So, you want to Strength Tank? The Full Comparison and Info!

[Donjo]

I did this stuff before I ever got melded i90 accessories. Just hit a dummy as a Pally until TP hit zero. The small difference between adding 90 STR and 60 STR makes me think I did something wrong, though.

Edit: Yep, these are bad tests because they screw with secondaries too much. They're still sufficient to show that STR counts, though.

Test: Halone until TP is gone and cooldowns on cooldown.

Situations

Shield Oath: Act as if I've just pulled: FoF -> Shield Lob -> Rotation

Sword Oath: Act as if I'm OT: FoF -> Rotation

Why PLD?

1. PLD DPS is super easy to optimize and impossible to screw up.
2. My WAR isn't at 50. (Well, it wasn't when I did this. It is now, at the time of this post!)

Weapon: Diamond Brand
Shield: Diamond Shield

Full VIT, Full Parry

Test 1: Shield Oath
428 STR
395 Crit
221 Det
410 SS
590 Parry

Duration: 2:38
Total DPS: 169.99

Test 2: Sword Oath
428 STR
395 Crit
221 Det
410 SS
590 Parry
21
Duration: 2:49
Total DPS: 259.72

Weapon: Diamond Brand
Shield: High Allagan Shield

All DPS Accessories. Four 110's and one 120 increase STR by 93. Switched HA Boots for Noct Boots for Determination boost. This emulates the decision to wear Pentamelded i90 + 30 bonus points in STR.

Test 3: Shield Oath
521 STR
418 Crit
280 Det
401 SS
499 Parry

Duration: 2:36
Total DPS: 215.67

Increase: 26.47%

Test 4: Sword Oath
521 STR
418 Crit
280 Det
401 SS
499 Parry

Duration: 2:48
Total DPS: 320.30
309.55

Increase: 23.32%

Conclusion: 93 STR and a small bit of secondary swapping increased my DPS by approximately 24-25% overall.

A mix of DPS Accessories and bonus stats to increase STR by 65. This emulates the decision to use pentamelded i90 while having bonus points in VIT.

EDIT: Well, it was supposed to be 65 but it's actually 68. Coincidentally proportional to the first set having 93 instead of 90.

Test 5: Shield Oath
496 STR
399 Crit
268 Det
401 SS
532 Parry

Duration: 2:40
Total DPS: 206.47

Increase: 21.46%

Test 6: Sword Oath
496 STR
399 Crit
268 Det
401 SS
532 Parry

Duration: 2:50
Total DPS: 305.68

Increase: 17.69%

Conclusion: 68 STR increased my DPS by approximately 19% overall.

Bonus finding: It would appear that the extra STR is increasing Shield Oath DPS more than Sword Oath DPS.

[Abbul_Stonecleaver]

Very good info. The difference that dps can make in an actual fight can be huge. Example: right now my group is in phase 3 of T12. Last time our best attempt we had the fifth Bennu spawn at I believe 54% without using LB. If we can LB2 and tanks figure out how to push that little bit more (maybe even not need LB if whole group can push hard enough), we may can start consistently hitting 4 Bennus.

[Donjo]

Moar STR Testing. This is again on Pally because the rotation is easy. I have a WAR at 50 but lack knowledge of the most efficient rotations, so if anyone has an idea of something I could do, feel free to call it out!

Starting Stats
446 STR
382 CRIT
228 DET
341 SKS

Equipment: Full Left Side Ironworks + a Parry/Accuracy Excalibur and Aegis Shield

Rotations: Halone for 3 minutes. Use FoF, CoS, and SW on cooldown. No Mercy Stroke or Fracture.

The accessories I will use are Topaz melded with just Accuracy and Strength. Parry and Accuracy are the only Secondary Stats that will change as I mess with stuff. Otherwise, I'll use Bonus Attributes to play with STR as well.

Test 1: 446 STR(base)

Shield Oath: 180.53
Sword Oath: 264.85

Test 2: 476 STR(+30)

Shield Oath: 195.35
Sword Oath: 278.38

Test 3: 511 STR(+65)

Shield Oath: 206.22
Sword Oath: 306.97

Test 4: 541 STR(+95)

Shield Oath: 214.80
Sword Oath: 310.29

Well, that's curious. Why such a small increase between 65 and 95?

[Abbul_Stonecleaver]

Hmm that is odd. Perhaps RNG with crits during FoF?

[ZDamned]

You failed to take into consideration how man Critical hits you actually had on each parse.

I have a Dragoon in my group who has gotten lucky, and dealt more damage with a very sub optimal rotation just due to crits vs the most logical, and highest potency rotation for the same time period.

Re-parse, and add in how many times you crit in each one, and you will see why the numbers are so different.

The fact that more strength consistently added damage still shows that even with the overall chance of bad damage rolls and bad Crit rolls, you STILL can pump out a significant difference in DPS with minor upgrades to the stat.

[Donjo]

The fact that more strength consistently added damage still shows that even with the overall chance of bad damage rolls and bad Crit rolls, you STILL can pump out a significant difference in DPS with minor upgrades to the stat.

Which is exactly what the test was meant to find out... so there's no need to really test further. Besides, I'm too lazy to repeat all of these enough times to normalize the crit rate

It does surprise me though that Crits would make such a difference when my Crit Rate for the tests equated to a theoretical rate of only 7-8%.

[ZDamned]

Which is exactly what the test was meant to find out... so there's no need to really test further. Besides, I'm too lazy to repeat all of these enough times to normalize the crit rate

With no offense to the work you did, that sort of invalidates the results for actual DPS increases... =/

If you saved your parse logs, it wouldn't be too hard to go back and check how many times you crit with each encounter, just add the number after the DPS.

Basically, if you got lucky, each crit adds an entire extra attack, so even as many as 5 lucky crits can actually make Shield oath look like it's almost on par with Sword Oath on DPS.

The reason your parses are not valid at that point, is because they only show a small sample of data. In order for an actual useful data set, you would need to macro an entire TP pool, and let it run fully parsed for at least 20 encounters on the dummy, then average ALL of those results together. That removes the spike in DPS from chance over time. (Mean Median Mode math from the past!)

Anway,I'm not trying to attack you, but you did sort of invalidate your findings by excluding/refusing to take this into account. =P

It does surprise me though that Crits would make such a difference when my Crit Rate for the tests equated to a theoretical rate of only 7-8%.

If your Crit rate was something more consistent, then you wouldn't need to care as much, but the fact that you think it's 1-2 in 10 hits makes it a very misleading data set. Crit-rating is on LOTS of fending gear, as well as DPS gear, so you haven't taken ANY of those factors into account for their value.

I know I keep saying the word Consistent, but thats the key word here as to why we use terms like DPS, instead DPH or Potency per second.
When you are trying to figure out your damage over time, it's ALL about how you remove RNG form being a factor.

[Lavieh]

it's ALL about how you remove RNG form being a factor.

Given how RNG works and functions you could just do longer parses to get more equalized results.

[ZDamned]

Given how RNG works and functions you could just do longer parses to get more equalized results.

Sure! Longer Parses, because we don't have any limiting factors to stop us from doing that.

Oh wait.. TP. Awwwwwww

=P hehe They need to seriously get rid of TP. it's pointless...

[Lavieh]

Sure! Longer Parses, because we don't have any limiting factors to stop us from doing that.

Oh wait.. TP. Awwwwwww

=P hehe They need to seriously get rid of TP. it's pointless...

Find a BRD friend and ask him to Paeon - AFK for 10 minutes.