Tank damage WAR vs PLD testing data and parser

[Set]

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.

[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.

[Set]

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.

[OPneedNerfs]

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.

[Hitokirinomad]

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.

[Apl_Juice]

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.

[Set]

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.

[Astralos]

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.

[Dotsusama]

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?