Cure Formula Changes

[Motenten]

Would still appreciate if you could do the mnd and vit tests. Would probably be easiest on Cure III. Anything above 231 skill (260 HP healed) takes 5 skill per HP. Do enough to find out exactly how much mnd/vit is needed to bump the amount cured up a point for at least 3 cure values.

I've got exact scaling values for all of tier 3, and most of tier 2 (Cure 1 only has data from the ffxiah post, and Cure 4 is a little peculiar), and amount cured for caps 3 and 2.

[Motenten]

Also, can you recheck Cure III @345 skill. Is it 283, or actually 282?

[Aleste]

I've coloured it in and stuck a note on it. I'll test it when I'm done with voidwatch tonight. doing it atm.



You're correct, its 282. Adjusted value on the spreadsheet.

Since it's going to be pretty much impossible to get the remaining gaps in my graph, I'll try doing some scaling by MND and VIT tomorrow for you mot.

[Economizer]

Apparently we're closing in on an answer. Or more accurately players on Fenrir (Seriously, Furen was on Fenrir too, what is is about Fenrir and discovering the cure numbers? The forums may never know.).

I tried to dig up more threads but this is all I found:

FFXIAH WHM Forums
FFXIAH General Forums

Kinda sad considering how important these numbers are.

Anyways if we're doing testing made to order can you test this too:

• Cure Potency and Healing Magic
  WHM 99 Capped Skill, with and without cure potency (24% should suffice however 50% would work better if possible).
  WHM 99 Capped Skill + Merits (although more would be acceptable), with and without cure potency.
  I'm thinking the results of this test will be very predictable but they'd completely destroy any theories for a multiplier, helping to alleviate any lingering doubts.
• Cure V/VI
  Just need a handful of numbers on this one, mostly to ensure that nothing too scary changed, especially with the threatening brew numbers. MND/VIT being the same would be a plus but not a requirement. I'm mostly interested in the natural cap number personally but extra data is always useful.
  Numbers to test at for skill: 500+, ~450, 424, 301, 232 (Cure V only).

[Aleste]

Motenten's and Economizers testing has been knocked out today.

I think we officially debunked econ's theory today... kinda sad really, I had hoped it was the lesser of 2 evils.


In other news. Take that cure5~ +50 MND adding only 7 hp when naked...

[Economizer]

Aleste's numbers for Cure V were at first comforting (they were a slight buff) around Healing Magic 424, MND 97, VIT 86. Then he added some MND. It looked normal. Then he added some more... and we noticed a trend... it wasn't improving as fast as it normally would.

The picture in question shows 97+50 and actually shows one more point of HP then the current formula but if you continue to increase the MND it gets worse. At a reasonable and perfectly achievable outside Abyssea amount of MND, the new formula will perform worse then the old one for the same value of Healing Magic. It is disturbing to say the least.

Considering that Cure V is already less MP efficient then Cure IV by a wide margin with the new formula I don't think a nerf of the effect of MND was neccisary.

I have more thoughts on this matter but I'm not sure I want to write them all out right now. I'm hoping the implications of this are clear enough to get some discussion going.

[Motenten]

Well, I'd hoped you could stretch the mnd tests out to at least +20 (and vit needs a wider span as well), but working with what's available..

And actually, if you could redo it with a skill value that's divisible by 10 (eg: add 3 merits to take it to 430 total skill, or even full merits to take it to 440 (divisible by 20)), that would eliminate one element of uncertainty for now (whether any factors are added together before flooring).


Relative values:

Cure I and II scale at 20 skill per point of HP, so the range isn't nearly large enough to show results for mnd and vit. Ignoring them.

Cure III scales at 5 skill per HP.
Cure III scales at 9 mnd per HP. Not enough points to check for rounding.
Cure III doesn't have enough range to check for vit scale.

Cure IV scales at 5 skill per 2 HP.
Cure IV scales at 9 mnd per 2 HP. Would like a few more points of mnd checked to be certain of this.
Cure IV scales at 12 vit per 1 HP. (so either 24 or 25 vit per 2 HP)

Cure V doesn't have any skill scaling data.
Cure V scales at 10 mnd per 2 HP.
Cure V scales at 12 vit per 1 HP. (so either 24 or 25 vit per 2 HP)

Slight discrepancy between IV and V that may be related to overall scaling. Since we don't have skill-based scaling for V, can't say for sure.

Cure VI doesn't have any skill scaling data.
Cure VI scales at 10 mnd per 2 HP.
Cure VI scales at 12 vit per 1 HP. (so either 24 or 25 vit per 2 HP)

Cure VI appears to scale at the same rate as Cure V.


Overall power formula appears to be approximately:

Power = skill + mnd*5/9 + vit/5

Or, alternatively you can think of it as:
45 skill = 45 power
45 mnd = 25 power
45 vit = 9 power

So let's see if that fits with prior test results..

96 mnd + 85 vit = 53.33 + 17 = 70 power

Cure II, cap 2 = 130 cured @ 331 skill
70 + 331 = 401

Cure III, cap 2 = 260 cured @ ~231 skill
70 + 231 = 301

Cure IV, cap 2 = 520 cured @ 331 skill
70 + 331 = 401

Estimated cap 1 for Cure IV
405 HP @141 skill, scaling rate of 2 skill/HP
390 HP would be reached at 111 skill
70 + 111 = 181 power


Would guess that I'm at +1 over actual power. Not sure which element is off.


FFXIAH data:
92 mnd, 83 vit, 67 power from stats

Would then be 3 power lower than Aleste's data, and expect to reach same amount cured as Aleste's with 3 lower skill.

Skill   Cure3   Cure4
334     280     520
A331    280     520
368     286     533
A365    286     533
404     294     548
A401    294     548
424     298     556
A421    n/a     n/a
464     306     572
A461    n/a     n/a
489     311     582
A486    311     582
504     314     588
A501    314     588

Aleste's data with 3 lower skill exactly matched all instances of the FFXIAH data, aside from the skill values that Aleste doesn't have data for.

[Motenten]

Making some estimates on power. Ignoring the +1 anomoly.

Cure
Scaling for tier 3: 20
Cure at cap 2: 45
Power at cap 2: 200? (somewhere between 198 and 218)
Max cure: 65
Power for 2 > 3: (65-45) * 20 = 400
Power at cap 3: 600?

Cure II
Scaling for tier 3: 20
Cure at cap 2: 130
Power at cap 2: 400
Max cure: 145
Power for 2 > 3: (145-130) * 20 = 300
Power at cap 3: 700

Cure III
Scaling for tier 3: 5
Cure at cap 2: 260
Power at cap 2: 300
Max cure: 340
Power for 2 > 3: (340-260) * 5 = 400
Power at cap 3: 700

Cure IV
Scaling for tier 3: 2.5
Cure at cap 2: 520
Power at cap 2: 400
Max cure: 640
Power for 2 > 3: (640-520) * 2.5 = 300
Power at cap 3: 700



Cures 2, 3 and 4 seem to cap at 700 power. Cure 1 caps at 600 power. No data to make a guess at cures 5 and 6 yet.

This is different from the current formula, where all cures reach cap at the same power (1240).


Old cap reached at: 400 skill, 250 mnd, 250 vit.
~~ (250 mnd + 400/5 skill) * 3 + 250 vit
== (250+80) * 3 + 250
== 990 + 250
== 1240

New cap reached at approximately: 496 skill, 270 mnd, 270 vit.
~~ 496 skill + 270 mnd * 5 / 9 + 270 vit / 5
== 496 + 1350/9 + 270/5
== 496 + 150 + 54
== 700

or

598 skill, 135 mnd, 135 vit.
~~ 598 skill + 135 mnd * 5 / 9 + 135 vit / 5
== 598 + 675/9 + 135/5
== 598 + 75 + 27
== 700



Aleste's graphs of Cure IV indicated another inflection at about what would be 300 power. So its progression would be like:

Min power: ??
Cap 1 power: 300
Cap 2 power: 400
Cap 3 power: 700

With scales of:
Scale 1: 2.0
Scale 2: 10/7 (1.429) (originally estimated, corrected to exact value)
Scale 3: 2.5

That means the amount cured at cap 1 should be: 520 - ((400-300) / 1.42) = 450

That matches the test data from Aleste. So...

Amount cured at:
Min: 270
Cap 1: 450
Cap 2: 520
Cap 3: 640

However that implies that the power required to reach cap 1 is at least (450 - 270)*2 = 360, when we're only at 300 power.

The above Scale 1, 2 and 3 are all verifiable as 2.0, 10/7 and 2.5, respectively. Therefore there has to be a fourth tier at work.


Given that the difference between Min and the above Cap 1 (450-270 = 180) is much larger than the difference between the other caps, we have a fair bit of room to work with in making a preliminary estimate. In the current model, the scaling rate is 1.0 for the very first tier for most spells, so I'll find a value set that matches that.

Revised:

Min cured: 270 HP
Min power: 40
Tier 1 scale: 1.0
Cap 1 cured: 370 HP
Power to reach cap 1: 100
Total power at cap 1: 140
Tier 2 scale: 2.0
Cap 2 cured: 450 HP
Power to reach cap 2: 160
Total power at cap 2: 300
Tier 3 scale: 10/7
Cap 3 cured: 520 HP
Power to reach cap 3: 100
Total power at cap 3: 400
Tier 4 scale: 2.5
Cap 4 cured: 640 HP
Power to reach cap 3: 300
Total power at cap 3: 700

I chose a min power of 40 to match a rough scale of min power of 10/20/30/40 for cures 1/2/3/4, similar to the current model. Obviously all of this needs testing to find the actual values, but the progression seems pretty clean.

What we see is an initial fast gain (1.0 scale) that slows down (2.0 scale) as it approaches the cure capability of someone subbing a mage job (thus ~150 skill and maybe around 250 power). It then speeds up again after power passes the 300 mark (10/7 scale), followed by another slowdown (2.5 scale) for the highest levels.


Testing the lowest end will require the no-subjob trick and the no-skill trick.

For the no-subjob trick, described here.
For the no-skill trick, as I understand it, you need to get the moogle to set your job level (41 whm for cure 4), and then choose one of the teleport options without exiting the menu. This should prevent auto-capping your skills.

Together this should put your total power in the 30-ish range. From there you can do minor tweaks to merits and such until your cure amounts start going above baseline.



As an aside....

Currently I cure in.. 88+40 mnd, 80+3 vit (from what I remember offhand). Light Arts puts me at 404 skill, plus 35 skill in gear. Total power would be approximately 526.

Assuming that's roughly in the neighborhood of a normal average-to-good build, that would put typical cures at about:
Cure 1: 60, 90 with 50% potency
Cure 2: 135, 200 with 50% potency
Cure 3: 305, 460 with 50% potency
Cure 4: 570, 855 with 50% potency





Checking the same with Cure III, but assuming 4 tiers to start with:


Min power: 20?
Cap 1 power: 170?
Cap 2 power: 200?
Cap 3 power: 300
Cap 4 power: 700

With scales of:
Scale 1: 2.5?
Scale 2: 1.0?
Scale 3: 2.5
Scale 4: 5

Amount cured at:
Min: 130
Cap 1: 190?
Cap 2: 220?
Cap 3: 260
Cap 4: 340


Much more difficult to make good guesses here due to lack of low-end data.

[Vyvian]

Don't know how helpful it is, but you can also manipulate skill levels by changing to 1 job to get a certain skill, and then going to a Nomad Moogle (Selbina, Mhaura, Kazham, Rabao etc) and changing jobs with them. The limitation is if you lower a job level, it takes every job down with it, so this is good for getting your skill reduced to subjob levels without lowering the level of WHM.

So for the sake of example, if you wanted to test like 100 healing skill, change to a job with no healing skill, and sub PLD (C skill), then set your main job level to 70, which will cap your skills. C skill with a level 35 job is 100 skill exactly. Leave the mog house as that job, and get to a nomad moogle, change jobs to whm, and you will still have 100 healing skill (although your whm will still be level 70).

It may not be that much more helpful than what you're already doing, but it may provide additional data for specific skill points after large increases in MND or VIT (because you're using a job that is twice as high a level for the same skill).

[Aleste]

Well, I'd hoped you could stretch the mnd tests out to at least +20 (and vit needs a wider span as well), but working with what's available..

And actually, if you could redo it with a skill value that's divisible by 10 (eg: add 3 merits to take it to 430 total skill, or even full merits to take it to 440 (divisible by 20)), that would eliminate one element of uncertainty for now (whether any factors are added together before flooring).

I'll change the healing magic to 430 and retest a larger range for you tomorrow.

Testing the lowest end will require the no-subjob trick and the no-skill trick.

For the no-subjob trick, described here.
For the no-skill trick, as I understand it, you need to get the moogle to set your job level (41 whm for cure 4), and then choose one of the teleport options without exiting the menu. This should prevent auto-capping your skills.

Together this should put your total power in the 30-ish range. From there you can do minor tweaks to merits and such until your cure amounts start going above baseline.

Unfortunately this doesn't seem to work for me, upon changing primary job level, the dialogue with the GM moogle closes. Skills are auto capped whenever you use the change main job level function.

Stat adjusting stuff

Actually, that's pretty much what I've been doing to get the lower bounds values, the trouble is that it becomes a very complicated juggling game of stats. Adjusting merits, trying to find magian trial staves with appropriate stats, and hunting through claim slips for gear.

The main trouble I've been having is that I'm attempting to vary one stat, whilst maintaining the other 2 so we have less variables to worry about. Although, that said, if we became 100% sure of the formula, in theory we could just work with cure-power and graph any combination of job/level/stats/skill.

I'm on nightshift tonight, so I'm afraid I won't have time to do it today... however tomorrow I'll make a much larger sample size of MND and VIT scaling, which'll hopefully help confirm Motentens formula. I'll then change my main spreadsheet to work with cure-power rather than with skill, and then work on getting the lower bounds.

I'm actually surprised that the change to the formula isn't being more discussed... it has a huge knock-on effect across the other jobs.