~First And Final Line Of Defense V2.0~

I think it is the kind of thing to go on the OP Page of a guide. At least logically speaking, as well as roll in the pertinent information to said guides with links to the specific post/posts.

Hello plds. I wanted to ask how that new sword from lillith looks for your job. I am not sure if it compares to burt. But was curious if that status resist might make it worth equipping sometimes over burt. I usually tank on my run so I do not know what you plds look for but I do remember that other than fealty pld have hard time eith status resists. So was curious if this sword helps to fix that somehow.

hello,
since for /blu u have to use SIRD.
you need SIRD/Enimty/DT(PDT)/HP
what is the best set for that...

So the Knight's Bead Necklace +2 looks good on paper, but I have Loricate +1 and a moonlight necklace. I find that the 1 better DT doesn't help me free up any other gear slots.

So, people with PLD experience beyond me, am I missing a specific context where this neck is the most desirable option? Is it niche or a core need?

Enmity and Turtle Sets currently:
ItemSet 368938

ItemSet 359750

neck gives a nice HP bonus in addition to the dt aug. Wouldn’t say it’s needed but can’t go wrong with extra hp. I use the +1 myself. looking at turtle set you have room to free up/change your gear around as your over the pdt by 7 ( set is a total of 57) and that’s not including what sword your using. Looking at your set there’s a lot of changes you can make to still cap pdt and raise your HP.

It would not replace your moonlight necklace in your enmity set as it gives you both enmity and SIRD.

I posted the wrong set before....sigh. I hadn't apparently updated this to FFXIAH.


Updated the Above Set with this. I use a fully augmented Brilliance for 3 more DT while I am building my Burtgang.

The PLD neck was, honestly, quite a disappointment compared to other job's necks. That said it is a straight up improvement over Loricate+1. Extra HP is nice and having an extra +10 enmity in the set you'll be taking dmg in will reduce your CE loss. Also it does have immense DEF+ for a neck. Along with 20 vit, adding even more def. Issue there is that physical dmg is rarely all that much of an issue, but.. it is still better.

That said, with the price tag and either significant dynamis D RP farm time or additional cost to augment, I find it hard to justify the +2 neck for PLD.

btw, Jarl. Is that enmity set up to date? It's kinda weird. C path Souveran +1 is going to be better in nearly every slot there. Which you clearly have looking at that GS export of your tanking set.

While there are stronger options than Sourveran+1 in terms of pure enmity for every slot but the body, Souv tends to be all around good. Keeps HP and DT high during casts.

Yeah that is currently my set. I...completely missed that. I had been thinking from the perspective of trying to nab every bit of enmity I could but I have been seeing that it doesn't matter as much because modern DDs are just going to rip hate anyway it seems.

ItemSet 368978

Seems just all around better. Thanks Martel.

Tbh could drop the neck in general for something like dualism collar+1. Your not counting in the 10DT set bonus from Souv+1.
Current set And gear you have -42Dt + -10DT from set bonus and 5pdt from feet.

Has anyone played around with Moralltach or either Su4 Divergent swords? The HP boost and HP>DEF seems like it would open some slots, reduce HP flux, and help with some much needed MP management. Just curious if anyone has done actual testing and what it brings to the table.

I have a post comparing Moralltach's DEF effect to Burtgang. But it doesn't go much into the HP or refresh sides. I'm not sure if the augments were known when I posted that.

As for actual testing, I don't have one and I'm not going to buy one. So I will have to leave any actual testing to others. Although I don't think there's really anything to test at this point, with the augments and values known.

https://www.ffxiah.com/forum/topic/46016/first-and-final-line-of-defense-v20/54/#3409978

was the only post I remember putting the dyna sword in any positive light, letting you clear out some potency in other slots for hp or easier sird, so i guess its good for that?

I appreciate the math and spreadsheeting, but new gear has been introduced and was just curious. Looking across servers, it doesn't seem popular and understandably so with the buy in. I was just hoping someone with it or a Su4 could put its existence into perspective.

So, I've recently taken a long dive down the rabbit hole of JP FFXI blogs after FaeQueenCory pointed out some interesting posts to me.

This post will be about the Atonement dmg formula as tested by the JP blog leaguemili.com.

Direct link to post: https://leaguemili.com/blog-entry-34.html
DISCLAIMER: I cannot read JP. I'm going off the google translate for the posts. If you can read JP, Please, double check my conclusions. Or better yet, provide a full translation. XD

This was tested in 2015. I feel a bit behind the times now.Most of this is fairly straight forward. CE/6-1 + VE/6-1 = dmg. The complications come with the various caps.

While there's a global cap that scales with player level and then mainhand weapon ilvl up til 1190 dmg at ilvl 119. There are apparently also individual caps for CE and VE. Which is what that mess on the right side of the formula is.

So you can only get up to 665 DMG via CE or VE alone. A combination of both is required to go any higher.

When VE is involved your results can vary +/-5 Atonement dmg even when using exactly the same timing. This seems like it could just be the result of lag, but I have some ideas about this I'll express in a later post.

There's also some possible rounding oddities. Results for CE only are sometimes off by 1~2. Most of these seem to be resolved by rounding up to the nearest integer after ve/6 and again after ce/6. And assuming a minimum VE of 1. Credit to FaeQueenCory for those solutions. I just test things, the math makes my head hurt.

Now. We all know that Atonement sucks as a WS. So why is knowing how it calculates its dmg important? Because it makes an amazing enmity testing tool. This seems to have been the testers original intention. To figure atonement out so it can be used to test other things.

Prior test methods required multiple characters and carefully controlled situations while you used actions with known values to slowly even out enmity with the character that tested the unknown action. This is a huge pain in the ***.

With atonement you get an instant snapshot of your total enmity at the time of the WS. The need to completely control the situation is nearly eliminated, and the results are quite accurate, give or take a bit of VE.

This testing method has already led to some interesting discoveries on the JP side. I plan to get into some of these in some future posts but needed to lay some ground work with this one first.

I was going to wait a bit to post the rest of these, but well, not much point in sitting on them when they're ready, even if it means double posting. Still going to keep each in it's own post for ease of linking as a reference.

Sentinel... Does not work quite how we thought.

It's long been understood that sentinel generates 1800 VE, and its own +100 enmity effect does not apply to that 1800. Turns out that this is not true.

It's just that Sentinels base VE is 900. Not 1800. And its own enmity buff does apply to itself. So when you have +100 enmity, 900*2=1800.

But since both of these models yield identical end results how the hell do you tell the difference?

Sentinel Job points. It is known, though not terribly well documented that sentinel JP give an enmity bonus that is separate from gear enmity and is multiplicative with gear. Such that base*gear*sentinelJP% is how you calculate the enmity.

During my earlier atonement testing I was getting higher Atonement values than I should have been. After some control tests where everything was normal without sentinel, FaeQueenCory pointed out that my numbers were about 20% worth of enmity higher, and mentioned Sentinel JP.

This seemed odd since we knew that sentinel's enmity+ didn't apply to it's own VE, so why would it's JP bonus?

So. Testing!

I calculated some possibilities.

If the base is 900, then with +200 in gear enmity(+111 in actual gear, and +100 from sentinel) and 20/20 sentinel JP, we should see a 528 Atonement +/-5 dmg. If the base is 1800, then the predicted dmg would be over 1k, but would get capped to 665 by the VE dmg cap on atonement.

Test procedure was to have my alt pull a mob, then tag the alt with protect(0CE) and wit a few seconds to bleed away the small VE form the spell.
Then hit a macro that would sentinel, wait 1, then atonement.

And so.

We have a WS that's within +/- 5 of the expected Atonement dmg for 900VE base.

TLDR: Sentinel is 900 VE base, but its own enmity+100 applies to it, so you get 1800 VE on use.

Thoughts on VE decay mechanics.

In some of the prior Atonement testing I did I'd noticed something odd.

It makes sense that Atonement dmg would vary as VE decreases, and due to lag even with controlled waits dmg can vary.

But shouldn't the dmg be changing in amounts equal to the -60 per second tick? Some of the dmg values I've seen shouldn't be possible if VE is decreasing only in increments of 60.

Examples using sentinel.
The VE from using Sentinel with +/-0 enmity gear and 20 JP should be, 900*2*1.2=2160
The below chart shows the enmity and resulting dmg of atonement at successive 1 second delays after sentinel use. The 0 second delay is of course impossible, but I wanted show the original pre-decay enmity as well.

So, we have 359, 349, 339, and 329 dmg Atonements predicted. They each increment dmg by 10, since the dmg formula uses VE/6 and we lose 60 VE/second.

So right after using Sentinel under the stated conditions we should only be able to see these variants(excluding the 0 delay entry) in dmg up to 3 seconds out.

However this is not what actually happens. Any of these dmg values Can occur. But they are by no means the only ones that can occur.

For example I have a test Atonement sample at 348 dmg, which doesn't match any of the expected values. 348 dmg would require a VE value of 2089~2094. And while the VE decay should pass over these values as it decays, it should never land in that range.

My thought here is that VE doesn't 'tick' a single time once a second but de-increments almost continuously at a rate that matches 60/second.

Say for example, 1 VE decays per 1 weapon delay. Since 60 delay = 1 second losing VE at 1 per delay would also be 60 each second. And this is a system that the game already uses.

This more incremental decay system would explain how I can get Atonement dmg values that fall in between the expected -60/second 'ticks'.

It also explains the +/-5 variance in atonement dmg. Every 1/6th of a second you'd be seeing visible changes in dmg. So you get an expected value 1/6th of the time, and off numbers the other 5/6ths. +/-5. lol.

So, another revelation via JP blog. This one is also from leaguemili.com.
Again. DISCLAIMER: I cannot read JP. I'm going off the google translate for the posts. If you can read JP, Please, double check my conclusions. Or better yet, provide a full translation.

So... About cure enmity...

According to the testing done in this post the enmity generated by cures is divided by the number of mobs on the caster's hate list. Note: It seems that static enmity cures like cure V are an exception to this.

/blank stare

RIP cure enmity.

I've reproduced the tests myself, although I only tried a single target test and then one with 4 mobs.

Test procedure. Had my alt pull 1 mob. Tagged alt with protect(0 CE) then cast a cureIV on myself for 1001 with enmity+96. Waited for VE to decay to reduce variance. calculated it at about 46 seconds wait time, but I gave it a minute just in case.

64 dmg Atonement. The CE from that cure is already not blowing me away.

Killed mob and waited for repop. Repeated test with same procedures except alt pulled a group of 4 mobs with diaga.


.... 16 dmg.
64/4=16. This seems.. very, very clear. And very bad for cures as an enmity source when there are multiple mobs around.

I'm curious as to whether there is a cap to this, or if it will just divide by however many mobs you have. Well, up until you hit whatever limit there is. I have noticed some weird hijinks with enmity when you pull too many mobs and some of them seem to ignore enmity entirely. Or maybe I was trying to use cures to pull hate at the time... and it did eff all cause it was divided by 30 mobs... /facepalm.

What kind of damage were you taking in the time you're waiting to bleed off VE, and it has been a theory of mine that the hate mechanisms of cures also isnt what we thought.
Edit:
Cure produced enmity seems to generate different depending on who is casting and who is targeted along with combined enmity totals, leading to believe the generation mechanisms are multiplicative and not just a base value on amount cured. Ergo, the more hate you have the more is produced.

He takes 0 damage, it's a low level mob and he's got phalanx/stoneskin on.

just for clarification, does this apply to only cures or is it for any hate generating spell? I.E. would reprisals VE get split among mobs, cant think of any other pld spells that have any decent chunk of ve/ce on them

It's just cures. And as I noted it doesn't apply to static enmity cures like cure V.

Cure enmity is based on the target's level and the amount cured. Then modified by enmity+/-.

If you have some testing you'd like to present regarding your cure enmity hypothesis, by all means have at. But I currently have no reason to believe that cure enmity varies with your current CE/VE.

That's unfortunate then, they really dont like pld do they

Feels that way some days. PLD update, {When?}!

I have no math, my play time is limited and what I have is spent playing, not testing. The hypothesis is based on experience and expectations vs what actually happens. It should also be taken into consideration that what you believe is how something works may not be correct, or you're so conditioned to think one way, it's hard to see another. Libra is a viable tool to read enmity, but it never seems to get used in testing I've seen. I plan on doing so since we like numbers to back observations. Back to topic, in situations where I've lost hate to a damage spike, curing the DD that pulled hate vs Flash+self cure for same or greater (47% cure potency with 70ish enmity on others VS 20%Pot+30%Cure Recieved with +90ish enmity) will get attention back. According to the information on enmity generation from cures, this shouldn't happen. I'm curing myself for a same or greater amount with a higher enmity multiplier so the only answer is, the math that has been used is wrong. There's also evidence that taking 0 damage counts as taking damage thus contributes to CE bleed off. I was half expecting your tests to have your alt put the target to sleep to remove that possibility but again, since we presume that a hit landing for 0 bleeds no enmity, this could throw off the test.
We really only have 2 known methods of checking enmity and they're terrible at best. Libra seems to take both and presents an average of the 2 or maybe ignores volatile enmity altogether. We dont know.
I think pigeonholing ourselves into a mindset or way of thinking isnt going to get us where we want to be. I've been trying new things to Make Paladin Tank Again, and by thinking outside the box, I feel I have been successful.

Well, firstly some recommended reading. https://kanican.livejournal.com/13235.html. This is a link to Kaeko's blog. The FFIX player who did much of the original enmity testing that brought us to our current understanding of the system, and finally quantified the enmity values for various actions.

It's a fascinating read, and can show you a lot about how to test things via the scientific method.

One of the first things tested in the first post is whether or not being hit for 0 or missed reduces the struck player's enmity.

Regarding your anecdote, there's simply insufficient information here to draw any conclusions. We have a DD with an unspecified amount of CE/VE, from whom hate was pulled via one method but not the other? I honestly can't tell if you're saying one worked or both from this. In either case, without knowing the CE/VE of the DD in both cases, there's no telling if they were 10~ CE ahead of you when they had hate or 1000 ahead. Thus you can't tell anything about the enmity from the cures from this story.

If you want to propose a change to the current enmity model, please provide sufficient evidence to demonstrate that your new model is correct.

It's hard to quantify something that's damn near impossible to measure and changes so quickly. There's no proposed change, but I think we can all agree that some things dont work how we thought. Your multiple target cure/enmity test lends credence to this as well as Rigamiri's tests/conclusions with Atonement and Sentinel. Atonement is probably a bad example since their tests was to solve a long held mystery, but Sentinel and your cure tests are pretty good examples. Lemme state this another way. You said from the single target test that the CE generated was disappointing and RIP cure enmity. Was this because it was less than expected, cuz it seems that way.
We're all on the same team, we desire the same goal, we like the same job. Let's work together to figure it out and show those GS wielding RDMs who the real tank is. We have Superior tools at our disposal to get punched in the face and laugh it off, we just need to figure out how best to use them.

Dont wish for it too hard, might end up with nin update 2.0 or whats looking like the bst update and they just say add more enmity to sentinel and call it a day

This may be a silly question but would it be possible to code an enmity Scoreboard? I know having an alliance wide hate list would be damn near impossible but like a personal measure of enmity would surely be possible, no?

It would sort of be possible to do one for enmity generation for yourself only using packets. It would involve capturing your equipped gear on cast completion (packet 4) and job ability use (not sure what action packet that is). You'd also have to grab merits/jp that modify enmity.

It would be pretty hard to do the part for enmity regression from damage and VE degen.

The cure enmity test actually did initially have lower CE gain that I had anticipated. At least until I realized that being on PLD/RDM, I had Tranquil Heart from /RDM. After recalculating the expected CE the numbers fit pretty well when accounting for tranquil heart's cure enmity- term.

But the comment in my test post was more a general statement of discontent with the potency of modern cure enmity.

And yes. We've been wrong about things in the past and will be so again. But the way we learn the truth of it is through testing and evidence. Not anecdotal hearsay. The JP blogs posts presented excellent testing and evidence, which I then retested to verify, and I am now convinced of the veracity of those claims.

So. If you think that the current cure enmity model is incorrect, how do you think it is wrong, and what test could be performed that would demonstrate that this is the case?