 seconds would be. I don't see any data that supports 20.5%, in fact I would go as far as saying that 20.5 is completely impossible. Assuming you magically started with the +3% absorb from three stacks of screens (which is impossible on its own) that brings us to 17.5% from the actual blast part. With a 1.5 second cooldown on a screen, you can get 6 stacks in 9 seconds, and 15 second cooldown - 6 stacks = 9 second cooldown...meaning 9 seconds the absolute minimum. at 9 seconds that brings you to 16.67%. Now if that happened, you would be at 0 stacks for 1.5 seconds, 1.5-3.0 you would be at 1 stack, 3.0-4.5 you would be at 2 stacks, 4.5-9 you would be at 3 stacks. Do the math, that averages 2% over the 9 seconds. So at BEST, the maximum absorb you can get from blast/screens is 18.67%(16.67 from blast + 2% average from screens). To get 17.5% you would need to get a blast every 8.5 seconds, which means you would have an uptime of 70.5% of the time as well as circumventing the internal cooldown. I challenge you to find me a log that has a blast uptime over 70%.
Here's how they broke it down:
PTs get +4 passive absorb
Presume a 0.5 uptime on energy blast/heat shield
Presume that the PT spends 25% of his time at +0 absorb, 25% at +1, 25% at +2, and 25% of his time at +3 absorb
0.04*1 + 0.5*0.25 + 0.25*(0+0.01+0.02+0.03) = 0.1800. And in my defense, I did think 50% uptime was too much and even supported it with data. But Kitru's 'math' indicated it had a 10.3 second effective cooldown, and he shouted louder, so we had to compromise.