New Heal Formula

Hi guys, I thought I'd pop in and share a few data points from my HP cleric. Only level 35, and only has level 14 heal (ie 140%), but I think the data points are worthwhile given the abnormal nature of the source.

Basically, I pumped pure HP from 1~30. At 30, I had it confirmed for myself that HP meant NOTHING, but I kept going with the project with a plan to leach experience using snail throwing. Well, long story short, that dream was killed when they made snail shell throwing require some sort of accuracy.

Using my Maple Staff, I had:

max hit: 42
min hit: 2
magic: 123
int: 66+2
luk: 4

Using a 0 magic weapon:

max hit: 19
min hit: 1
magic: 73
int: 66+2
luk: 4

Also notice: I still miss quite a bit due to my low magic accuracy (see: low int stat), even though I tested on Zombie Mushrooms and that's the lowest undead thing. This may account for an outlier min hit, such as the min of 2 when I was using my full magic gear, because if I recall, along with missing, sometimes hits will come off as UNBEARABLY weak.

So, guys, does that break this formula too? A quick test for me tells me that my max/min with Maple Staff should be 16~59, and it should be 9~35 with my second set of stats posted. Am I wrong?

EDIT: By the way, don't be heartbroken if this proves your formula isn't 100% accurate. None of the formulae are. Being very accurate for normal builds is good enough, really.

Only the max dmg was observed, because I'm like that, a powergirl.

(changed ign to LagIess, if anyone wonders)

MAX = (INT * 4.8 + LUK * 4) * Magic / 999 * Heal %
Max = (412 * 4.8 + 4*4) * 543 / 999 * 3 = 3250

Removing damage from mdef: 3145.

Doesn't make correct damage. A more correct way would be to actually include the fact that there's a mastery percent (30%, which may be multiplied with 0.9 or 0.8 from all I know) affecting the minimum damage, like all other damage formulas:


MAX = [(INT * K + LUK ) * Magic /1000 * Heal % - mdef*0.5]*heal %
MIN = [(INT * K *0.3 * 0.9 + LUK ) * Magic /1000 * Heal % - mdef*0.5]*heal %

That would make more sense, since all other formulas use that way of calculating.

Not all others. L7, Venom, Phoenix/Frostprey, to name a few from my other thread.
And 1.2/4.8 is the best fit I've found so far, given all the data points.

I don't usually consider not reaching the supposed max damage enough evidence to disprove a formula, unless it's by a wide margin like over 10% of the range (300 damage off, in your case). However, if my equation predicted, say, 3081 for your max damage, that would be valid counterevidence right there. Since a 3083 would be a physically impossible number if the formula was true.

In this case, 3083 is definitely possible, even if it's somewhat unlikely that you didn't hit anything higher. 4.8 is already a pretty tight fit for some of the other data points.

Okay, I think something is strange.

Level: 50
INT: 262
LUK: 6
Magic: 350

Enemy level: 57
Enemy MDEF: 180

3 targets, observed damage: 138-840

I'm relatively certain the low end is close to actual minimum damage - but I'm not so sure on the high end.

It seems like I'm hitting way fewer numbers outside a normal distribution (most of my hits end up 250-600 or so) than I would with 1 or 2 targets. Perhaps 1 out of every 20-25 hits exceeded 800. It could just be my luck for today, but I find that to be a little strange.

Hi guys, I thought I'd pop in and share a few data points from my HP cleric. Only level 35, and only has level 14 heal (ie 140%), but I think the data points are worthwhile given the abnormal nature of the source.

Basically, I pumped pure HP from 1~30. At 30, I had it confirmed for myself that HP meant NOTHING, but I kept going with the project with a plan to leach experience using snail throwing. Well, long story short, that dream was killed when they made snail shell throwing require some sort of accuracy.

Using my Maple Staff, I had:

max hit: 42
min hit: 2
magic: 123
int: 66+2
luk: 4

Using a 0 magic weapon:

max hit: 19
min hit: 1
magic: 73
int: 66+2
luk: 4

Also notice: I still miss quite a bit due to my low magic accuracy (see: low int stat), even though I tested on Zombie Mushrooms and that's the lowest undead thing. This may account for an outlier min hit, such as the min of 2 when I was using my full magic gear, because if I recall, along with missing, sometimes hits will come off as UNBEARABLY weak.

So, guys, does that break this formula too? A quick test for me tells me that my max/min with Maple Staff should be 16~59, and it should be 9~35 with my second set of stats posted. Am I wrong?

EDIT: By the way, don't be heartbroken if this proves your formula isn't 100% accurate. None of the formulae are. Being very accurate for normal builds is good enough, really.

Hahaha~ just inputted your data. When you account for the Zombie Mushrooms' defense, the formula actually fits quite well.
Observed | Projected
1-19 | 1-20
2-42 | 1-44

On the contrary, rather than proving it inaccurate, this is very interesting data. It helps confirm that Heal % is multiplied over everything. And it's also helpful as far as order of operations go.

Lemme explain. You didn't see many 1's, right?
And you probably saw a lot more misses than you would've seen with magic claw, right?

Maple does something I call "1-correction" to most damage after defense is applied and whatnot. Basically, it checks whether the damage is within the range of 1-99999, and adjusts it if not. So if a high amount of defense makes your damage go into the negatives, 1-correction turns it into a 1. Hence why most attacks do 1 damage when they can't penetrate defense.

With Heal, however, I believe there either is no 1-correction or it happens before subtracting defense. So after defense, your damage is occasionally negative. This makes no sense to the game, so it turns it into a miss.

It happens with a couple other skills, notably at low levels; I'm not entirely sure whether it's the same case or not.

And Bui, I don't find that strange assuming the damage is a uniform distribution, you should only see 800+ damage 1 in 17 times anyway. But I haven't really looked at it yet, so we'll see.


Anyway, replaced the data with an image. Cleaner.

Well isn't that interesting.

I'd say this is pretty much it then. How did you get this again?

Sap, think of Iron arrow and how it misses on the second bonefish when it had wdef up.

Heal is the same logic, a skill with a multiplyer that brings it lower.

I'm curious about the multi-target numbers.

When I apply the formula like this they seem to fit neatly:

M = (4.8*int+4*luk)*matk/999
Max = (M*t - 0.4*mdef)*%heal
t = multiplier for number of targets,
1->1
2->0.8
3->0.7
4->0.66
5->0.6

There's not really enough data to compare this straight across on anything except the 1 character with 201 int and I don't have an excel spreadsheet set up with the formula to quickly compare this to your other multi-target monsters. Was wondering how well this works out for the other values.

0.4? That's a new figure. I've so far been assuming M*t*%heal - 0.5*mdef to project the maximum damage, and 0.6*mdef for the minimum.

I'll have to look into it later; I'm having my wisdom teeth out tomorrow, and I have to sleep early.

I got that from the one time I saw actual numbers posted as I couldn't figure out how you were doing it, it could be lower but that's how it is for most skills (applied before %heal, and times around 0.4-0.5)

Also as indicated it works out ok for the 1 character that's got all 5 target amounts max damage noted, as the numbers in previous post result in within 3 damage on every amount.

This is what I have, and it's 100% accurate so far:
MAX = (INT * 4.8 + LUK * 4) * Magic / 999 * Heal %
MIN = (INT * 1.2 + LUK * 4) * Magic / 999 * Heal %

So INT carries more weight than LUK? Interesting to know that...
Hmm... do you think that DEX is a factor in this as well? Because the 'hands' stat is a combination of DEX, INT, and LUK, which suggests that all three are related in some way. It may account for the slight discrepancies in your data and predictions.

So INT carries more weight than LUK? Interesting to know that...
Hmm... do you think that DEX is a factor in this as well? Because the 'hands' stat is a combination of DEX, INT, and LUK, which suggests that all three are related in some way. It may account for the slight discrepancies in your data and predictions.

Unlikely, since all holy magicians will have ~10 or less dex, usually 4.

Hands likely does nothing more than whatever it does with the Maker skill. It's even been confirmed that it has nothing to do with magic accuracy, which was a popular theory for a short while.

Unlikely, since all holy magicians will have ~10 or less dex, usually 4.

Many people don't start perfect and until the MTS came out, can't afford to reset that stat to four. Even with MTS out, some people refuse to reset it, thinking that it takes too long to merchant.
Then some have zhelms and some don't. The extra 13~17 DEX from that zhelm might add damage. The horntail pendant, which adds yet another 21~23 DEX may make it worthwhile investigating.
It's like Dits and Shadowers getting more Atk from a zhelm because of the STR on the helmet.

I got that from the one time I saw actual numbers posted as I couldn't figure out how you were doing it, it could be lower but that's how it is for most skills (applied before %heal, and times around 0.4-0.5)

Also as indicated it works out ok for the 1 character that's got all 5 target amounts max damage noted, as the numbers in previous post result in within 3 damage on every amount.

I've investigated weapon defense before. It's pretty much confirmed to be 0.6 off the max and 0.5 off the min before skill percentages, when level is not a factor.

I haven't tested out magic defense much, though my limited samples seem to support the 0.6/0.5. Then again, that was more assumption than actual conclusion, since I was working with a mere 10 defense.

Sap, is that magic calc on H-S correct?

If so I'll test out the formulur for you on some crazy stuff.

I'm not sure. Go to my thread; the 'fraction version' seems to be the most accurate.

Many people don't start perfect and until the MTS came out, can't afford to reset that stat to four. Even with MTS out, some people refuse to reset it, thinking that it takes too long to merchant.
Then some have zhelms and some don't. The extra 13~17 DEX from that zhelm might add damage. The horntail pendant, which adds yet another 21~23 DEX may make it worthwhile investigating.
It's like Dits and Shadowers getting more Atk from a zhelm because of the STR on the helmet.

Nothing you said matters. The point still stands that there is very little difference in dex between any Magicians, and it is illogical to think dex might do something like that.

Next Tespia, if one of my data slaves gets in, we'll test with 4/999/4/4 and 4/4/4/999 to make sure.

Are those multipliers agreed on? If so, I'll slap this thing in for v1.00, which you should expect around early August. Possibly later because I have to redraw things. (because someone nags about colours not looking well together. I forget who.)

I'm curious about the multi-target numbers.

When I apply the formula like this they seem to fit neatly:

M = (4.8*int+4*luk)*matk/999
Max = (M*t - 0.4*mdef)*%heal
t = multiplier for number of targets,
1->1
2->0.8
3->0.7
4->0.66
5->0.6

There's not really enough data to compare this straight across on anything except the 1 character with 201 int and I don't have an excel spreadsheet set up with the formula to quickly compare this to your other multi-target monsters. Was wondering how well this works out for the other values.

I tried your way of calculating it.
It's very close for many of the samples, but it just doesn't fit that evenly. No matter what I put for the defense multiplier, there are always some discrepancies if I subtract it before multiplying Heal's %. Whereas they all fit pretty evenly, if not 100% accurately, if I subtract afterward.

That one HP cleric's sample makes this clear.

@ Joe, the multipliers are still a bit of a mystery at this point. I'll play around some more to see what I can get, but we don't have definite answers yet. Other than the approximations in the main thread.


Oh and Bui, if you haven't leveled yet, I've found the problem.
The problem wasn't with the 967. It was with the 167. I see you screened a 176, but is the 167 for sure?
I can't put any sort of multiplier for 2 targets without infringing on that 167.

Anyway, ignoring the 167 (though I hope I don't have to ignore it), I've been able to place bounds on the target multipliers:
2: 0.7885 - 0.8012
3: 0.6905 - 0.6944
4: 0.6210 - 0.6406
5: 0.5699 - 0.5890
Need more data. With enough samples, we'll be able to pin them within two or three decimal places, and then it'll be pretty apparent what they are.
So far, though, it's quite an enigma. It's definitely not an exponential of any sort, nor is it a reciprocal of something.