Warriors - How to fix our damage mitigation

[Hachiko]

I just did the math for *in* tank stance. It's a slight edge to PAL, and I've got *no clue* where you got the 8-16% number from.

I've done the math too. In tank stance a PLD will be at ~295 Potency, while a WAR will be ~312 Potency, without counting Berserk, Unchained, Internal Release, or Inner Beast. Without any of that, WAR has ~6% lead. If you add those things in, and use them on cooldown you can extend the lead to 18% or higher, but it comes at a high cost to the WAR and isn't realistic. I just made a (long) post on Reddit about it, I'd be happy to link it. (Note that I also considered PLD using fight or flight on Cooldown).

It's pretty much the same numbers as the tank stance stuff with a couple changes: WAR no longer has a .75 modifier and PAL loses the .8 modifier. On top of that, PAL gets some bonus potency in the form of Sword Oath. Baseline potency per GCD of an autoattack is 83.33 (auto-attack / ((damage * delay) / GCD)). Sword Oath provides an additional 50 potency to each of those, such that PAL auto-attack is 133.33 additional potency. Bring in the previous formulas modified for the given values (using the Fracture math for WARs):

WAR:
(213.85 + 83.33)*1.2 * 1.0556 / .9 = 418.27

PAL:
(203.3 + 25 + 20.8 + 133.33) * 1.05 = 401.55

PAL (with WAR present):
(203.3 + 25 + 20.8 + 133.33) * 1.05 / .9 = 446.17

So a PAL does either 96% of what a WAR does or 106.6% while in their DPS stances. Once again, we're talking about what is effectively no difference (balanced within 5% is balanced).

This is slightly off. First, both circle of Scorn and Spirits Within work out to 25 Potency per 2.5, so I don't know where 20.8 comes from.

Second, you're not normalizing sword oath and instead adding 50 potency to the auto attacks which you've normalized to 2.5 seconds. Using Curtana as an example, that means an additional 50 potency attack every 2.32 seconds, not every 2.5.

So Sword Oath will contribute ~53.9 potency per 2.5 second global.

This puts the PLD (solo) calculation at:

203.33 + 53.9 + 25 + 25 + 83.333 = 390.563

With Fight or Flight used on Cooldown, it works out to 429.619, which is already higher than the WAR before you add in Storm's Eye for the PLD. But with Internal Release and Berserk used on CD for the WAR they would slightly overtake the PLD, but still be lower than a PLD in a situation where a warrior is present and using storm's eye.

I'm not sure why you were multiplying by 1.05 (Fight or Flight is a 10% buff, averaged out).

[Kitru]

I just made a (long) post on Reddit about it, I'd be happy to link it.

If you wouldn't mind, I'd like to see how/where our math deviates. My math is in this very thread.

[Hachiko]

If you wouldn't mind, I'd like to see how/where our math deviates. My math is in this very thread.

I went and looked at your math. I think the main difference is that you assumed Storm's Eye would be present for a PLD tank.

But I'll also note that I assumed full health on Spirits Within while you seemed to assume ~70%ish.

Also your calculation for Berserk and Fight or Flight seem to be off. Berserk is more than 20% increase in damage (from my testing), and Fight or Flight averages out to 10% boost, not 5% (33.33% uptime, 30% bonus = 9.99 repeating). It's the best damage dealing cooldown in the game by a huge margin.

(also you should see the previous page for why PLD do more damage than you calculated, while not tanking).

No, it is not a different construct. The problem is you are looking at static EHP which is insufficient and putting too much emphasis on "the burst." They're is no difference between Pal and War here. Typical fights throughout the past decade of MMOs have boss damage outputs that can be categorized roughly in terms of linear, small spike, and large spike. The large spike is not the real problem- by numbers posted on the forums, War can handle this. It is much more likely, as has almost always been the case, it is small and large spikes occurring in close proximity. The tank HP is never enough to sustain both. You need healing to get through these burst phases. It is more appropriate to approach HP in dynamic terms rather than simply claim it is a different issue.

EHP is a very different construct than Healing required to stay alive. For the purposes of overall durability it is more appropriate to approach it in dynamic terms, but in doing so it's also appropriate to consider other factors (i.e. block, parry, debuffs, etc.) as well.

[Coramac]

EHP is a very different construct than Healing required to stay alive. For the purposes of overall durability it is more appropriate to approach it in dynamic terms, but in doing so it's also appropriate to consider other factors (i.e. block, parry, debuffs, etc.) as well.

It's not an overall durability issue. It's a survivability issue. Evasion statistics are not meaningful here. They reduce the probability of severity, but unless you can stack them very high, their effect on the probability is too low to be relied upon and should be dismissed. It would be meaningful to take into consideration Bulwark's effect. As a general rule of thumb, unless your total evasion is somewhere in the vicinity of 45%, it doesn't have enough impact to be considered impactful for a full fight. For cooldown usage and for singular occurrences, 70% is a good mark. Typical numbers for typical fights, not universals.

[Kitru]

Berserk is more than 20% increase in damage (from my testing)

Berserk provides 50% increased damage for 20 seconds followed by 5 seconds of only auto-attack damage every 90 seconds. Split this up into auto-attack contribution and ability contribution.

Auto attack contribution gets 11.11% (50% for 20 seconds every 90 seconds). The ability contribution is substantially lower because you have 5 seconds of Pacification. As such, you get 20 seconds of 1.5x performance, 5 seconds of 0 performance, and 65 seconds of 1.0x performance, which totals out to 5.56% better than average. I was only doing the ability damage contribution (which means I need to go back and fix that math to account for auto-attacks).

[Kiteless]

Berserk provides 50% increased damage for 20 seconds followed by 5 seconds of only auto-attack damage every 90 seconds. Split this up into auto-attack contribution and ability contribution.

Auto attack contribution gets 11.11% (50% for 20 seconds every 90 seconds). The ability contribution is substantially lower because you have 5 seconds of Pacification. As such, you get 20 seconds of 1.5x performance, 5 seconds of 0 performance, and 65 seconds of 1.0x performance, which totals out to 5.56% better than average. I was only doing the ability damage contribution (which means I need to go back and fix that math to account for auto-attacks).

This is not true. Berserk increases your attack power by 50%, not your damage. The end result is that the damage increases is actually lower than 50%.

[Kitru]

This is not true. Berserk increases your attack power by 50%, not your damage. The end result is that the damage increases is actually lower than 50%.

So what's the conversion? I find it kind of annoying that everything that increases WAR attributes (Berserk, Foresight) increases a baseline attribute rather than an end value whereas everything a PAL gets apparently increases end values. It's not only annoying as a WAR but it's annoying to theorycraft since there isn't a standard mechanism.

[Kitru]

First, both circle of Scorn and Spirits Within work out to 25 Potency per 2.5, so I don't know where 20.8 comes from.

Did you not read any of the explanations I did for the math beforehand? I used 250 for Spirits Within instead of 300 because, if you're using it on CD, you can't guarantee you'll always be at max hp. If you wait until you're at max hp, you're not using it on CD and, as such, are getting less out of it.

Second, you're not normalizing sword oath and instead adding 50 potency to the auto attacks which you've normalized to 2.5 seconds. Using Curtana as an example, that means an additional 50 potency attack every 2.32 seconds, not every 2.5.

You don't have to normalize for attack speed. The attack speed math already does that. Sword Oath doesn't provide any more for a fast weapon than it does for a slow weapon. Potency is a percentage increase to damage dealt. 50 additional potency means that you're dealing 50% more (additional). If you swing twice as fast for half of the damage, getting 50% more damage adds up the same. The only way that you would get *more* damage out of a fast weapon rather than a slow one is if the damage added is flat, which Sword Oath is not. As such, regardless of the weapon you're using, Sword Oath increases the contribution from your auto attacks to 133 from 83.

I'm not sure why you were multiplying by 1.05 (Fight or Flight is a 10% buff, averaged out).

FoF provides 5% over time, not 10%. It's 30% while it's active with a 16.67% uptime (30 seconds every 180 seconds). Multiply that out and you get 5%.

Edit: Just realized that the wiki I was using for info was using out of date stuff. The CD on FoF is 90 seconds, not 180, which is why it's 10%, not 5% like I was using.

I've done the DPS math for WAR and PAL twice. First was done for tank stances, and second for DPS-mode. I suggest you read up on them because I explained everything I did explicitly.

[Hachiko]

I've done the DPS math for WAR and PAL twice. First was done for tank stances, and second for DPS-mode. I suggest you read up on them because I explained everything I did explicitly.

I did read. I still prefer assuming 100% health rather than significantly less health for spirits within. But that's beside the point.

You are wrong on Fight or Flight. It has a 33% uptime, not 16.667. http://xivdb.com/?skill/20/Fight-or-Flight

And second, you do have to normalize for sword oath. Sword oath adds a 50 potency hit on auto attack. Auto attacks happen more often than once every 2.5 seconds for a PLD. With Curtana (one of their slowest weapons) it's once every 2.32 seconds, which puts sword oath at ~54 potency every 2.5 seconds. You've normalized the auto attack contribution to 2.5 seconds, but you didn't normalize sword oath.

Auto Attack potency is (Delay / 3)*100. Sword oath will add 50 to that number, but that number will not always be 83.333 for a PLD.

Sword oath absolutely provides more for a fast weapon than a slow weapon.

[Wyndam]

And second, you do have to normalize for sword oath. Sword oath adds a 50 potency hit on auto attack. Auto attacks happen more often than once every 2.5 seconds for a PLD. With Curtana (one of their slowest weapons) it's once every 2.32 seconds, which puts sword oath at ~54 potency every 2.5 seconds. You've normalized the auto attack contribution to 2.5 seconds, but you didn't normalize sword oath.

Auto Attack potency is (Delay / 3)*100. Sword oath will add 50 to that number, but that number will not always be 83.333 for a PLD.

Sword oath absolutely provides more for a fast weapon than a slow weapon.

Oh, duh. The wording on Sword Oath is weird. It does not increase the potency of the attack. It adds a second hit at 50 potency that occurs when the auto attacks occur. It's like Enfire or abilities like that. You'll actually see it pop up on screen as two separate numbers. I'd been wondering why it was doing that.