Warrior Fantasy XIV: 1.0 Strikes Back [WAR/MRD 2.1 Update Thread]

[LinktheDeme]

The idea that the ninja would be giving Goad to the WAR instead of another DPS is foreign to me.

This is a perfect example:

there's some circumstances (T10 an example as i burn out pretty hard since we don't do tank swap anymore). only need one goad to bridge the gap to when our brd normally plays paeon, in which time the melees are just hitting empty.

This scenario applies to any real fight. All DPS should have invigorate up to carry them through the first instance otherwise you're just sitting on a goad CD unnecessarily. Warriors do not get invigorate so they need the goad to carry their TP amount to when the BRD would play paeon for everyone. If paeon is completely absent from the fight, then the situation changes. It becomes a question is if it's not being played because the DPS is stopping due to mechanics or if the BRD and melee don't need paeon.



Keep in mind in 99% of situations you can early goad a warrior and have it off CD for when a DPS would need it. If you early goad a melee they sit on invigorate, if you wait until after the invigorate you sit on a goad. Unless the physical attacker has AOE spam early on they'll always have enough TP until the 3 minute timer is up. Even then they may not even need it depending on the paeon timing.

Any decent warrior, especially with esuna/leeches, will run out of TP rather quickly. Even quicker with selene out. (For example: with no selene , no goad, and esunas on T13 I am generally out of TP by the end of gigaflare depending on earthshaker RNG)




Random Math:

In the span of about 30 seconds of continuous DPS a warrior will expend about 800 TP
In the span of that 30 seconds 500 TP will be regenerated due to TP regeneration in combat. The goad will make up that difference of 300 and give an additional 100 which would make that last 30 seconds of DPS moot from a TP perspective and a gain.

Which is why you can goad a warrior at 800-900 or so TP (This is what I mean by early goad. Mileage will vary by WAR, selene vs eos, berserk and esuna/leeches timing.) without fear of overflowing TP or clipping goad / invigorate CDs.


Below is an illustration of how the TP value of the war would vary, math includes two of the basic SS builds right now for WAR.(without Selene) If you have Selene you can see how you can get away with goading a war at even 900 TP.

Spoiler: show

A more realistic illustration

Spoiler: show

Note: There are certain assumptions at when the TP regen ticks at that had to be made. This will vary results slightly. Yellow indicates a natural 1000 whereas gray indicates and overage.

[Obsidian]

Soooooooooo, after I posted WAR's enmity mods quite some time ago with a comment about doing PLD's "some time soon", I figured 7 months was long enough to get around to it.

For reference this was the post for WAR's mods:

http://www.bluegartr.com/threads/119...=1#post6188661

Everything was done the same way, same methods, same test partner, etc.

The results:

Savage Blade     DMG x 3
Rage of Halone   DMG x 5
Shield Lob       DMG x 3
Circle of Scorn  [Initial Damage x 3] + DoT damage

Shield Oath      DMG x [WS mod] x 2

And for comparison:

Skull Sunder     DMG x 3.5
Butcher's Block  DMG x 5.5
Tomahawk         DMG x 3
Overpower        DMG x 5
Steel Cylone*    DMG x 6

Defiance         DMG x [WS mod] x 2

Of note:

The mod on CoS:
A bit surprising but made sense in retrospect. The ability description has the "Additional Effect: Increased enmity" following the text regarding the potency of the initial hit, so it stands to reason it served as an addendum that applied to only that damage and that the DoT effect did not benefit from it (or it would have been at the very end of the description below the DoT damage). We figured it out by using a bunch of crit gear to get a larger variance on the DoT portion of the damage, came up with the above model then re-tested a few times to verify its consistency.

Difference in PLD/WAR mods:
PLD's mods being lower than WAR's on their hate combo makes a certain amount of sense as well. The higher mods on Sunder and Block are most likely there to make up for the fact that WAR is juggling 3 combos, only 1 of which has an enmity bonus, compared to PLD who uses nothing but Halone.

More notes on Flash:
The patch that added the hate mod to CoS also increased the hate on Flash, once again we just kinda screwed around with it a bit without trying to get it's actual potency. Got a few loose declarations regarding Flash's effect and what increases it.

-Increased by DET
-Increased by STR
-Does not seem to be effected by any other primary stats
-Not effected by general ilvl
-Does not seem to have a variance (static amount generated every time)
-Have yet to see it crit, but not enough data to say whether or not it can with any certainty

It's worth noting that after the boost to Flash it should consistently be more hate than an un-combo'd Halone or Block, which was not true before the buff.

These are a few numbers from the very limited testing we did, spoilered as they're not complete tests just some examples to give people an idea:

Spoiler: show

Only had Aug.Ironworks sword on hand since there wasn't any real plan for rigorous testing, so the "baseline" numbers have +26 DET from the weapon.

[Stats] ------------------->  [Damage before and after hate was transferred]

D 57  STR 243  DET 228  --->  537-548
                              535-545

D 57  STR 243  DET 282  --->  551-561
                              546-557

D 57  STR 325  DET 228  --->  675-686

+54 DET gave an increase of ~12-13 enmity
+82 STR gave an increase of ~139-140 enmity

Rigorous Flash testing can be expected "sometime soon" (see: fucked if I know, wouldn't hold my breath).

Did you check your results against the Enmity plugin for ACT?

[Niiro]

I've never used or heard of it.

[Obsidian]

There's a link to it here, if you care.

[Dammerung]

The idea that the ninja would be giving Goad to the WAR instead of another DPS is foreign to me.

Yes, since it should be going to the PLD first.

[Yugl]

Niiro's values seem to work using the ACT plugin. Good work there.

As for Flash:

Assumptions

1. The potency of flash, like other actions, will be some multiple of 10 (Meaning, no 603 or 605 values).
2. The equation for damage and flash looks similar to this:

(((Base Damage)*Potency)*Randomization)

3. Randomization is +/-5%
4. Flash has no randomization
5. Rounding occurs after Randomization (Damage) unless no randomization occurs (i.e. Flash). If no randomization occurs, then rounding occurs after Potency.
6. There is a "Real" and "Displayed" Flash threat. The displayed flash threat is the rounded integer displayed to the user. The real flash threat is what the game calculates before rounding that value.
7. The base damage for Flash is the same as other damage skills (Halone/Savage Blade/etc)

Idea

If Flash threat acts like any other damage spell other than no randomization or critical hits, then dividing the Flash threat by its potency should give us a pretty accurate base damage. If we know the base damage, we can accurately predict the minimum and maximum damage from a skill of known potency.

First, we want a potency range to select from. In this case, I used Halone which has an uncomboed 100 potency. If (3) is true, that makes the maximum potency 105 and the minimum potency 95. Since the base damage for Halone and Flash should be the same, we can set them equal to each other in the following equations:

(Flash Threat/Flash Potency) = (Halone DMG/Halone Max Potency)
(Flash Threat/Flash Potency) = (Halone DMG/Halone Min Potency)

Second, we will want to test this range of potencies by tens (See assumption [1]) using the following equation:

(Flash threat/Flash Potency)*(105) = Maximum Halone Damage
(Flash threat/Flash Potency)*(95) = Minimum Halone Damage

Results

Test: 1
Class: GLD
Gear: http://bucket.bluegartr.com/4a2e1b8d...b7159ed127.png
Halone results: https://dl.dropboxusercontent.com/u/...mityValues.act
Halone Max/Min: 36/33
Flash DMG (Threat from ACT Plugin): 212

Test: 2
Job: PLD
Gear: http://bucket.bluegartr.com/6c175779...a75f533ea6.png
Halone results: https://dl.dropboxusercontent.com/u/...Halone1000.act
Halone Max/Min: 133/120
Flash DMG (Threat from ACT Plugin): 771

Math

Test 1:
(212/Flash Potency) = (36/105)
Flash Potency: 618.333333

(212/Flash Potency) = (33/95)
Flash Potency: 610.303030303

Test 2:
(771/Flash Potency) = (133/105)
Flash Potency: 608.6842105

(771/Flash Potency) = (120/95)
Flash Potency: 610.375

Based on these results, a range of 600-620 potency. This leaves us 600, 610, and 620 to try out.

Potency at 600

Test 1:
(212/600)*(105) = 37.1
(212/600)*(95) = 33.5666667
Predicted Max/Min: 34-37 [False]

Test 2:
(771/600)*(105) = 134.925
(771/600)*(95) = 122.075
Predicted Max/Min: 122-135 [False]

Potency at 610

Test 1:
(212/610)*(105) = 36.49180328
(212/610)*(95) = 33.01639344
Predicted Max/Min: 33-36

Test 2:
(771/610)*(105) = 132.7131148
(771/610)*(95) = 120.0737705
Predicted Max/Min: 120-133

Potency at 620

Test 1:
(212/620)*(105) = 35.90322581
(212/620)*(95) = 32.48387097
Predicted Max/Min: 32-36 [False]

Test 2:
(771/620)*(105) = 130.5725806
(771/620)*(95) = 118.1370968
Predited Max/Min: 118-131 [False]

Cleaning this up with more Math

There is, of course, a fair counter to all of this. The equations only used the Displayed Flash threat (See assumption 6) instead of the real Flash threat. For example, using 212 Flash threat may not let 600 potency accurately predict Halone damage, but using 212.49 (Which would get rounded down to 212 displayed Flash threat) may do so. Lets see what happens if we take both extremes of this (211.5/212.49 and 770.5/771.49).

Potency at 600

Test 1:
(211.5/600)*(105) = 37.0125
(211.5/600)*(95) = 33.4875
Predicted Max/Min: 33-37 [False]

(212.49/600)*(105) = 37.18575
(212.49/600)*(95) = 33.64425
Predicted Max/Min: 34-37 [False]

Test 2:
(770.5/600)*(105) = 134.8375
(770.5/600)*(95) = 121.99583333
Predicted Max/Min: 122-135 [False]

(771.49/600)*(105) = 135.01075
(771.49/600)*(95) = 122.1525833
Predicted Max/Min: 122-135 [False]

Potency at 620

Test 1:
(211.5/620)*(105) = 35.8185
(211.5/620)*(95) = 32.40725806
Predicted Max/Min: 32-36 [False]

(212.49/620)*(105) = 35.986
(212.49/620)*(95) = 32.55895161
Predicted Max/Min: 33-36

Test 2:
(770.5/620)*(105) = 130.4879032
(770.5/620)*(95) = 118.0604839
Predicted Max/Min: 118-130 [False]

(771.49/620)*(105) = 130.6555645
(771.49/620)*(95) = 118.2121774
Predicted Max/Min: 118-130 [False]

In another attempt to salvage these potencies, one might argue that the value is between these extremes. However, as we can see from these values, there is a problem with that argument.

Problem with 600 Potency:
The minimum value to come out with 212 displayed Flash threat would be 211.5 real Flash threat. If you look at Test 1, this minimum value overestimated the maximum Halone damage.

Problem with 620 Potency:
The maximum value to come out with 771 displayed Flash threat would be 771.499999999999 real Flash threat. If you look at Test 2, this maximum value underestimated the minimum Halone damage by a significant margin. Even if you used 771.5 real Flash threat, which would not show 771 displayed Flash threat anyways, the damage is underestimated.

Conclusion

I'm going to go ahead and place Flash at 610 potency based on this evidence.

[jibini]

I can't tell if I'm overlooking something obvious or applying numbers incorrectly or something else here, so I'm probably about to make an idiot of myself.

Does Flash suffer from the penalty of Shield Oath (and presumably Defiance) even though it's not really damage?

If not, it seems like Flash would have a higher potency than the Halone combo in Shield Oath:

Fast Blade = 150 potency * 1 WS mod * 2 Shield Oath enmity mod = 300 potency worth of enmity
Savage Blade = 200 * 3 * 2 = 1200
Rage of Halone = 260 * 5 * 2 = 2600

Average(300 + 1200 + 2600) = 1367
1367 * 0.8 = 1093 average potency worth of enmity

Compared to 1220 average for Flash (610 base, doubled under Shield Oath).

[Yugl]

It does. When I tested Shield Oath + Flash using the ACT plugin, the increase was 1.6x instead of the expected 2x value.