Cure Formula Changes

[Aleste]

I've spent the majority of the day with econ collecting values and bouncing back and forth ideas and theories.

HP healed varied by healing magic skill


Old/new cure ratios

[Economizer]

Reported test server values of cures under a Primeval Brew compared to older values:

• Cure
  40->65
• Cure II
  105->145
• Cure III
  210->340
• Cure IV
  450->640
• Cure V
  810->780
• Cure VI
  1140->1010

[Fredjan]

While I may not be able to test stuff as well as others, I figure, may as well voice my opinion on the changes on the test server which'll most likely be live as well.

Reported test server values of cures under a Primeval Brew compared to older values:

• Cure
  40->65
• Cure II
  105->145
• Cure III
  210->340
• Cure IV
  450->640
• Cure V
  810->780
• Cure VI
  1140->1010

Or in other words, with 50% potency:

Cure I = ~97.5 HP
Cure II = ~217.5 HP
Cure III = ~510 HP
Cure IV = ~960 HP
Cure V = ~1170 HP
Cure VI = ~1515 HP

As Cure V+ got slightly reduced in potency, considering the major increase in the other 4 spells (I and IV by as much as 62-63%), I'm quite content to be honest. They needed the most attention by far. People weren't really capable of reaching the maximum amount on Cure V+ anyways and the difference will be pretty minimal on Cure V. I may think Cure VI may have been reduced a bit too much, but it's also become much less worthwhile to cast (as if it was amazing to begin with as it's still an abyssea only spell to me OR for an emergency cure).

Definitely gotta say -enmity setups just became that much more important with the stronger HP healed.

[SpankWustler]

I have to imagine that "emergency use only" is the idea that the Development Bros have in mind for Cure VI, since the amount was reduced alongside Cure V rather than increased. Well, assuming they have ideas and minds and stuff.

These changes seem like a big win for every job involved. Red Mage and Scholar will have a substantial Cure in Cure IV and even Cure III will be workable. White Mage will more-or-less have three flavors of Cure V after these changes go live. Traditional Vanilla (V), a quirky Dark Chocolate with Toffee Bits (IV), and Sugar-Free Low-Fat Gooseberry Licorice that melted and refroze (VI).

Everybody's MP efficiency will go way up, and White Mage's MP efficiency on lower cures will be huge. It's hard to care that Cure VI is apparently intended to be mostly superfluous when, well, Cure VI is going to be mostly superfluous.

[Motenten]

Current best guesses that I've made. General informational update on my work so far. I lack any solid information on mnd/vit scaling, so no guarantees there yet.

Power = Skill + Mnd/2 + Vit/4

Scale N: x == how much power (x) is required to raise amount healed by 1 (valid up to cap for N)

Cure
Min power: ?? (10 cured)
Scale 1 (to cap of ??): ??
Scale 2 (to cap of 45): 12.5 ?
Scale 3 (to cap of 65): 20
Max power (guess): ?? (637 if scale 1 is 2.5 to cap of 25)

Cure II
Min power: ?? (60 cured)
Scale 1 (to cap of 100?): ?? (5?)
Scale 2 (to cap of 130): 10
Scale 3 (to cap of 145): 20
Max power (guess): 640 + base

Cure III
Min power: ?? (130 cured)
Scale 1 (to cap of 150): 2.5
Scale 2 (to cap of 250): 4/3
Scale 3 (to cap of 340): 5
Max power (guess): 633 + min

Cure IV
Min power: 60? (270 cured)
Scale 1 (to cap of 325): 1
Scale 2 (to cap of 480): 7/6
Scale 3 (to cap of 640): 2.5
Max power (guess): 636 + min

No data on Cure V/VI

I know that some of the values don't quite add up, but I'm missing data (especially at the low end) to solidify the results.

[Economizer]

I've been mostly trying to hold off my analysis on the Cure V/VI numbers for a brew until we see the lower numbers for them at realistic stats, although I certainly have an earful if the brew numbers are the norm for Cure V/VI.

However an interesting trend is that they were the same at 424 and at a higher skill number when they were tested.

Since I've been relying on essentially one (brilliant) tester to get these numbers it has been a pain to get more information on this and the added difficulties that are coming up for testing, particularly for getting Cure V/VI numbers.

I've been digging through various forums here and there and it seems like this is the authoritative and almost nearly only source, although I did see a handful of numbers from an FFXIAH thread. I grabbed those numbers and referenced them against Aleste's numbers and while it is a pretty small sample and I was unsure of the testing method that person used it gave an interesting (although frustratingly hard to read) chart as a result. (Also regretfully I forgot to plot in Daniel_Hatcher's numbers he gave with this until I started typing this, but I have a sneaking suspicion that this numbers are starting to look more and more consistent with something.) I'm not sure if my sudden urge to post charts is a result of latent progress or lack of sleep so I'll have to leave it at that.

As you can see the same tier cures, despite having slight different MND/VIT numbers practically hug each other to the point of being hard to tell apart. I'm not sure if this is a result of similar MND/VIT (92/83 from the FFXIAH poster and 96/85 from Aleste) or an actual trend.

I also dug through the JP forums and didn't find much in the way of testing but perhaps my lack of results was due to the reliance on an abusive machine translator.

I can't really seem to fit any sudden inspiration or humor into this post so I'll leave it at this for now.

[Aleste]

My thoughts for today:

Upon sleeping on it and thinking some more, it seems to me that the cure4 formula has at least 3 different gradients in it; initially observed yesterday and eyeballed with some rulers.

Here is a very terrible and quick mspaint-job to get the point across.


Mathematically we can approximate where the lines meet by calculating the gradient of both lines and figuring out the intersect point between them. This should give a good idea where to further investigate for testing.

NOTE: Working to 4 significant figures. It's an approximation after all.

We can do this by approximating the 2 lines using the standard straight line formula y=mx+b.
Gradient(m) of the 141-157 line is:
(413-405)/(157-141) = 0.5

We can then substitute in a few values to obtain our b value:
(405)=(0.5)(141)+b
b=405-(141*0.5)
b=334.5

Thusly Line A (141-157) can be approximated by:
y=0.5x+334.5

Now for the second line:
Gradient(m) of 264-299 is:
(497-473)/(299-264) = 0.6857

Substitute in some values:
497=0.6857(299)+b
b=291.96

Line B can be approximated by:
y=0.6857x+291.96

Aaaaand now its graph time.


These 2 lines intersect around the ~240 mark.

---------------

It would appear that the effect of healing magic on cure4 seems to scale differently (at least, lower bounds not tested) in the following areas:

ADDITIONAL EDIT:
<240
240-340
340-???
???-???

My aim for testing today is to physically test values around the 240 mark, above and below in the hopes of observing the precise point at which the gradients change.

If there is a gradient change at the 240 mark, we can then pick a few more scattered skill values between the 2 to confirm or deny extrapolating the line.
If there isn't a gradient change at the 240 mark, then we can assume there is at least 2 more points at which the graph changes.

Additional edit note: maybe it's just me, but the gradient of the highest tiers seems to be different than the 340-400 ish results. I'll do the gradient math in a sec to see if its just my eyes playing tricks on me.

Though, it would be handy if it was 0-140, 140-240, 240-340, 340-440, 440+

[Aleste]

Tested more values on the spreadsheet. Managed to get the 225-260 values, and a few gaps dotted here and there.

I'm starting to become incredibly inventive (read: stupid) trying to figure out how to get the 100-220 values...


Edit:// I've given up for today, all this stat jiggling is giving me a headache. I /think/ i can get some of the 120-140 values, but it's being a complete pain in the arse...

[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?