Reasonable yes, but not accurate. Almost every number differs from what is stated in first post. They are however in 1%-2% range (in percentage points)

As in the original post is not accurate or the new math is not accurate?

I think scalars are reasonable. They just need to be adjusted.

Yep, if you're just going from 0% armor reduction/penetration/ignorance to 20% armor rating reduction, then the scalar is (1-0.3)/(1-0.35) = 1.0769. This applies to Balance Shadows, Balance Sages, Telekinetic Sages, Watchman Sentinels, and Dirty Fighting Scoundrels.

Still looking at the Full Auto data

Still looking at the Full Auto data

Full auto benefits from 20 % armor debuff and 35 % armor penetration from cell and 30 % ignore from the Target Lock talent.

You calculate the equivalent armor rating

7538 * (1 - 0.2) * (1 - (0.35 + 0.3)) = 2111

Insert that into formula for DR

2111 / ( 2111 + 240 * 55 + 800 ) * 100 = 13.1 %

                         |  min  |  avg  |  max 
full auto tooltips | 2780 | 3030 | 3280
after 13.1% DR  | 2416 | 2633 | 2850
your parse data | 2421 | 2657 | 2844

Your parsed full auto values are within the tooltip range from what I can see.

Alright I did a new parse today since the Pyro Fix: Pyro Parse

And I'm feeling pretty confident with my new calculations for scalars, although this is subject to change if you all can find any errors : Pyro Scalar Spreadsheet

I didn't get fancy with the sub 30 burn, I just adjusted the time assuming I was solo'ing the boss with an armor debuff applied and recalculated DPS based on that adjusted time.

Boss HP = Adjusted damage done w/armor debuff applied

Total Adjusted Time = Length of parse - A + B

Adjusted DPS = Boss HP / Total Adjusted Time

*A = Length of Parse Sub 30 (in seconds) = .3 * Length of parse

*B = Sub 30 Adjusted Time = .3 * Boss HP / Sub 30 DPS

Summary: Based on this parse and the new verified (more or less) method for calculating armor debuff, the gain Pyro gets from Armor Debuff is ~4.65% and a total gain of ~6.64% when factoring in Sub 30% burn. This brings the 3550 dps parse to 3786. I think the sub 30 burn may be slightly undervalued in this model based on most group compositions, but I'd rather not complicate things...someone else is welcome to try and modify it based on whatever this thread came to an agreement on for sub 30 scalars.

I find this very believable as I ran DF HM last night. On Nefra, the new tunnel boss, I missed very few GCDs (had to cleanse myself here and there) as I was never targeted by the bomb guy and never very close to the one who was. I hit 3720, no telling for sure what that would have been on a dummy, but I felt good about my rotation in this encounter.

Full auto benefits from 20 % armor debuff and 35 % armor penetration from cell and 30 % ignore from the Target Lock talent.

 

You calculate the equivalent armor rating

7538 * (1 - 0.2) * (1 - (0.35 + 0.3)) = 2111

Insert that into formula for DR

2111 / ( 2111 + 240 * 55 + 800 ) * 100 = 13.1 %

                         |  min  |  avg  |  max 
full auto tooltips | 2780 | 3030 | 3280
after 13.1% DR  | 2416 | 2633 | 2850
your parse data | 2421 | 2657 | 2844

 

Your parsed full auto values are within the tooltip range from what I can see.

Ahh okay I was getting stuck at 1-2657/3030 = 0.123 and trying to understand why 0.123 and 0.131 were so different after 100 samples, but looking at it like that that difference seems pretty reasonable.

Similarly the VG hib data came out like this:

Average damage for non-crit HIB samples = 1874. Base 7538 armor * (1-0.45-0.3) = 1884, DR from armor = 0.118, spread 1995-2265 gives center 2130, 2130*(1-0.118) = 1878!

I'm going to skip parsing a Commando 'Gunnery' skill tree that foregoes Demo Round in favor of HIB arpen because the numbers do look reasonable.

Here's a table of dummy-to-boss scalars (which you can apply to Torparse logs) and Tooltip scalars (which you can apply to the average of a tooltip's reported damage). Combat sentinels and Scrapper Scoundrels were left undone; Combat Sent should be pretty simple if you assume base DPS done during the 100% arpen window is the same as base DPS done outside the arpen window and if you know the effective cooldown on Precision Slash; then you can figure out how much damage was done during non-arpen window and increase it by 1.075. Of course the assumption that base DPS done during the 100% arpen window is same as base DPS done outside the window is completely false, but whatever.

I don't know the Scrapper arpen buff's uptime so I couldn't do a table for it. EDIT: in response to Vai's post below this one, if the Scrapper Scoundrel gets 30% arpen for 100% of the time then they have the same arpen as a Focus Sentinel. That's a 1.0579 dummy-to-boss scalar at all times.

For ACs with varying armor penetrations, like Gunnery/Assault Commando, you can do a dot of scalar i * damage done by attack i. So for example if an Assault Commando parses 20k HIB damage (scalar 1.0367), 30k Full Auto damage (scalar 1.058) and 30k Hammershot damage (scalar 1.075) then you could expect him to do ~84724 damage on a boss instead of the 80k that you parsed on a dummy.

I forgot to do a column for Gunslinger Illegal Mods. Oh well.

Here's a table of dummy-to-boss scalars (which you can apply to Torparse logs) and Tooltip scalars (which you can apply to the average of a tooltip's reported damage).

Nice work, thanks

There are a ton of replies and this is a new thread to me, so I was hoping to pop a question or 2 in here and bypass the 14 pages.

I am a gunnery commando on Begeren Colony. I noticed that for gunnery commandos the adjustment is 0%. I do believe that our in-fight dps will be higher. See below.

1. Sentinels inspiration

2. Any other armor debuff

3. Combat dummies don't move, therefor our electro nets do less damage

If these have already been accounted for then I guess I just need to try harder.

P.S. I like that you guys segregate those darn sabo's

1. Inspiration isn't counted for any spec, so it isn't for Gunnery either.

2. Armor debuffs don't stack since 1.3 (or 1.2?)

3. Electro net works different in PVP and PVE. in PVP it stacks up to 10 but only if the target moves, in PVE it stacks up to 5 on it's own, no movement required.

1. Inspiration isn't counted for any spec, so it isn't for Gunnery either.

2. Armor debuffs don't stack since 1.3 (or 1.2?)

3. Electro net works different in PVP and PVE. in PVP it stacks up to 10 but only if the target moves, in PVE it stacks up to 5 on it's own, no movement required.

1) Sounds fair

2) I was more curious about other classes armor debuffs that would take precedence over grav round because of the higher armor reduction. I am not an expert on all classes by any means but if I do recall correctly there is a class that has a 35% or 40% armor reducing debuff which is better than grav round's 20%. But... after researching I have found that everyone is equal across the board of 20%.

3) Interesting.... I will test this. The skill does not specify "up to 5 on it's own, no movement required"

Come on man...! Throw gunnery a bone! Let me get like 2%....

Nearly all DPS specs gain some advantage from raid buffs and/or execute phases on a boss (Gunnery Commandos / Arsenal Mercs are the notable counter-example). These advantages are not reflected in a dummy parse, which means that class balance will dictate that these classes parse (on a dummy!) somewhat lower than other classes will. The most notable example of this is Shadow/Assassin DPS, which benefits substantially from both an armor debuff and an execute phase.

 

The "DPS Scalar" referenced in the title of this thread is simply the percentage increase seen by a particular damage dealing spec when moving from a dummy to a boss. Literally, it is a number which can be multiplied into a player's dummy DPS to predict how well they would do on a static boss with an armor debuff and an execute phase. This information is useful from a theory crafting perspective, as well as useful to raid leaders everywhere who need to correctly evaluate which players will do the most damage in actual boss fights based on their dummy parses.

 

For example, if the scalar for a particular spec is 7%, and a player of that spec parses a 2900 on a dummy, you can expect them to parse 2900 * (1 + 0.07) = 3103 DPS on a real boss.

 

This is a collaborative thread! There are a lot of DPS specs in the game, and it's going to take me quite a while to get through all of them. Also, I've been known to make mistakes, and getting extra eyes on my math never hurts. If you see a damage dealing spec which doesn't have an associated scalar in the list below, feel free to grab a top-tier dummy parse from any one of the leaderboards, your local copy of torhead, and math out the exact DPS jump that class will see on a boss. Please include the mathematics which justifies your result. Don't just throw a percentage at me and walk away. :-)

 

Listings

 

Advanced Classes are listed alphabetically by their Republic mirror, with each accepted spec archetype itemized under the sub-heading.

 

Commando/Mercenary

 

• Gunnery / Arsenal - 0.0%
  
• Assault / Pyro - 10.1%
  

 

Guardian/Juggernaut

 

• Vigilance / Vengeance - 0.56%
  
• Focus / Rage - 3.1%
  

 

Gunslinger/Sniper

 

(note: Saboteur/Engineering is omitted due to the difficulty of reliably calculating the DPS effect of the roll on a real boss)

 

• Sharpshooter / Marksman - 2.3%
  
• Hybrid - 5.24%
  
• Dirty Fighting / Lethality - 5.0%
  

 

Sage/Sorcerer

 

• Telekinetics / Lightning - 7.3%
  
• Balance / Madness - 6.7%
  

 

Scoundrel/Operative

 

• Scrapper / Concealment - 6.2%
  
• Dirty Fighting / Lethality - 6.5%
  

 

Sentinel/Marauder

 

• Watchman / Annihilation - 7.1%
  
• Combat / Carnage - 3.8%
  
• Focus / Rage - 12.0%
  

 

Shadow/Assassin

 

• Hybrid - 13.7%
  
• Infiltration / Deception - 11.3%
  
• Balance / Madness - 15.1%
  

 

Vanguard/Powertech

 

• Tactics / Advanced Prototype - 3.6%
  
• Hybrid - 3.6%
  
• Assault / Pyro - 8.2%
  

This would be an interesting thread to wave under the noses of the devs in my quest to get vigilance guardian some love. Oh, but that MS root, yeah, that's going to be huge. <eyeroll>

YES!!!!!!

http://imgur.com/a/VVX29

YES!!!!!!

 

http://imgur.com/a/VVX29

so, this thread is essentially pointless now. nice

so, this thread is essentially pointless now. nice

And its for the best. No more complex maths and no more BS by devs. Just pure DPS (obv taking into account players of equal skill and gear.)

Inspiration still won't be accounted for but that should not matter too much.

And its for the best. No more complex maths and no more BS by devs. Just pure DPS (obv taking into account players of equal skill and gear.)

 

Inspiration still won't be accounted for but that should not matter too much.

Indeed, and honestly, I find it perfectly fair that the only class able to provide that buff gets to apply it to themselves during parsing, to account for the value added by it to the rest of the raid.

so, this thread is essentially pointless now. nice

This thread was pointless from the moment that KBN decided to use the wrong armor penetration formula... so from its time of posting. I tried to turn it around but he has tactically ignored our work.

The calculations are fairly easy to verify in part. Once we're done with the Vanguard/Powertech numbers, it will be pretty easy to verify their value quite precisely simply by parsing with an armor debuff. In any case…

 

If you would like to point out specific flaws in the math we're using, please feel free. That's why this is a thread. Otherwise, posting that you're considering all this work invalid until you see combat log verification is extremely counter-productive and borderline ignorant. All theory crafting proceeds in the manner we have followed, and it does so because statistical models are much smoother and produce far more precise results than practical testing. Of course, if theory crafting fails to predict actual bosses, then naturally the work is useless, but the point of good theory crafting is to account for all factors in a smooth, stochastic environment. If you can show that we have failed to account for the factors involved in a statistically sound manner, then please do so and we will adjust the work. Otherwise, randomly sniping the math just because it's math is not welcome.

-_-

Not even close.

Although he MAY have gotten the armor debuff formula wrong, he touched on many points that were right.

Shadows, are STILL underpowered, but their best players are getting to 3.4

Vigilance guardians, are still around the same place, and their execute is crap. They are sitting around 3.3-3.4, once Kisheksun gets his crap together and gets us a good proc fished parse

Pyro/Assault has shown that its wayyyy ahead and I believe there is a 4k parse now.

I feel like class balance is looking a lot worse than before now.

The next class rep round will be interesting. I have a lot of questions. Guardians need their buff. We have a **** execute.

Oh and the 500k dummy doesn't count. Thats like parsing on the old dummy for 1 minute and calling it good representation of dps.

1.5 million is where its at.

so with 20% armor debuff and 500k HPs applyed at dummy all classes are at the same lvl? (without counting ispiration)

Falver, the organiser of this thread, has decided to set the standard of 1 million with armour debuff to compare the dps between classes, so that's where you should look to see how the various classes will do. So far merc pyros (and mandos, I'd presume, but no parses from them yet) are the undisputed winners, having already reached 4k dps. Hybrid slingers continue with their high numbers. Annihilation marauders, lethality operatives and lightning sorcerers are seeing some positive, and significant, changes. Carnage marauders are doing better, but not the the extent of the previously mentioned specs. Same with marksman snipers, it seems. Although there's only really one assassin/shadow dps posting some really good numbers, he has proven what numbers they can achieve. Guardians/Juggs will probably see a small change due to execute phase, whereas arsenal mercs will remain unchanged. Not seen vanguard/powertech parses yet (be it hybrid or not). Same with madness sorc.

My overall impression is that melee is not where they should be, given the fact that they should generally doing more damage due to having to deal with more mechanics and movement. DF and DP certainly do not deviate from the melee unfriendly attitude. Fact of the matter is, however, that it is the ranged classes generally dominating the leaderboards. Annihilation may compete with MM snipers/slingers and lightning sorcs, but there's no chance they will be able to compete with pyro mercs.

Obviously there will be some differences in actual boss fights, but on average the ranged classes reign supreme. There's obviously players that can deal significant damage as melee, but in a lot of fights they have less uptime than ranged. Knight/warrior is the only class with a charge ability that closes distance. Socundrel/operative can do it, but at a significant energy cost. Shadow/sin can use force speed, but not nearly as effective, in my opinion. So in general I feel BW have some serious balance issues to correct. Dummy parsing should heavily favour all melee classes over ranged, not the opposite.

dps of infiltration doesn't mean so much, I mean the same player can do 3300/3200/3100 dps, depending by lucky, just luck.

I'm min/maxed and I do perfectly the rotation but just 3145 for noe, I can take over 3200/3300 but I get so much boring at dummy lol and I don't try mayn times.

infiltrtion is just a joke.

Max dps www.youtube.com/watch?v=IgU0X73

Do not click the above link. It is SWTOR referral program fraud.

Also I want to take this time to alert everyone of a forum rule. Due to the possible "trick" nature of using your referral link to trick someone into unintentionally becoming your referral, we are not allowing use of your referral links here on the forums until this issue is resolved. Whether that is being honest about what the link is, or trying to hide it, it will be removed. Any egregious breach of this rule can result in infraction points.