PSA: Altmer "Spell Recharge" passive is providing less mitigation than intended

Reorx_Holybeard

TheYKcid wrote: »

@Reorx_Holybeard

All vulnerabilities are additive with each other—and this summated modifier is then applied multiplicatively to incoming damage. This is intended and always worked that way, so that's another story.

Actually, I meant the vulnerability is additive to damage taken mitigation...I did extensive testing last month on PTS for our build editor but haven't gotten around to posting about it due to the complexity and strange results. Basically the way it looks like it is calculating vulnerabilities is like:

Vulnerability = Vulnerability1 + Vulnerability2 + ...
DamageTaken = DT1 * DT2 * DT3 *...
Mitigation = Resist * (DamageTaken + Vulnerability) * Block * Player * ....

In fact, it gets even stranger in that the exact calculation depends on the order you apply the effects and then gets "reset" to the above calculation when you logout and back in. Depending on the order you can get everything going additive or everything going multiplicative. Took a long time to figure that all out. Should test on different mobs to make sure it is repeatable and still occurs. If anyone has recent data that either confirms or refutes this I'd love to set it.

If you have no Vulnerabilities then everything (mostly, some strange rounding errors probably) works as you'd expect multiplicative stacking to work.

TheYKcid

-deleted-

starkerealm

TheYKcid wrote: »

So this is really weird.

After relogging due to a disconnect, I went to test against the same NPC under the same conditions, and now I was getting full mitigation against all their attack types. Wrathstone seems to have messed-up damage calculation in some very fundamental way and I don't know where to begin elucidating how...

I'm going to throw a guess out there: Some buffs are persisting after you've removed them from your build (however you had them to begin with), which is then lost on a relog.

TheYKcid

starkerealm wrote: »

TheYKcid wrote: »

So this is really weird.

After relogging due to a disconnect, I went to test against the same NPC under the same conditions, and now I was getting full mitigation against all their attack types. Wrathstone seems to have messed-up damage calculation in some very fundamental way and I don't know where to begin elucidating how...

I'm going to throw a guess out there: Some buffs are persisting after you've removed them from your build (however you had them to begin with), which is then lost on a relog.

Hmm... that's very possible. Raises some scary possibilities, like if it also applied to debuffs. Or going into PvP with PvE set buffs despite changing gear.

Reorx_Holybeard

TheYKcid wrote: »

So this is really weird.

After relogging due to a disconnect, I went to test against the same NPC under the same conditions, and now I was getting full mitigation against all their attack types. Wrathstone seems to have messed-up damage calculation in some very fundamental way and I don't know where to begin elucidating how...

Might be the same thing I observed with mitigation effects depending on the order you add/remove them until you log out/in. My guess is the game has 2 different types of calculation for mitigation:

1. Complete recalculation (happens at login and perhaps other times)
2. A difference recalculation whenever you add or remove a mitigation effect.

In theory these should be identical but with rounding and/or bugs it seems that they are not.

TheYKcid

Reorx_Holybeard wrote: »

TheYKcid wrote: »

So this is really weird.

After relogging due to a disconnect, I went to test against the same NPC under the same conditions, and now I was getting full mitigation against all their attack types. Wrathstone seems to have messed-up damage calculation in some very fundamental way and I don't know where to begin elucidating how...

Might be the same thing I observed with mitigation effects depending on the order you add/remove them until you log out/in. My guess is the game has 2 different types of calculation for mitigation:

1. Complete recalculation (happens at login and perhaps other times)
2. A difference recalculation whenever you add or remove a mitigation effect.

In theory these should be identical but with rounding and/or bugs it seems that they are not.

That could be it too. And it wouldn't be the first time the game had trouble handling mitigation sources depending on the order of application. This happened during Morrowind PTS and I'm actually still unsure if it ever got fixed:

https://forums.elderscrollsonline.com/en/discussion/347576/selective-noncommutivity-of-warrior-cp/p1

TheYKcid

TheYKcid wrote: »

starkerealm wrote: »

TheYKcid wrote: »

So this is really weird.

After relogging due to a disconnect, I went to test against the same NPC under the same conditions, and now I was getting full mitigation against all their attack types. Wrathstone seems to have messed-up damage calculation in some very fundamental way and I don't know where to begin elucidating how...

I'm going to throw a guess out there: Some buffs are persisting after you've removed them from your build (however you had them to begin with), which is then lost on a relog.

Hmm... that's very possible. Raises some scary possibilities, like if it also applied to debuffs. Or going into PvP with PvE set buffs despite changing gear.

To expand on this, persisting buffs alone would not explain the discrepancy. As mentioned earlier, any other resistances or %-mitigation further up the equation would have no bearing on the results on my test, since I was specifically doing a before/after comparison with one source of multiplicative mitigation, and measuring the change in proportion.

The only thing that could reduce a 5% value to 4.1 or 4.3% (or any of the other whacky figures @Reorx_Holybeard found) would be if certain multiplicative modifiers got mis-factored as additive ones, in the same way Reorx observed vulnerabilities to be doing.

Very, very weird stuff...

TheYKcid

Okay I finally figured it out, it was the exact issue of vulnerability mis-factoring that you described. It just occurred to me that I was testing against a fire NPC (Flame Atro) as a stage 3 vamp. So here is more testing, naked, using it's basic attack "Flare".

Stage 1 (no vulnerability) - 777 damage
Stage 1 + Minor Protection - 715 damage (777*0.92, rounded-up)
Stage 3 (20% vulnerability) - 933 damage (777*1.2, rounded-up)

Everything looks good so far, until you include minor protection (8% mitigation) as a stage 3 vamp:
870 damage

You would expect a result of 933*0.92 = 859 rounded-up, because vulnerabilities and %-mitigation are supposed to interact multiplicatively, but you only get 870 is it gets calculated as 777*(0.92+0.2), indicating an additive operation.

Minno

TheYKcid wrote: »

Okay I finally figured it out, it was the exact issue of vulnerability mis-factoring that you described. It just occurred to me that I was testing against a fire NPC (Flame Atro) as a stage 3 vamp. So here is more testing, naked, using it's basic attack "Flare".

Stage 1 (no vulnerability) - 777 damage
Stage 1 + Minor Protection - 715 damage (777*0.92, rounded-up)
Stage 3 (20% vulnerability) - 933 damage (777*1.2, rounded-up)

Everything looks good so far, until you include minor protection (8% mitigation) as a stage 3 vamp:
870 damage

You would expect a result of 933*0.92 = 859 rounded-up, because vulnerabilities and %-mitigation are supposed to interact multiplicatively, but you only get 870 is it gets calculated as 777*(0.92+0.2), indicating an additive operation.

Yea sounds like vulnerability isnt working right.

usmcjdking

starkerealm wrote: »

TheYKcid wrote: »

@starkerealm naked as far as the mechanic is question was concerned, yes.

@TheYKcid, so you're still wearing armor?

lol math is clearly hard for you.

TheYKcid

Tested with both Ferocious Leap (vs. vamp) & Incap vs another player. The vulnerability was treated as additive in both cases, and was doing far more damage than it should. I'm guessing ZOS did a ninja revamp of the damage equation, and this is the norm now in Wrathstone, even with just 1 vulnerability effect and no stacking.

Prior to this update it was most definitely multiplicative (tested Incap extensively last patch).

Still haven't tried to replicate Reorx's order-of-application bug that turns it multiplicative, though, that's a whole other issue.

MashmalloMan

TheYKcid wrote: »

Now that I think about it, is the Dragonguard + Powerstone interaction even an anomaly?

0.85*0.85 = 0.7225, rounded up to .73 which is exactly a 27% cost reduction.

This is assuming the rule for ult cost reduction is typically multiplicative, though IDK if that's the case for ult cost redux specifically.

Similar thing happens with Redguard's -8% cost on weapon ultimates with sorc's powerstone -15%, never checked how that interacts with other sources of ult cost reduction.

It should be 1 of 2 outcomes, but it isn't.

Additive
0.8 + 0.15 = 0.23
1 - 0.23 = 0.77
Ballista 175 ult * 0.77 = 134.75

Multiplicative
0.92 * 0.85 = 0.782
Ballista 175 ult * 0.782 = 136.85

So tell me how the ult cost of Ballista comes out to 139?

starkerealm

Marshall1289 wrote: »

TheYKcid wrote: »

Now that I think about it, is the Dragonguard + Powerstone interaction even an anomaly?

0.85*0.85 = 0.7225, rounded up to .73 which is exactly a 27% cost reduction.

This is assuming the rule for ult cost reduction is typically multiplicative, though IDK if that's the case for ult cost redux specifically.

Similar thing happens with Redguard's -8% cost on weapon ultimates with sorc's powerstone -15%, never checked how that interacts with other sources of ult cost reduction.

It should be 1 of 2 outcomes, but it isn't.

Additive
0.8 + 0.15 = 0.23
1 - 0.23 = 0.77
Ballista 175 ult * 0.77 = 134.75

Multiplicative
0.92 * 0.85 = 0.782
Ballista 175 ult * 0.782 = 136.85

So tell me how the ult cost of Ballista comes out to 139?

When power stone is mixed with Dragonguard, the 1pt 8% discount drops to 6%, and the 15% drops to 12%. (Not sure which one.)

From memory, the Imperial discount stacks to 18% with either one.

Sounds like the same thing here. 15% + 6% = 21%. Then 175 * .79 = 138.25 (which, if it's forced to round up) would get you 139. But, I mean, this is a guess at this point.

Normally, I like to use Storm Attronach as my gold standard, because the base ult cost is 200, so you get nice, clean, .5% readouts. But, it's still incredibly strange.

starkerealm

TheYKcid wrote: »

Tested with both Ferocious Leap (vs. vamp) & Incap vs another player. The vulnerability was treated as additive in both cases, and was doing far more damage than it should. I'm guessing ZOS did a ninja revamp of the damage equation, and this is the norm now in Wrathstone, even with just 1 vulnerability effect and no stacking.

Prior to this update it was most definitely multiplicative (tested Incap extensively last patch).

Still haven't tried to replicate Reorx's order-of-application bug that turns it multiplicative, though, that's a whole other issue.

In theory, any truly multiplicative resist system shouldn't care about order of operations at all. So, that leads to the possibility that there's something even stranger going on.