Tank damage WAR vs PLD testing data and parser

I am in the midst of a conversation on another tank thread about damage on war vs pld and to help prevent theory craft from overriding empirical data I decided to do some real work and get evidence of the truth in the argument of war vs pld in damage.

Stats
Both tanks are in full Darklight and same jewelry so no armor is effected in this
Both tanks use their relic weapons
Both tanks have +8 str and +22 vit

This is as close as I can get to them being the same.

Because I didn't want to write a interactive scraper and I wanted to get it done quickly I decided to use python to parse the log data to do my calculations.
I supplied my parsing script so others can do the same test. If you are not aware how to use it google python scripting language.
I could have made this cleaner but meh I just wanted the results.
http://pastebin.com/hvp02wGh

The 1 nuance to this is CoS, The FF devs decided Dots do not need to be added to the log window so I did a simple calculation to add this as follows. The calculation does not account for crits but this should be so small it should not effect the overall data.
CoS hits for 26-29 making the median 27.5
CoS ticks 5 times making 27.5*5= 137.5
I simple detected if CoS was cast and if it was I split the damage of CoS between the current rotation and the next rotation (+68.75 per) with a 25 second cooldown this should never cause an issue of collision and should average out fine.
In reality some things like the opening damage should be higher in the calculation but again I didn't want to write a memory scraper and reverse some of FFXIV I just wanted a decent set of emperical data.

One thing to understand here is this is a comparison between War and PLD and does not attempt to max DPS. Because of this I left out any overlapping abilities from the rotation. For example Fracture would overlap and removing it from both will nullify the damage it would cause.

I did use the Damage CDs as soon as they came up and ran through 2 full cycles of CD on the damage buffs, berserk and fight or flight. Also warrior gets Internal Release.

I added the logs of the fights here
Warrior : http://pastebin.com/gdPGmWYU
Paladin : http://pastebin.com/8CJdje96

I used rotations to benchmark the damage instead of a damage over time calculator. This would have been much more complicated and with the mechanics of FFXIV not a noticeable difference.

Rotations
PLD with sword oath
Fast Blade->Savage Blade->Rage of Halone
I mixed in Circle of Scorn and Spirits Within the moment they came up

Warrior without Defiance
Heavy Swing->Maim->Storm's Eye->Heavy Swing->Skull Sunder->Butcher's Block
I mixed in Brutal Swing as soon as it was up


So to the conclusion.
Output from both parsed.

WAR OVERVIEW

>>> Rotation damage = 1337
>>> Rotation damage = 1546
>>> Rotation damage = 1996
>>> Rotation damage = 1311
>>> Rotation damage = 1020
>>> Rotation damage = 1223
>>> Rotation damage = 1231
>>> Rotation damage = 1369
>>> Rotation damage = 1052
>>> Rotation damage = 1084
>>> Rotation damage = 1270
>>> Rotation damage = 1094
>>> Rotation damage = 1606
>>> Rotation damage = 1789
>>> Rotation damage = 1545
>>> Rotation damage = 1612
>>> Rotation damage = 1034
>>> Rotation damage = 1090
>>> Rotation damage = 1181
>>> Rotation damage = 1251
>>> Rotation damage = 1086
>>> Rotation damage = 1562
>>> Rotation damage = 1104
>>> Rotation damage = 1228
Median = 1241.0
Average is =1317
total = 24


PLD OVERVIEW

>>> Rotation damage = 1591.75
>>> Rotation damage = 1255.75
>>> Rotation damage = 1293
>>> Rotation damage = 1519.75
>>> Rotation damage = 1390.75
>>> Rotation damage = 1050
>>> Rotation damage = 944
>>> Rotation damage = 1075.75
>>> Rotation damage = 1344.75
>>> Rotation damage = 1109
>>> Rotation damage = 1085.75
>>> Rotation damage = 990.75
>>> Rotation damage = 1442
>>> Rotation damage = 1591.75
>>> Rotation damage = 1273.75
>>> Rotation damage = 1204
>>> Rotation damage = 1192
>>> Rotation damage = 1104.75
>>> Rotation damage = 908.75
>>> Rotation damage = 911
>>> Rotation damage = 1605.75
>>> Rotation damage = 1011.75
>>> Rotation damage = 885
>>> Rotation damage = 921
Median = 1150.5
Average is = 1195.9375
total rotations = 24

There is 1 outlier here, the first part of the rotation you can see the warrior crit and the pld never crit. The conclusion still remains the same over 10 different comparisons. It was close but War was always slightly higher damage in long run than pld.

The 1 thing that did change was 1st and 2nd rotation damage. Paladin usually did more 1st rotation and it was a toss-up on 2nd rotation.

Just to clarify, you did NOT use fracture, but you DID use CoS/SW, and you didn't control for time, just for the number of rotations. And you did NOT attempt to max DPS, while comparing DPS. I can't tell if the lack of control hurt you or cancelled itself out, and I don't understand why you just didn't use a parser.

It might also be worth noting that with their respective threat sustainables (Shield Oath/Defiance), The gap widens further in Warriors' favor.

Sword Oath adds an auto-attack for free. Compared to a Defiance-free Warrior, the skill is better obviously, yet Paladins still fall short. With Shield Oath, Paladins not only get the -20%, but they also lose that auto-attack, whereas with Defiance you merely lose damage but gain 10% Crit (which can be significant with Crit speccing and Internal Release) in addition to Unchained and Inner Beast. With the new 2.2 gear, the Crit cap for Warrior is something like ~620, which equates to about a 25% chance to crit. Couple that with Wrath's 10% and Internal Release's 10%, it comes to 45% which is certainly nothing to sneeze at. Let's also assume Vengeance will be used at some point, which is another (albeit small) DPS boost in Warrior's favor.

I can't give specific numbers, of course (parsing is against the rules) but, with these considered, you can safely say Warrior trumps Paladin in DPS, and absolutely murders them against multiple targets.

I would expect WAR to deal more damage than PLD - but I would also expect PLD to take less damage than WAR due to the equipped shield. That's the trade-off.

Quote:

---

Originally Posted by Set

Also warrior gets fight or flight.

---

Except they don't... Unless you mean that you added the percentage damage increase to moves after you parsed the log for both classes, to simulate if WAR did have FoF. Which is kind of pointless either way. Was this just a typo?

Quote:

---

Originally Posted by Hitokirinomad

Just to clarify, you did NOT use fracture, but you DID use CoS/SW, and you didn't control for time, just for the number of rotations. And you did NOT attempt to max DPS, while comparing DPS. I can't tell if the lack of control hurt you or cancelled itself out, and I don't understand why you just didn't use a parser.

---

I didn't use fracture because that is in both rotations for war and pld and as I explained that nullifies it's damage. I did use CoS and SW because they are unique to pld. Time control is set via globals and as long as I did relatively well in my rotations it shows the true perspective of damage.
I didn't max DPS because as I said it is a comparison not an attempt to max DPS.

I didn't know there was a parser, I didn't look for one, I didn't want to write a hook into the game and take time to do that, and I wanted to run the controls for myself.

Did you even read the post above I explained almost all of this.

Yes I know read the post I have that data in there. This had nothing to do about who does most damage tanking, it was out of the 2 who has the higher output out of tanking stance.

That was a mistype. I fixed it to "Also warrior gets Internal Release."

As a pure DPS check this works fine. In a real fight where you'll need to save brutal swing for stuns they should be about equal. Also in a fight where the tank needs to focus solely on taking the least amount of DMG possible, pld wins. They have more defensive cool downs and an extra mitigation proc with shield. War excels in pure DMG and pld excels in pure DMG reduction as we've all known for a while.

Also, fracture lasts longer for war than it does for pld which would allow war to do more DMG rotations before reapplying. So they don't actually cancel each other out.

Only gotten MRD to 20 or 21, can't remember right now, and I myself have known for awhile that after it got buffed, WAR had superior damage capability. PLD is about mitigation and WAR is about sustaining. It's at least good to see an "official" confirmation to know it was true.

this was simply to test potential not reality, that is too hard to test because it would be situational.

you are very correct about fracture and I didn't account for that. Oh well script is out there and can be modified to work with it.

Thanks for the work op, war FTW!

Quote:

---

Originally Posted by Set

I didn't use fracture because that is in both rotations for war and pld and as I explained that nullifies it's damage. I did use CoS and SW because they are unique to pld. Time control is set via globals and as long as I did relatively well in my rotations it shows the true perspective of damage.
I didn't max DPS because as I said it is a comparison not an attempt to max DPS.

I didn't know there was a parser, I didn't look for one, I didn't want to write a hook into the game and take time to do that, and I wanted to run the controls for myself.

Did you even read the post above I explained almost all of this.

---

Umm you can't simply discount fracture for warriors. Its 30s compared to paladin's untraited 18s. The sole fact that you wouldn't need to refresh it as often would make it significantly superior for warriors.

Quote:

---

Originally Posted by Set

I didn't use fracture because that is in both rotations for war and pld and as I explained that nullifies it's damage. I did use CoS and SW because they are unique to pld. Time control is set via globals and as long as I did relatively well in my rotations it shows the true perspective of damage.
I didn't max DPS because as I said it is a comparison not an attempt to max DPS.

I didn't know there was a parser, I didn't look for one, I didn't want to write a hook into the game and take time to do that, and I wanted to run the controls for myself.

Did you even read the post above I explained almost all of this.

---

Of course I read the post, that's why I opened with FOR CLARITY, and listed all of the ways you did NOT account for randomness in your experiment. People have already mentioned the Improved Fracture, so that's fine. If you didn't want to use a parser that's on you. Time control is set via globals for special attacks, what about auto-attacks? Why not just set a stopwatch, and do the rotation until it beeps instead of counting rotations.

Also you said you wanted to compare DPS, I don't understand how you compare POTENTIAL(theoretical upper limit) without attempting to maximize. If you were to compare tanking rotation, then you have the WAR rotation wrong.

And I'm not arguing your results btw, I have no dog in this fight. I have both tanks at lvl 50 with relic, and ST damage is not something that usually concerns me. Just pointing out some flaws in the methodology. But I seriously can't wrap my head around trying to compare damage potential w/o attempting to maximize that damage, in tank stance or out. Heck, as OT war I almost never use Butcher's Block that often because there is no need to fight for aggro and I maintain maximum uptime on the Eye/Path debuffs.

Quote:

---

Originally Posted by Set

PLD OVERVIEW
Average is = 1196

WAR OVERVIEW
Average is = 1317

---

Hmm.. lets see.. so you are telling me that their DPS was within roughly 10%? Interesting... Storm's Eye anyone? Keep in mind that if the dummy had SE for PLD, their DPS is virtually identical (actually slightly in favor of PLD) as SE is an 11% damage increase for both PLD and WAR. Yes on it's own WAR will DPS higher then PLD because of this, but all this really means is a PLD/PLD tank combo will do less damage then PLD/WAR combo, which would be roughly equal to WAR/WAR combo.

As far as people talking Fracture, please just stop. Fracture is a TP starve for *minimal* benefit. It also increases risks of debuffs falling off (SE/SP) which are better kept up in nearly every scenario. It isn't bad to tack on for moments of transitions (when you aren't dumping TP into add pickup), but all in all you are splitting hairs if you think it is actually useful.

Hmm, really trying to find how and where this situation with tank doing DPS becomes relevant. I suppose it's like Twintania (pre-2.2) where you have 1 tank that is not taking actual hits, and need to burst DPS something (4 stack conflag in the Twin example). Also, has to be a situation where a tank needs to switch from taking hits to dishing dmg: If it was pure DPS the whole time, then you would just bring a DPS class.

Would be great if SE designed more fights that actually need this. But at this point I almost never see it happening.

This was strictly an exercise in could not implementation because of the people that kept fighting over the damage different I wanted to supply empirical data. The usefulness and likelihood this stands on it's own is dumb to even consider but just shows the numbers.

I only did this because of a conversation unrelated on another thread where someone got all pissy because I did some small test that showed paladin doing more damage on low health mobs and their theoretical numbers where dumb. I figured I did the numbers might as well share the results.

Hey, how bout we not make PSA about using bannable third party software -.-