Gaming
 

Talk:Enmity Numbers

From FFXIclopedia, the free Final Fantasy XI encyclopedia

Requested Move

Move to Subpage/Merge This should either be a subpage of Enmity or merged with the actual article itself. I'm leaning towards the latter as I think this information could well enhance the article itself, perhaps after it's finished make it into a nice little table. --CharitwoTalk 08:02, 9 August 2007 (CDT)

I'd like to request nothing of the sort happens until the testing methodology is better explained and some verification on its validity can be done. --Itazura 04:34, 11 August 2007 (CDT)

Resolution: Do not move for now, pending further verification and testing. --◄Chrisjander► t/ c 19:10, 6 September 2007 (CDT)

Source?

Before we decide about incorporating this into the main Enmity page, where is the source of this info? Where can we find the data on the tests that provided this? --◄Chrisjander► t/ c 10:01, 9 August 2007 (CDT)

Unfortunately, there are no other sources of this info. I've written what I felt to be accurate based on some tests I've done. I didn't add this info to the Enmity page because there were some stuff I didn't agree with and I didn't know how to or if I should change it. --JKL 21:58, 9 August 2007 (CDT)

Do you have your testing methods to post? Otherwise many people are going to be skeptical without knowing where this is coming from. It's one thing to say "Last Resort produces slightly less enmity than Provoke" (which anyone can easily see) and another to start quoting numbers. I'm not saying your information is wrong, I'm just requesting that the method it was obtained be a little more transparent for everyone. --◄Chrisjander► t/ c 22:46, 9 August 2007 (CDT)

I've always believed in a binomial nature of what players generally consider as "hate"; my theory divided hate in two components, "dynamic" and "static" hate, the former noticeably increased (spiked as many people like to say) by abilities such as Provoke and Flash whose decay time must be considerably small for obvious reasons and the latter (which may or may not be somewhat affected by CHR, doubtful though) slowly and gradually until a "plateau" is reached, gained performing actions on the mob for extended periods of time that apparently decreases greatly upon taking damage, have always been part of my own personal way of depicting hate management.

Just my two cents, I foresee some sort of potential in the discussion that could stem out of the OP article.

--SearainGaruda 22:40, 9 August 2007 (CDT)


On a definately distinct yet quite inherent note I'd say that Provoke generates much more than just "1" unit of "static" or "residual" enmity: a War/Nin wearing full enmity gear and using only Provoke and the occasional Warcry would usually take quite a few minutes to build a reliable amount of hate while kiting a mob like Kirin, however, despite the fact he'd have many troubles to keep up at the beginning, he might be capable of holding hate almost flawlessly in a magnet-like way after a while until he eventually takes enough damage to get pushed back in the hate list or doesn't perform any action on the mob for some time (case in point being the extremely noticeable loss in hate you could recognize in the aforementioned situation right after surviving a Stonega that did above 1000 damage almost instantly).

--SearainGaruda 23:05, 9 August 2007 (CDT)

I think it's possible for the war/nin to accumulate temporary enmity if he has +enmity and spams provoke but you mention that the war/nin loses hate if he takes damage, which contradicts my belief that taking damage doesn't lower temporary enmity... so I don't know....

The temporary enmity of provoke with +0 Enmity wears off after 30s, which is the same as the ability recast time, so spamming it cancels out the natural decay of temporary enmity. With +Enmity > 0, you'll gain enmity over time.

--JKL 04:48, 10 August 2007 (CDT)

The tests were done with 2 characters and low level mobs but it's hard to describe all the tests in detail because they were numerous and most of them were done differently. For most of the tests, I start with both characters with the same amount of enmity, make one character do an action, whose residual enmity I want to measure, and then use the other character to repeatedly do actions with known enmity values until I can tell they are both equal. For estimating temporary enmity, I use a stopwatch. When doing these tests, I have to make sure the characters don't take damage (+def, +vit, phalanx, stoneskin).

Curing a LV75 character gives the caster [Amount Healed x ~0.728] REP (residual enmity points) plus 4+ or 5+ times the amount of REP in TEP (temporary enmity points).
Cure for 0 HP = 1 REP + ? TEP
Cure for 6 HP on a LV75 = 4 REP + ? TEP
Cure for 30 HP on a LV75 = 21 REP + ? TEP
Cure II for 55 HP on a LV75 = 40 REP + ? TEP

Anyone who wants to test this stuff is welcome to post their findings.

--JKL 04:48, 10 August 2007 (CDT)

What testings were done to prove: 1. "Temporary enmity" exists as a separate quantity from "Residual enmity"? 2. "Residual enmity" is a variable quantity? 2. Enmity decays over time? --Itazura 04:42, 11 August 2007 (CDT)


Have a +0 Enmity BLM claim a LV1 Wild Rabbit with Sleep and then follow with Sleep II, Sleep and Sleep II (200 + 320 + 480 + 320 + 480 = 1800 residual enmity). Wait until it wakes up. Have a +0 Enmity WAR Provoke it repeatedly (1 res. enmity + 1800 temp. enmity). The WAR will hold its attention for less than 1s and may not even turn it. The WAR's provoke will hold the mob one additional second for every 60 provokes he's done. Time has no effect on the BLMs enmity, since his temporary enmity is already at its minimum. The temporary enmity of provoke decays to nothing after 30s, which is the recast of ability. Keep in mind that the BLM and the WAR should not take damage in order to maintain their hate.

If someone else could reproduce the test and post the results, that would be great.

--JKL 02:05, 13 August 2007 (CDT)

Why so complicated with so many Sleeps and Provokes? If one Sleep spell supposedly has much greater residual enmity than Provoke, all you need is one worm (test subject), one Sleep spell (Player 1), one Provoke (player 2), and on player to keep the worm Silenced (player 3). Over time, it should stop pointing at the Provoke user (player 1), and swing over to point at the Sleep user (player 2). (Seeing that worm changing target would be evidence for enmity decay, BTW, if residual enmity exists and is at different value for Sleep and Provoke.)
The more complicated the testing, the less confidence I have in the interpretation of the outcome. --Itazura 13:55, 13 August 2007 (CDT)

I think that's the opposite, using more spells introduces more variables which should make it harder to predict the exact behavior and lessen the chance of a coincidental prediction. I didn't base the numbers shown on the article page just to make this test work. I measured each value of residual enmity separately and mostly by matching it with known enmity values of a few Cures. I didn't choose nice numbers such as 180, 240, 320, 480 just to make them look nice. They were exact matches.

The suggestion you are making is fine. I would avoid worms though, because they lose the hate of its current target if it spends more than 30s without touching someone. You can even pull with a Slow for (200 + 1) res.enmity. The res.enmity of Provoke is far too low to overcome it in practice. I still think that using 4 sleeps method is easier to understand since the provoker will only hold hate for < 1s, compared to the 20s+ of just using one sleep.

--JKL 21:29, 13 August 2007 (CDT)


Looks like we were onto something:

somewhat official explaination of the two components of Enmity taken from a JP guide and posted on BG

--SearainGaruda 07:00, 14 August 2007 (CDT)


Thanks Searain, that was a very interesting thread. I'm thinking about changing the terms I used for the ones used in the JP guide ("Volatile" and "Cumulative"). I've figured out the numbers of the most interesting spells/abilities except for Bind, Cures volatile enmity and Damage enmity.

--JKL 14:46, 14 August 2007 (CDT)

It would raise the community's confidence level in the numbers and the methodology if you can at least describe in detail how you get the "first" number. Given that you rely on known values to derive enmity numbers for each action, how did you "know" the first value you used? --Itazura 14:19, 16 August 2007 (CDT)


Well, at the beginning I used 'Provoke units' to measure cures until, after having tested other stuff, it became evident that provoke generated 1 unit of cumulative enmity. I guess Provoke is the first 'number'. Cures for 0 HP produces the same amount of cumulative enmity as one Provoke but it doesn't have a 30s recast.

I redid an old test for clarity. The following describes how it went:

  • A = +20 Enmity WAR/NIN
  • B = -1 Enmity NIN/WAR
  • mob = Wild Rabbit


  1. B voked >> held hate
  2. A voked right after B >> B still held hate
  3. B zoned >> A held hate
  4. A voked 5 times more in a row
  5. waited 1min
  6. B voked >> B held hate for 30s
  7. B voked >> B held hate for 30s
  8. B voked >> B held hate for 30s
  9. B voked >> B held hate for 30s
  10. B voked >> B held hate for 30s
  11. B voked >> B held hate
  12. 8mins later
  13. B took 1dmg crit, A held hate


From this test, I'd say that:

  • 6x Provokes with +20 Enmity = 6x Provokes with -1 Enmity (ignoring volatile enmity).
  • Provoke cumulative enmity = 1 point.
  • Provoke volatile enmity lasts for 30s, same as the ability's recast.
  • Getting hit for 0dmg doesn't lower enmity.

--JKL 18:53, 17 August 2007 (CDT)


That's not exactly accurate, the first action used on the mob gets more hate, especially with something like provoke. you'd be best to have a third char pull the mob, or atleast aggro pull, or use something else that gets very little hate. 
The stat 'Enmity' multiplies the amount of enmity gain by (1 + Enmity / 100).

i can gaunrentee you that that's wrong. everything in the game is done in measures of 256, not 100. why would enmity be any different? --rOg \ TaLk 00:08, 13 October 2007 (UTC)


Step 3 of the test above takes care of the extra initial hate. Zoning isn't even needed if your "pulling" action misses or is resisted. This is based on some tests I've done.

About "1 + Enmity / 100", it's basically saying that ±1 Enmity = ±1% hate. There are other stats that are percentage increases without explicitly saying so such as Double/Triple Attack, Killer traits, Conserve MP, Store TP, Subtle Blow, ...

--JKL 06:14, 13 October 2007 (UTC)

Specify model

It may be a good idea to preface this guide with an explanation that it is based on the "Volatile and Cumulative" model. While it is explained on the bottom, an introduction paragraph would be nicer for the readers. (I'd be even happier if it's also explained that "Volatile and Cumulative" is not the only possible model, but that decision should be left the primary author, JKL.) --FFXI-Itazura 23:35, 27 September 2007 (UTC)


This guide is closer to a table with a few notes than a guide and I like it this way. I like pages with just raw data with not too many explanations, similar to VZX FFXI Doc website. I'm not a good writer so I don't think I would have done a good job writing an explanation anyway. --JKL 06:14, 13 October 2007 (UTC)

Retrieved from "http://wiki.ffxiclopedia.org/wiki/Talk:Enmity_Numbers"