Home

Advertisement

Previous Entry | Next Entry

Enmity Testing (Part VI)

  • Nov. 19th, 2007 at 9:59 PM

Overview

There are only 2 tests in this section; however, they are pretty important, so I've decided to go ahead and post them seperately.  After this section, we will have essentially found most of the basic concepts of Enmity other than Cure, Damage Taken, and Damage Dealt.  We're already starting on that data with the help of some other groups; however, it is proving to be pretty difficult at this point.

In this post, we'll go over 2 major concepts...

      -  Does Volatile Enmity (VE) "cap"?  If so, where?
      -  What is the distance at which CE and VE gain are no longer active?


The first of these is critical in developing both tanking strategy and understanding damage dealing.  While Cumulative Enmity has been the major focus of most of these tests and a lot of discussions around the community regarding tanks, I believe it is actually VE that may play a more vital role in most situations.  This is simply because most fights are too short to accumulate massive amounts of CE.

The second finding is something I think BSTs and SMNs are pretty familiar with already.  It is well known just throw experiencing these jobs that at a certain distance from the mob, any action you do like /healing or curing yourself no longer grants enmity.  We simply find this "fringe distance" (made this term up) through a series of very simple tests.

In addtion to these 2 tests, I have also gone over some methodolgy on how to perform testing on Cures I through IV.  Ashira and I need some help on these and we really can't do it alone due to the dependence on level.  Anyone help would be greatly appreciated on that.  I go over the process of how to perform the testing in the post.


Special thanks to Tour (Midgar) for running the "VE cap test" for us.  We knew this test was a big deal, but Ashira and I don't have access to a level 30+ PLD, so were unable to run it.  I would also like to thank all of those contributing to the growing table.  We're roughly just over half way done at this point


Previous:  Enmity Testing (Part V)

Next:  Enmity Testing (Part VII)

Enmity Table

____________________________________________________________________________________________________
____________________________________________________________________________________________________

Test 17

Is there a Volatile Enmity (VE) Cap?  If so, what is it?

      H0:  Volatile Enmity (VE) does not cap
      H1Volatile Enmity (VE) does cap

Testing Setup:

- 1 puller character with Silence access (BRD/WHM)
- 2 level 75 members with Dispel, one of these a PLD over 30 (PLD/RDM, RDM/WAR)
- 1 character must have access to enmity gear, and both need to have access to "distance plug-in"
- Use a "level 0" monster - we chose Bumblebee in West Sarutabaruta
- This test requires the use of a stopwatch!

- All characters must have either Stoneskin or Phalanx buffed prior to starting to ensure 0 damage
- Utsusemi shadows must be off when this experiment takes place

- Puller pulls with Silence while the 2 test characters move to 0.7 distance (same distance)
- Both test character 1 and 2 cast Dispel once (CE1 = 320, CE2 = 320)
- Player 1 (non-PLD) casts Cure 1 for 0 HP 5 times (CE1 = 325, CE2 = 320)
- Player 2 (PLD) uses Sentinel (CE1 = 325, CE2 = 322)
- Player 2 (PLD) uses Invincible with active Sentinel Effect (CE1 = 325, CE2 = 324)
- Document the take taken for Player 2 to lose hate back to Player 1

      H0:  The time taken for Player 1 to regain hate will be greater or equal to 4 minutes
      H1The time taken for Player 1 to regain hate will be less than 4 minutes

Time Taken to Regain Hate =  ~167 seconds

The results show that the Invincible wore off in roughly 2 minutes and 47 seconds.  If the Volatile Enmity had not been capped, the Invincible with Sentinel Effect (+100 Enmity, or double CE and VE) should have granted 14400 VE, which would have taken about 4 minutes to completely wear.  Because the actual time fell well short, we must conclude that the Volatile Enmity does cap.

Other than just finding that the cap exists, the actual VE cap can be estimated quantitatively using the same data!  We know that the decay rate is -60 VE/sec, so we just multiply the time taken in the test by 60 VE/sec and we have our estimated VE cap...

                    [Time Taken] x [VE Decay Rate] = [167 seconds] x [60 VE/second] = 10,020 VE

Notice how amazingly close this number comes out to be to the Cumulative Enmity cap at 10,000 CE.  At this point, I think it would be safe to assume that SE capped these both at exactly 10,000 E.  The margin of error on this test, due to the fact that we are using a stop-watch is probably within that realm.  I think it is a pretty safe assumption to say that Volatile Enmity does cap, and does so at exactly 10,000 VE, the same cap as CE.


Conclusions

Volatile Enmity does cap and does so at 10,000 VE - the same cap as CE.
The maximum Enmity any player can reach is 20,000 TE - 10,000 CE and 10,000 VE.

____________________________________________________________________________________________________
____________________________________________________________________________________________________

Test 18A

What is the minimum distance a player can perform an action and not receive enmity gain?

- 1 puller character with Silence (BRD/WHM)
- 1 WHM with access to Cure V (WHM/WAR)
- Both need to have access to "distance plug-in"
- Use a "level 0" monster - we chose Bumblebee in West Sarutabaruta

- Puller pulls with Silence and stands still
- The WHM casts Cure 1 for 0 HP on the puller to get on the hate list.
- Have the WHM go to a distance of about 35' and cast Cure V on himself.
- Move in closer and closer, repeating cures until the WHM is finally able to take hate.
- Document the closest distance at which the WHM could Cure V without taking hate.


Maximum Distance that still got Hate =  24.9
Minimum Distance that did not get Hate =  25.0


Based on this test, we can say that any action you perform on yourself at a distance of 25' or greater from the mob gives no Enmity.  I call this distance of 25' the "Fringe Distance" - if you are outside this distance from the mob, nothing you do can give you more Enmity.  Note that it is possible to still get hate outside this distance - if the player with hate is shedding CE or VE and you then overtake hate, the mob will go after you, even if you are outside the 25' distance.  I think that is pretty intuitive.

This really just pertained to actions you can perform on yourself.  Performing an action on the mob from greater than 25' would most certainly still give you Enmity - I don't know if you can actually perform an action on a mob at this distance - maybe a ranged attack?  Either way, this test only shows how things work on actions you perform on yourself.  The next test goes into details about performing actions on an ally.

               ________________________________________________________________________________

Test 18B

Is the "Fringe Distance" dependent on the player performing the action or the player the action is targetting?

- 1 puller character with Silence (BRD/WHM)
- 1 WHM with access to Cure V (WHM/WAR)
- 1 character with access to Cure 1 (BLM/RDM)
- Non-pullers need to have access to "distance plug-in"
- Use a "level 0" monster - we chose Bumblebee in West Sarutabaruta

- Puller pulls with Silence and stands still
- The non-WHM casts Cure 1 for 0 HP on the puller to get on the hate list.
- Have the non-WHM go to a distance of exactly 24.9' from the mob
- Have the WHM stand at the maximum casting range relative to the other player
      - The WHM should be roughly 21' from the player and about 46' from the mob

- Have the WHM now cast Cure V on the player 24.9' away from the mob
- Document if the mob runs for the WHM or stays on the puller.


Result =  The mob will run after the WHM in this scenario

This test shows that it doesn't matter where exactly the caster is from the mob - only where the target of the action is relative to the mob.  If you really think about this in practice, this makes a lot of sense.  If the caster's distance is what mattered, you could really abuse this situation in certain endgame fights.  In many cases like self curing, the target of the action is also the caster, but it is still important to differentiate and pin-point exactly what is going on.

I think this test is most important to BSTs and soloing SMNs, as this distance is incredibly important in making sure the pet does not lose hate to the mob.  Even so, I was quite surprised at how close this distance is from the mob itself.  I had heard many theories on where this distance was and 30' and 50' were the most common things I had heard.  25' is actually just outside the range of some mobs.  I know from soloing experience that Behemoths can be cast on anywhere 23.8' and under, so the "Fringe Distance" is really only 1.2' out from the maximum casting distance.

               ________________________________________________________________________________

Test 18C

This test only serves as further verifcation for Test 18B...

Repeat Test 18B, but this time, have the target of the cure stand 25.0' from the mob, while the caster of the cure at about 6.0' from the mob (but still in range to cure the target).  You'll find that the caster will not get any Enmity in this case.  Now move the target to 24.9' instead of 25.0' and repeat - the curer will now take hate.

This shows that, indeed, the target of the action is all that matters when calculating the fringe distance.  The caster himself can be really close to the mob, but still get no enmity as long as the target is 25'0 or greater away from the mob!


Conclusions

If an action is performed on a player at a distance 25.0' or greater from the mob, no Enmity is gained.

____________________________________________________________________________________________________
____________________________________________________________________________________________________

The Cure Formula - Help Needed!

If anyone wants to help us out, that would be greatly appreciated.  Ashira and I are going to work on the Cure formula first; however, we have found that we need players of different level - something we really don't have much access to, sadly.  I want to quickly go over the testing methodolgy we're using to find the CE of Cures I through IV.  Note that we are not touching VE until we can clearly establish how CE works.

This testing methodolgy works for 2 or 3 characters, although 3 makes this faster and easier in my opinion.  I recommend using RDMs if you have access to them over level 40 for Convert.  Using Convert makes it easy to lose a lot of HP for Cures.  RDM/WHMs are ideal for this testing in my opinion, since they get Divine Seal Cure IV.  I'll go over why it is beneficial to Cure large amounts of HP for this testing.

Step 1
- Have the character that you want to cure (with variable LVL and MAX HP) lower his HP as much as possible
- Try to get the lowest HP possible so you can get off a maxed Cure - 400ish normally or 800ish with Seal
- Convert is generally the easiest way to do this if you have a RDM greater than 40.

Step 2
- Pull a mob and then have both players cast a decently high level CE spell (we use Dispel).
- If you do not have access to 3 characters, you can zone the puller than cast that high level CE spell and check.
- Verify that both characters for testing are at the same CE value before proceeding.
- Note that if you are using Divine Seal, it grants 1 CE so make sure to balance that as well

Step 3
- Once CE is equal and verified, have either player cast the largest Cure possible on the player with low HP
- Note that the TARGET being cured is what matters in the formula, not the caster
- The caster will get some amount of CE from the Cure
- Use various methods to find the CE of that cure
      -  I personally like using Utsusemi Shadow Loss to determine differences
      -  BRD songs are good ways to add CE in smaller units
      -  This requires a bit of creativity on sub
      -  Note that if you use Cure to step up, make sure you are Curing for 0 HP only! (we messed up a lot on this)


Document the following pieces of information:

      (1)  The level of the player cured
      (2)  The maximum HP of the player cured when the cure occurred
      (3)  The amount of HP cured
      (4)  The CE obtained from the cure by testing


I can calculate the estimated multiplier with the information given above.

Using this information at various levels and HPs, we can try to determine if LVL, MAX HP, or even both play a role in the formula.  What we do know is that the amount of CE appears to scale down per HP cured as level increases (although this could be an effect of having higher HP).  I personally think that only level of the target matters; however, I won't rule out MAX HP having an effect at this point.

Our Preliminary Numbers for LVL40, 70, and 75 TARGETS are...
  
                    At 40:  1.1440 x [HP Cured] = [CE Gained]
                    At 70:  0.7700 x [HP Cured] = [CE Gained]
                    At 75:  0.7275 x [HP Cured] = [CE Gained]

The number at LVL75 is pretty exact to 4 digits; however the other 2 may only be accurate within +/- 0.01.  You can use this information to kind of predict around where the CE will fall when testing other levels.  I did not document HP values for these; however please do for this.

If anyone wants to try to verify if MAX HP even has an effect on the multiplier, that would be great as well.

____________________________________________________________________________________________________
____________________________________________________________________________________________________

Conclusions

Here is the summary of conclusions reached in this post...

      - Volatile Enmity does cap and does so at 10,000 VE - the same cap as CE.
      - The maximum Enmity any player can reach is 20,000 TE - 10,000 CE and 10,000 VE.

      - If an action is performed on a player at a distance 25.0' or greater from the mob, no Enmity is gained.

This was a pretty short post, but I think the 2 tests performed in this post are of particular importance in rounding out a good base knowledge of how Enmity works.  From here on out, we'll be dealing almost strictly with real formulas for Cures, Damage Taken, and Damage Dealt.  These will most certainly be the hardest things to really nail down in this Enmity Testing.

As stated previously, Ashira and I want to work on the Cure Curve for now in addtion to continuing the table.  I hope that we can get more help still on this, since the hardest 3 things are still undone.  Once again, thanks to everyone that has contributed to the testing thus far.  We're really coming a long way very quickly.  Recently, my blog has also been getting a lot of hits from Japan, and I was able to get the link using Statcounter...

Japanese Version of the Enmity Table

Apparently, this made 2chan, one of the most popular JP BBS.  It'll be interesting to see how the information plays out.

____________________________________________________________________________________________________

Pic of the Day

OMG MOOGLE CAPS!!

img249/3244/img20071117120948hy3.png 

img211/4607/img20071117121951fw2.png 

Some of the JPs in LS all pitched in and got every NA in TK their very own Moogle Cap from the Tokyo Game Show.  Super thanks to Gyamada, Ryagi, Bing, Tabite, Shall, Radish, Forever, Nakamoon, Mizz, and Yuchi. <3<3



Drama Thread of the Day

http://killingifrit.com/forums.php?m=posts&q=140225

If you're on Carbuncle and read KI, I think you know about this.  I can't tell if this is a drama thread or not... all I know is it has got the be the freaking longest topic I have ever seen... over 270 pages and 8000 replies... I have no idea what is going on in it either - there doesn't seem to be any real topic to the post at all.



____________________________________________________________________________________________________
____________________________________________________________________________________________________

Comments

( 31 comments — Leave a comment )
[info]aesthetik_dekay wrote:
Nov. 20th, 2007 06:04 am (UTC)
No drama thread of the day!?!?
[info]kanican wrote:
Nov. 20th, 2007 06:25 am (UTC)
I didn't think people read those lol...

I'll edit one in, you're like the 5th person to already mention it :(
[info]aesthetik_dekay wrote:
Nov. 20th, 2007 06:30 am (UTC)
I usually look at those before the rest haha. Like the picture of the day..and drama thread...and w/e else you usually put!
(Anonymous) wrote:
Nov. 20th, 2007 08:15 am (UTC)
fringe distance
some test on fringe distance vs some NM would be cool :D I'll pay more attention to stuff like that and post. I think kirin has a greater than 30' i went to 40 or so and cure IV myself and it came after me, coudl be coincidence though!
[info]kanican wrote:
Nov. 20th, 2007 08:33 am (UTC)
Re: fringe distance
I can try that, but keep in mind, just because it goes after you does not mean you necessarily gained hate. You could have been high on the list to begin with and people above you dropped down.

25' does seem a bit low basd on experience with some kited fights like Kirin. We'll have to look into that.
(Anonymous) wrote:
Nov. 20th, 2007 10:16 am (UTC)
Re: fringe distance
The fringe-distance thing bugs me minorly, only because the fringe casting distance is variable based on the size of the model. Bees are about 20, but you can get 25ish on larger mobs like HNMs

Just would like to see clarification, to be sure. It's entirely possible/likely that it's same as agro detection, static regardless of model.

Much respect for all this testing

--Xanthu, Asura
[info]sarialseraph wrote:
Nov. 20th, 2007 06:47 pm (UTC)
At 40: 1.1440 x [HP Cured] = [CE Gained]
At 75: 0.7700 x [HP Cured] = [CE Gained]
At 75: 0.7275 x [HP Cured] = [CE Gained]

You had 2 at 75s, I assume the second is 70?

What is the reasoning for curing for health on the first cure? Would it not be easier to test by doing a 0 HP cure, and then using Cure 1 until it unbalances?

Also, I was interested in the hate lost by logging off. How much info would be needed to determine that? I was thinking if two people used Dispel, one logged off and on, timed the log off time, waited for all VE to diminish, and then attempted to unbalance the hate. Would that be sufficient? I assume you would have to do multiple tests to determine whether the time spent offline changes the amount of hate lost?

Great work by the way. Awesome that someone is doing this, and it's extremely helpful for both my jobs, NIN and BLM^^ Keep up the work, and I hope to be able to assist.
[info]sarialseraph wrote:
Nov. 20th, 2007 06:51 pm (UTC)
I tried to edit my post, but I don't seem to be able to. The reason I was confused by the Cure IV actually healing HP was that would that not cause VE, which you are trying to not account for yet?
[info]kanican wrote:
Nov. 20th, 2007 11:09 pm (UTC)
It causes a ton of VE, it's just impossible to test VE when you don't know CE, so we don't worry about it yet.
[info]sarialseraph wrote:
Nov. 20th, 2007 11:43 pm (UTC)
What about the log out hate lost? Think it's possible to test?
[info]kanican wrote:
Nov. 21st, 2007 12:05 am (UTC)
From my experience if you were to log off and then back on immediately, it would not cause any CE loss at all. I think it's an all or nothing loss rather than a decay. It wouldn't be too difficult to test, but I already have so many requests on things, I don't really know when and where to fit that in.
[info]sarialseraph wrote:
Nov. 21st, 2007 12:17 am (UTC)
I was thinking I could attempt it and help out if you thought it would be worthwhile/possible to test, actually.
[info]kanican wrote:
Nov. 21st, 2007 12:52 am (UTC)
It's definitely a good test. Logging out is something people still use sometimes so it is important.

Let me know how it goes if you get around it testing it out!
[info]oliveman40 wrote:
May. 13th, 2008 07:34 pm (UTC)
I will need more data, but the best equation based on this data I can gather, CE gained from cures, is:

CE = 3.596116263-.6647849851*ln(level)

or, doing some rounding...

CE = 3.6-(133/200)ln(level)

Again, we'll need some more levels to strengthen the formula before we can make it official.
[info]kanican wrote:
May. 13th, 2008 09:54 pm (UTC)
I really have a hard time seeing what you did ;o

Also what is "ln(level)". Did you base this off the #s in here or extra numbers?
[info]oliveman40 wrote:
May. 13th, 2008 10:49 pm (UTC)
I used the backwards equation rendering (don't know the exact name) available as a function of any TI-83 calculator, and probably many others. I plugged into a table both the values for level and proportional CE, and then created multiple types of equations from that, to see which fit best. So far, the closest one was the ln, or natural log, equation that I listed above. Basically it is the best fit for that type of equation, and that type is the best fit overall for the original numbers. Try it out, you'll see.

Still, I will need more numbers (lvl 1 being the most important I think) to get the best equation possible. With the current equation the CE gained per hp cured would be 3.6. So the real number will get us the real equation (maybe this is it)
[info]aurik_ffxi wrote:
Nov. 20th, 2007 07:51 pm (UTC)
Another test you can run is to find the maximum distance away that a monster will consider you as a new target:

In Tahrongi canyon, over 50' from the crag, have player A use dispel 10+ times.
Have player B cure player A for 0.
Have player A teleport-mea.

Player A will retain his massive amount of CE. However, the mob cannot find A and will target B. Have A slowly approach the monster. Record at what distance the monster turns off of B and moves towards A.

[info]kanican wrote:
Nov. 20th, 2007 11:09 pm (UTC)
We'll try this out along with the other distance related tests next time.
[info]kanican wrote:
Nov. 21st, 2007 04:46 am (UTC)
Just tried this at about 100'+

The mob still went flying for the player at the crag. I have a feeling if we had done this on an NM, it would be flying across the entire map if we seperated far enough. I don't know exactly what to make of it at this point.
(Anonymous) wrote:
Nov. 21st, 2007 04:22 am (UTC)
Could you edit the "Cure test help" section, to ensure that they are done without the use of any equipment that may enhance the potency of the spell (Light Staff, Noble's Tunic, etc).

It may or may not have an effect, but just incase the additional +% (or MND, even) isn't accounted for when generating Enmity, it may be best to exclude the use of them from said tests.
(Anonymous) wrote:
Nov. 21st, 2007 04:27 am (UTC)
In fact, it could be more complicated.

The fact as you're now dealing with spells that vary depending on certain stats, it could make these spells a lot harder to test.

So perhaps removing these variables initially, or discovering them, will help with the final equations.
[info]kanican wrote:
Nov. 21st, 2007 04:43 am (UTC)
I don't see the point in this honestly.

The only variables I am really testing right now are LVL of target cured, and the MAX HP of that target. I don't see what things like MND have anything to do with enmity. That would be so horribly and cryptically complicated. Nothing we've found so far would suggest stats like MND would have any affect on it.
(Anonymous) wrote:
Nov. 21st, 2007 04:53 am (UTC)
I simply meant the enmity generated could be calculated on the amount cured before adding the % bonus. The same could have applied to Cure V with say, MND gear.

Just throwing it out there is all. Someone could just run two identical tests with and without a light staff, if only to eliminate it as a possible variable.

If it affected it in some way, you would be getting some very confusing results from tests, making the results somewhat confusing.
[info]kanican wrote:
Nov. 21st, 2007 05:18 am (UTC)
I can verify that really quickly using LVL75 characters for you
(Anonymous) wrote:
Nov. 21st, 2007 04:56 am (UTC)
I need to proof read. That final paragraph made me lol.
[info]kanican wrote:
Nov. 21st, 2007 06:13 am (UTC)
lol I've done so much worse... I just get to edit =P
(Anonymous) wrote:
Nov. 21st, 2007 02:39 pm (UTC)
ask elipse about the test we did on if max HP effected the hate gain from cure. going from about 1k to 1.7k HP with the same amount hp return from cure1 i believe it showed the same enmity gain. so it's a function of lv only. (at least that's how i remember it, i didn't write anything down).
(Anonymous) wrote:
Nov. 21st, 2007 02:41 pm (UTC)
comment above by ciz^_^
(Anonymous) wrote:
Nov. 27th, 2007 06:13 pm (UTC)
Just a thought, maybe try with a bigger cure?

If Cure 1 healed 30 HP, 30/1000 = 0.03, 30/1700 = 0.0176. Does the game truncate here to get 0.0 for both? Not trying to say the game views this example as 0 hate obviously, just maybe it groups both scenarios into it's smallest hate bucket based on it's number system.
(Anonymous) wrote:
Jan. 16th, 2008 07:44 pm (UTC)
High/Super Jump
Hey, sorry for posting as Anonymous just too lazy to make an account.
Would it be possible that you could do a few tests on High Jump/Super Jump.

Altho it's commonly believed that super jump removes all hate or close to all hate there seem to be a lot of sources saying that it just puts you last on the enmity list.

As for High jump it has never been entirely clear how much hate it removes or even if it's a % or set amount. The only thing I can feel at least a little sure about saying is that the hate removal goes before the hate gain from the actual damage.

//Darkani of Cerberus
(Anonymous) wrote:
May. 29th, 2008 10:53 am (UTC)
Distance
Hey, can you show me how far the distance 25?
I am a PS player, I can't use plugin.
( 31 comments — Leave a comment )