Awesome testing done on enmity

[ItazuraNhomango]

Re: Awesome testing done on enmity

I'll make sure to include all the details once I have a hard formula. I just wanted to share what I quickly observed in an hour of experimenting last night. The 3 person setup with two testers using Stoneskin and a refresher/puller is definately optimal.

I understand, but in the greater scheme of things, the information on experimental setup and methodology is as just important as what is observed. (I would go as far as to say the observation data is nearly worthless without them.)

Talk with Kaeko about your setup and methodology; it will make your hard work even more valuable.

[Yellow Mage]

Re: Awesome testing done on enmity

Uh...guys? I just realized Chainspell + Dispel just became the most absurd thing ever.

I mean, all it takes is 32 Dispels to hit the hate cap. Chainspell Dispel can pretty much leave everyone in the dust hate-wise in the first 10 seconds of any fight, and after that, well, Dispel recast is only 10 secs. Between RDM's levels of Fast Cast, and some Haste...you can pretty cast Dispel every 6-7 secs and stay at the hate cap.

Sounds, excessive, but if you think about it, it would allow DRKs to go ape shit crazy with Souleater Bloodweapon WITHOUT TAKING HATE within the first 15 secs of a fight.

I'm scared.

Already thought it up: Sentinel + Chainspell + Dispel (among other things)

= You ain't stealin' hate from me!

( . . . ****, why do I level so slowly? I want my /PLD now! ;_; )

[Akashimo]

Re: Awesome testing done on enmity

Amusing. By theory, any rdm gear up could out tank a pld in terms of keeping hate and to later extent, lessen damage. I must do tests myself... if i wasn't sick of playing rdm at this time...

[Taskmage]

Re: Awesome testing done on enmity

Something I thought I'd point out just to have on the table, though I'm sure most already realize this. It isn't as though a player can cap out on enmity and then just sit pretty on the top of the hate list. Whenever damage is taken or a shadow is lost there is a drop in enmity that has to be compensated for. What I'd really like to see the values for are the drops from losing a shadow or taking X damage.

Also, the blog post has been edited to reflect the fact that the test does not prove that Dispel provides exactly 320 units of enmity, only that 320 units is the non-decaying component of Dispel's enmity.

[LilithAngel]

Re: Awesome testing done on enmity

What I'd really like to see the values for are the drops from losing a shadow or taking X damage.

That's pretty much what I was trying to get at when asking my questions. I'm interested in how much hate is lost when damage is given and taken, and at what rate each is expected to run at. It'll take a lot of testing, but eventually such information can pay dividends in the long run. It could help change the way the game is played if we begin to understand in depth how enmity in this game works.

[Armando]

Re: Awesome testing done on enmity

Whenever damage is taken or a shadow is lost there is a drop in enmity that has to be compensated for. What I'd really like to see the values for are the drops from losing a shadow or taking X damage.

This is true. But even so, once a RDM/NIN reaches the hate cap there's no reason to get out of there. Between Fast Cast and the Haste spell they're already at -30% recast, so they can cast Dispel every 7 secs. The hate loss from two shadows is most likely negligible compared to Dispel's hate. And god knows if there's any other spell that gives fixed hate that you could cast between Dispels (Sleep/Gravity/Bind...?)

Already thought it up: Sentinel + Chainspell + Dispel (among other things)

= You ain't stealin' hate from me!

( . . . ****, why do I level so slowly? I want my /PLD now! ;_; )

That would set hate faster, but it's kind of innecessary, and then you can't have shadows to keep that hate efficiently.

[Icemage]

Re: Awesome testing done on enmity

Another interesting point is that hits for 0 obviously do NOT deplete your enmity rating at all. Important to note for those who are doing end-game as WHM or BLM and stacking Stoneskin. Not that using Stoneskin isn't a good idea, mind, just that if you end up taking 0 damage from the first hit (unlikely but possible) from an HNM, you haven't bled off any enmity, which is a bad thing. Of course, this is still preferable to getting clocked for 350 points of damage, but still important to note.


Icemage

[Ryoii/Nonomii]

Re: Awesome testing done on enmity

Initial Boost Test

I did a quick test of Boost with the help of a MNK/WAR (Fruit, thank you!). Our testing suggested that Provoke and Boost go into a single pool which decays at a rate of 60HU per second. While Boost is weak alone (~300iHU), when combined with Provoke in a cycle, approximately 2 Dispels worth of hate is built each cycle. Very soon all those WAR/MNK and MNK/WAR tanks might actually have some proof that spamming Provoke+Boost+Boost is effective.

UPDATE: I just tested that I could pull hate from Nonomii after she casted 31 Dispels while using a Provoke+Boost+Boost cycle with my WAR/MNK. It only takes about 9 minutes to hit maximum hate. However, after 32 Dispels, I was unable to maintain hate for the whole cycle. This might open up more options for kiting HNMs.

Provoke Update

Terms:
HU: Hate Unit - a unit of hate defined as the hate (emnity) left from casting Blind and waiting 2 minutes
iHU: Initial Hate Units - initial hate generated from a spell or action
fHU: Final Hate Unit(s) - hate remaining from a spell or action after a passage of time which fully decays the hate

Given:
Blind, Stoneskin, Refresh, and Barspells all give 1fHU. I tested this myself my casting Blind followed by one of the other spells. It would take two Provokes to maintain hate.

Method:
Ryoii would pull an enemy with Provoke. Nonomii would cast Blind on the enemy and let Ryoii zone. Nonomii would cast Blind again, and Ryoii would zone back. Now Ryoii has 0 HU and Nonomii has 2 HU. This was tested by having Ryoii use Provoke twice and not loose hate the second time.

Now Nonomii would cast Dispel 5 times and cast 1 fHU spells until she has 1801 HU. Now Ryoii would zone out and back to reset her hate to 0. Now it would require that Ryoii use Provoke twice to get the enemy to turn away for Nonomii for only a second. I waited about 5 minutes and tried several more Provokes which still caused the enemy to turn for only a second.

This would be much easier with another RDM casting Refresh and Phalanx II. I'll make 100% sure this weekend that I have this all correct. I'm still trying to figure out the decay. It appears that it might be 60 per second, which would fit reasonably well with my previous data except for the error about 5 Dispel > 1 Provoke.

5 Dispel = 1600 fHU, so hate is lost from Provoke after 4 seconds.
4 Dispel = 1280 fHU, so hate is lost from Provoke after 9 seconds.
3 Dispel = 960 fHU, so hate is lost from Provoke after 14 seconds.
2 Dispel = 640 fHU, so hate is lost from Provoke after 20 seconds.
1 Dispel = 320 fHU, so hate is lost from Provoke after 25 seconds.

Current Theory:
Provoke = 1800iHU and 1fHU with a loss of 60HU per second

Comments:
You will see below that I never got to exactly 1801 HU, but I believe some of the following may have happened: I miscounted my Barspells, Dispel is 321fHU, or Provoke is 1801iHU.

Raw Data for Attempt #1:
Spell, Nonomii's Total HU
Blind 1
Stoneskin 2
Refresh 3
Blind 4
Dispelx5 1604
Refresh 1605
Stoneskin 1606
Barx6 1612
Barx6 1618
Barx6 1624
Barx6 1630
Barx6 1636
Refresh 1637
Stoneskin 1638
Barx6 1644
Barx6 1650
Barx6 1656
Barx6 1666
Refresh 1667
Stoneskin 1668
Barx6 1674
Barx6 1680
Barx6 1686
Barx6 1692
Barx6 1698
Refresh 1699
Stoneskin 1700
Barx6 1706
Barx6 1712
Barx6 1718
Barx6 1724
Barx6 1730
Refresh 1731
Stoneskin 1732
Barx6 1738
Barx6 1744
Refresh 1745
Stoneskin 1746
Barx6 1752
Barx6 1758
Barx6 1764
Barx6 1770
Barx6 1776
Refresh 1777
Stoneskin 1778
Barx6 1784
Barx6 1790
Barx6 1796
Required 2+ Provokes

Raw Data for Attempt #2:
Spell, Nonomii's Total HU
Blindx2 2
Refresh 3
Stoneskin 4
Dispelx5 1604
Barx6 1610
Barx6 1616
Refresh 1617
Stoneskin 1618
Barx6 1624
Barx6 1630
Barx6 1636
Barx6 1642
Refresh 1643
Stoneskin 1644
Barx6 1650
Barx6 1656
Barx6 1662
Barx6 1668
Refresh 1669
Stoneskin 1670
Barx6 1676
Barx6 1682
Refresh 1683
Stoneskin 1684
Barx6 1690
Barx6 1696
Refresh 1697
Stoneskin 1698
Barx6 1704
Refresh 1705
Stoneskin 1706
Barx6 1712
Barx6 1718
Barx6 1724
Blindx2 1726
Refresh 1727
Stoneskin 1728
Barx6 1734
Barx6 1740
Barx6 1746
Barx6 1752
Blindx2 1754
Refresh 1755
Stoneskin 1756
Barx6 1762
Barx6 1768
Barx6 1774
Barx6 1780
Refresh 1781
Stoneskin 1782
Barx6 1788
Barx6 1794
Required 2+ Provokes

[Armando]

Re: Awesome testing done on enmity

Wait. Are you saying Provoke's initial hate spike will pull hate from 5 Dispels?

[hongman]

Re: Awesome testing done on enmity

Uh...guys? I just realized Chainspell + Dispel just became the most absurd thing ever.

I mean, all it takes is 32 Dispels to hit the hate cap. Chainspell Dispel can pretty much leave everyone in the dust hate-wise in the first 10 seconds of any fight, and after that, well, Dispel recast is only 10 secs. Between RDM's levels of Fast Cast, and some Haste...you can pretty cast Dispel every 6-7 secs and stay at the hate cap.

Sounds, excessive, but if you think about it, it would allow DRKs to go ape shit crazy with Souleater Bloodweapon WITHOUT TAKING HATE within the first 15 secs of a fight.

I'm scared.

Holy shit.

[ItazuraNhomango]

Re: Awesome testing done on enmity

I dunno... I may prefer Chainspell Stun + Aspir. That way I'd be alive when Chainspell ends, and have MP, too. (Only works for stuff which has MP and isn't an Undead, of course.)

[Kirsteena]

Re: Awesome testing done on enmity

The mp usage on CS Dispel is phenominally high - 32 Dispels = 800 mp. You would be risking an awful lot on a convert when doing this.

Also, thinking about when I CS stun, and there is a zerg going on, I can't pull hate for the life of me, and in our shell for DL, I'm generally the first to go, with the rest of the rdm taking their timing from me.

[hongman]

Re: Awesome testing done on enmity

Althought I think the whole point of this is to have the mob pretty much dead within the time limits for 2hr running out, otherwise people wouldnt be all going apeshit to start with anyway.

EDIT: CS Stun is great and all, but mobs still do moves through, and if they have any kind of stun move you're gonna break that Cs cycle. Added to the fact if you are tanking something that others are zerging, likely hood its going to be a higher lvl mob, your Aspirs are going to be neglible at best.
------------------------------------------

The mp usage on CS Dispel is phenominally high - 32 Dispels = 800 mp. You would be risking an awful lot on a convert when doing this.

Also, thinking about when I CS stun, and there is a zerg going on, I can't pull hate for the life of me, and in our shell for DL, I'm generally the first to go, with the rest of the rdm taking their timing from me.

Dont forget though, our PLD's are all popping Invincible and have +Emn gear. Presuming Invincible puts them straight at their max hate cap (which is higher than yours sue to Emn gear), thats why you cant pull hate.

I know from CS Stunning things like Suzy where noone pops Invincible, I pull hate nearly everytime, with my poor buffer as witness.

[Armando]

Re: Awesome testing done on enmity

If you're going /NIN, you could spam Utsusemi as needed during the Chainspell (along with a Stoneskin in there for good measure,) and it'd be better than Stun at ensuring that you don't take damage. Plus, on anything that's actually a threat, going /DRK for Stun and Aspir would ultimately mean a lot less hate after Chainspell is over since you're going to get hit, and getting hit hard loses more hate than losing a shadow as /NIN.

That aside, odds are Stun and Flash are both decaying hate. Reason Flash is so good for PLD is probably the fact that it prevents you from getting smacked for about 12 seconds, and that's 12 seconds you didn't lose permanent hate.

There's something I believe we need to look into. I recall an interview a year or two ago with S-E, and they inquired about Enmity and Provoke. If my memory serves me, S-E replied that Enmity doesn't affect Provoke, but that due to the nature how Provoke works, a higher level player makes more hate with Provoke than a lower level one. This sounds extremely fishy, but we need to check to make sure. Could have something to do with Ryoii/Nonomii making more than 5 Dispel's worth of hate with Provoke, too.

[Karinya]

Re: Awesome testing done on enmity

I think what you're suggesting for chainspell dispel is substantially what already happens with chainspell stun. The rdm/drk will rip hate off anything and everything, and it's typically only the fact that the target spends most of its time stunned that keeps it from snacking on you. (In some cases they manage to do something anyway; I've eaten flares from Suzaku while chainstunning him during his own chainspell, for example.)

Although it's possible that stun decays more, and it's only the fact that the stuns are all done in rapid succession that pushes them up to such a ridiculous hate value.

In any case, if this does become abused somehow, it would be simple to fix: greatly lower the enmity of dispels that don't dispel anything. Then you can't cast 32 dispels for full hate unless the target has 32 dispellable effects on (not bloody likely), but the enmity of dispel in ordinary situations wouldn't be affected (you're only going to cast it when you expect it to have an effect).

P.S. "Non-decaying" hate is somewhat misnamed: in real fight situations, it will still be reduced when the monster takes a swing at you and does damage or removes shadows (the part that is too complex to easily see by this type of test). So the hypothetical dispel-tank would also need to kite and consistently stay out of range/never have their stoneskin broken to maintain hate (/nin wouldn't really help because, although you might stay alive by using it, you'd shed the hate you 2houred to build). Regularly renewed spike hate can be just as important as "permanent" hate over the long term when hate is being lost to hits faster than it can be lost over time.

While this information *is* useful, don't try to apply it beyond its actual applicability. The test scenario contains several simplifications not present in the typical real fight.

P.P.S. It's unfortunate that their testing methodology requires illegal hacks. Maybe there's some setting where distance can be calibrated by other means (tiles on the floor?) that would allow non-hacking players to contribute data?