Meteor

[Yugl]

More tests
- Double recovery makes meteor hit for zero damage
- Brew gives well over 900 attribute, so I couldn't test the INT

[Yugl]

More Magic Tests (Can check the same link; see INT Modifier)

I used the most MAB I could with gear, but no atmas. The SSs show the amount of INT I used (Used Kirin, Clawed Butterfly, and Minikin as atmas). The goal is to test whether INT affects the 32 damage bonus. Since the damage is much less than under brew, we won't be able to see single digit increases (ex: won't see the '2' if the damage increases by 32), but you will see the three at least. I used mandies to artificially increase my MAB:MDB ratio because previous tests suggest MAB has no effect towards the 32 damage increase.

Results:
- No noticeable difference going from 0 to 4 skill (1878 v 1878 DMG; could be single digit increase)
- 413 damage difference going from 0 to 424 skill

Possible explanations:
- INT does affect the 32 DMG increase and single digit increments weren't visible

Concerns:
Remember that this is cutoff damage, so if there were a 32 DMG increment, we would see +3 DMG instead of +32 DMG. However, if we apply that logic to the 413, the damage bonus from elemental skill actually surpasses the gain while under brew (2464 bonus total difference between no elemental magic and 424 skill when using brew). How the hell does that make sense? Second, we don't know whether changing the INT changed the tiers, so null difference could occur due to single digit leaps or because the tier changed and we wouldn't know.

At the very least, we can say INT does affect the elemental magic damage component. This totals our findings to...

- MAB does not affect the damage increments per tier
- INT does affect the damage per elemental skill
** Either changes the tiers or changes the damage yield per tier
- Having no MAB nets a total of zero damage
** Implies that MAB is multiplying somewhere where no addition is taking place
- When brewing the damage increment per tier is 32 regardless of MAB
- With no elemental magic 34944 damage using brew; no image, but the damage is 348 with 1:1 MAB:MDB, no elemental magic, and 100 INT

[Yugl]

More tests using 300 and 333 INT:
- Same results regarding 0 v 4 skill
- Didn't do 300 INT tests after that since I wanted to use 333 for 1/3 Brew INT
- Difference between no elemental magic and 424 was 247 DMG (1164 v 1411 DMG); interestingly, 10% of what the difference is when using brew
- Used no MAB gear and this NM has 140 MDB, so MAB:MDB is 1:1

[Byrthnoth]

And when you use brew, you're doing it against a 0MDB target with +60 MAB? That would seem like pretty conclusive evidence that the skill term is multiplied by MAB/MDB, to me.

[Yugl]

I never used a brew for the last two tests. If you mean the 10% comment, the damage going from no elemental magic to 424 skill is 2464 DMG with brew and 247 during that test.

[Yugl]

Working equation for base damage: [(INT*3.5) + ((Elemental Magic Skill -2)*(0.58))]
Magic Attack Bonus works as normal using this.
Variability: By shifting the Elemental Magic reduction (The bold part) between 2/3/4, you can get +/-2 damage at most

https://www.dropbox.com/sh/79tvj5mpx19qp4m/1rBmEIpzKk

This link has an excel sheet I added summing up the latest test and images with folders detailing the stats for each image. In short:
- Elemental Skill Values 0/100/200/300
- INT Values 100/150/200/250/300
- MAB/MDB Values (140/140) and (140/100)

The key to matching to getting close to the value is the bold portion. With the written equation, the difference can hit double digits, but by adjusting the 2 to 1/3/4, you can get within single digits or exact range of the real value. I'm guessing this inconsistency is the reason the 32 damage seemed constant regardless of MAB. I already found some tiers earlier, but if we find a pattern for determining tiers, that would probably fix the issue.

[CDF]

Regarding Meteor, I did this on the test server so I hope these apply to the current game... the base damage approximation on BGWiki doesn't seem to work? https://www.bg-wiki.com/bg/Meteor

(need more data points for elemental skill and magic damage but too tedious waiting on ES recast or resetting timers and then zeroing out elemental skill)

Summary:
Meteor damage still depends on total INT, not dINT.
Meteor damage still depends on elemental magic skill.
Meteor damage is dependent on the magic damage attribute.

Data:

            Target   MAB   INT   dINT   M. Dmg   Elemental skill   Meteor dmg   Water II dmg (control)
*  Tiny Mandragora    50   100     94        0                 0          521                      514
   Tiny Mandragora    50   106    100        0                 0          550                      532
   Tiny Mandragora    50   106    100        0               424          924                      532
   Tiny Mandragora    50   116    110        0                 0          604                      547
*  Tiny Mandragora   100   100     94        0                 0          764                      686
   Tiny Mandragora   100   106    100        0                 0          809                      710
   Tiny Mandragora   100   116    110        0                 0          886                      730
   Tiny Mandragora   100   100     94       86                 0         1248                      858
   Tiny Mandragora   100   106    100       86                 0         1286                      882
   Tiny Mandragora   100   106    100       86               424         1785                      882
   Tiny Mandragora   100   116    110       86                 0         1356                      902
*    Desert Spider    50   100      ?        0                 0          521                        -
*    Desert Spider   100   100      ?        0                 0          764                        -

Notes:

• Targeted Tiny Mandragoras and then Desert Spiders at the end to determine whether Meteor damage depends on dINT or total INT. It appears still to depend on total INT (I marked the data rows with asterisks.)
• Meteor damage still appears to depend on elemental magic skill (so more data is needed to determine its contribution). (I was hoping it wouldn't so I wouldn't have to zero out my skill on the test server.)
• I checked Water II damage (without day bonus or penalty) as a control after each Meteor cast. Damage appears consistent with the changes from the recent version update. See BGWiki: https://www.bg-wiki.com/bg/Magic_Damage
• Ratio of damage between MAB+100 and MAB+50 (holding dINT and M.Dmg fixed) is unexpected for Meteor damage (~1.47 vs expected 1.33 as seen with Water II)??? Maybe MAB is not a simple multiplier here?

Equipment used to vary INT, MAB, M.Dmg (Baseline MAB 40 from BLM trait, INT 93, M.Dmg 0):

Refraction cape (INT+8)
Goetia sabots +2 (INT+10)
Laevateinn 99 (MAB+60)
Witchstone (MAB+2)
Stoicheion medal (MAB+8)
Dark rings (INT+1)
Omega ring (INT+3)
Cognition belt (INT+7)
Psystorm earring (INT+4)
Voay Staff (MAB+60, INT+7, M.Dmg+86)


Also Weatherspoon Souliers don't have MAB+10 on the test server...

[Motenten]

Don't suppose you could get a quick sample of dInt 100, 100 MAB, 424 skill, 0 Magic Damage?

[CDF]

1308 Meteor with dINT 94 (INT 100), 100 MAB, 424 skill, 0 magic damage (686 Water II)
892 Meteor with dINT 94 (INT 100), 50 MAB, 424 skill, 0 magic damage (514 Water II)

1308/892 = 1.466 vs 686/514 = 1.335 so same thing going on with MAB

[Motenten]

Er, dInt 100, not Int 100; the same as the other two samples you have in the above chart that include skill.

[Motenten]

Tried to approach things in several different ways, but kept having things not add up.

1)

OK, one that we should be certain of: 86 magic damage vs 0 magic damage should have exactly +86 base D.

+00 D @ 100 dInt, 100 MAB, 0 skill = 809
+86 D @ 100 dInt, 100 MAB, 0 skill = 1286

There's a difference of 477 final damage from the increase in 86 base damage.

@ 94 dInt, the increase is 484
@100 dInt, the increase is 477
@110 dInt, the increase is 470

Seems to drop in increments of 7 damage as dInt changes. Possibly tiered? (eg: @94 dInt, 100 dInt, 106 dInt, 112 dInt, etc)

However the real issue is that the difference between the two final damage values is *going down* while Int is increasing and Magic Damage is constant. If Magic Damage were capped the way it was on normal nukes pre-patch, the damage difference should be increasing. If uncapped, one would expect the damage increase to be constant or increasing, depending on other multipliers. The only way for it to go down is if the multiplier is decreasing as Int (or dInt) increases.

2)

Assuming MAB is not itself part of the base damage formula (further testing would be needed to rule this out), that implies...

Base damage (D) @100 Int = 487
@50 MAB, D * 1.07 = 487 * 1.07 = 521.09 => 521
@100 MAB, D * 1.57 = 487 * 1.57 = 764.59 => 764

Which would mean the spell has an innate -43 MAB. An odd number, and slightly more than enough to offset blm's native MAB (40). Trying again on the 106 int values:

Base damage (D) @106 Int = 515(?)
@50 MAB, D * 1.07 = 515 * 1.07 = 551.05 =/> 550 (too high)
@100 MAB, D * 1.57 = 515 * 1.57 = 808.55 =/> 809 (too low)

We see that the MAB value doesn't work for the next Int increment, and in fact the values move in different directions from the actual result, meaning it's not just an improperly chosen base damage value. We'd need to change the MAB offset itself.

Base damage (D) @106 Int = 519
@50 MAB, D * 1.06 = 519 * 1.06 = 550.14 => 550
@100 MAB, D * 1.56 = 519 * 1.56 = 809.64 => 809

So, again, we have an increasing penalty being applied as Int increases: -44 MAB in this case.

One further check at 116 Int:

Base damage (D) @116 Int = 565
@50 MAB, D * 1.07 = 565 * 1.07 = 604.55 => 604
@100 MAB, D * 1.57 = 565 * 1.57 = 887.05 =/> 886

In this case we're back to the -43 MAB penalty, however no configuration will match both final damage values, only one of them. So we can't trust that the MAB penalty itself is changing, but we can probably go back to the apparent general penalty seen with increasing Int from part 1.

3)

Can we get the MAB penalty effects solely from a separate multiplier that's dependent on Int? Or perhaps some sort of offset effect?

Let's try looking at the MAB differences as pure offsets first. In that case, the difference between the 50 MAB and the 100 MAB results will be half the base damage, regardless of what our apparent final damage is.

@100 Int, 50 MAB = 243, so base = 486-487
@106 Int, 50 MAB = 259, so base = 518-519
@116 Int, 50 MAB = 282, so base = 564-565

Interestingly, those values match the estimated base damage values when I tried to calculate using the offset MAB.

So what happens if we apply standard MAB multipliers to our extracted base damage values?

@100 Int, +50 MAB = 729-730 [actual=521, diff=208-209, ratio=0.7137-0.7159]
~ +100 MAB = 972-974 [actual=764, diff=208-210, ratio=0.7844-0.7869]

@106 Int, +50 MAB = 777-778 [actual=550, diff=227-228, ratio=0.7069-0.7090]
~ +100 MAB = 1036-1038 [actual=809, diff=227-229, ratio=0.7794-0.7818]

@116 Int, +50 MAB = 846-847 [actual=604, diff=242-243, ratio=0.7131-0.7150]
~ +100 MAB = 1128-1130 [actual=886, diff=242-244, ratio=0.7841-0.7863]

Here we see that the differences change at different Int levels, however the difference at +50 MAB is the same as at +100 MAB for each particular Int. Since MAB is a multiplier that should have increased the differences if it applied to these extra differentials, they have to be part of an entirely separate term that is unaffected by MAB.

I also checked the ratios at the different int levels for actual vs simple prediction. The ratios are very close together at the different Int levels, however are different at different MAB levels.

This is a bit of a conundrum. Absolute differences are consistent across MAB levels (but not Int), while ratios are consistent across Int levels (but not MAB).

Also of note, compared to #1, is that the differences are increasing at higher Ints.


4)

Going to fiddle with various ways the ratios and offsets could fit together.

@100 Int/50 MAB, ratio appears to be 0.715 (143/200)
@106 Int/50 MAB, ratio appears to be 0.7075 (141.5/200)
@116 Int/50 MAB, ratio appears to be 0.715 (143/200)

@100 Int/100 MAB, ratio appears to be 0.785 (157/200)
@106 Int/100 MAB, ratio appears to be 0.780 (156/200)
@116 Int/100 MAB, ratio appears to be 0.785 (157/200)

Because of the dip in the ratio at 106 Int, it would appear that Int has -some- effect on the ratio value. However because the lower and higher Int values both seem to have the same ratio, it's hard to say how.

@100 Int, Diff=208-209; ratio w/Int=2.08-2.09
@106 Int, Diff=227-228; ratio w/Int=2.15
@116 Int, Diff=242-243; ratio w/Int=2.09

Similarly, the difference could be a multiplier applied to Int, however the spike at 106 Int makes things difficult.


5)

Let's see what happens when we work in the samples with +Magic Damage.

@100 Int, +86 MDmg = 1248; +100 MAB (base) = 972-974, diff = 274-276; +100 MAB (+MD) = 1144-1146, diff = 102-104
@106 Int, +86 MDmg = 1286; +100 MAB (base) = 1036-1038, diff = 248-250; +100 MAB (+MD) = 1208-1210, diff = 76-78
@116 Int, +86 MDmg = 1356; +100 MAB (base) = 1128-1130, diff = 226-228; +100 MAB (+MD) = 1300-1302, diff = 54-56

Here we want to look at a couple different points.

First, comparing the observed +Magic Damage results with the predicted base using our extrapolated base damage above. Of particular note is that the difference in damage goes down dramatically as Int increases, similarly to what we looked at in #1.

The second is what the predicted damage would be if the +Magic Damage were added directly to the extrapolated base damage. Difference in damage continues to go down, but its absolute value is much smaller than in other comparisons.

If we used the Int-based flat deduction from #4, it wouldn't work here. If we add Magic Damage to base damage, the resulting Int ratio would have to be drastically reduced.



Anyway, no conclusions yet. Just putting my thoughts so far out for people to read and maybe comment on.

[CDF]

Don't suppose you could get a quick sample of dInt 100, 100 MAB, 424 skill, 0 Magic Damage?

dINT 100 (INT 106)
MAB 100
skill 424
M.Dmg 0

Meteor 1356 (incidentally this is the same value when skill = 0 and M.Dmg = 86)
(Water II 710)

[Byrthnoth]

The formula on BGwiki is from testing that was done when they first added Meteor to the test server. IIRC, it was up to 10 damage off and was just the best fit we could find. I think someone pointed out that it didn't really add up anymore when the spell was released on the real servers.

[Yugl]

Re: Meteor

I have some data you can use here. Note that "tiered" folder should have an image labeled 6, but I *think* this is actually 7 based on the SS. Best to use the SS + Chat to determine the skill just in case there were other mislabels. If you see a skill up reach a new tier before the elemental magic damage, you should subtract a skill level from the skill level given by the chat. For example:

[Skill chat shows 3]
[Chat says "Elemental Magic Skill increases to 3]

Most likely that was 2 skill.

Edit: http://www.bluegartr.com/threads/108...=1#post5417488 (The MAB/MDB isn't right of course, but the tiered idea still works).

[CDF]

More Meteor (I decided to hold INT fixed at 100 for new data points)


There appears to be actually decreasing return to magic attack bonus for some reason (holding other factors fixed) so there can't be a constant MAB penalty

INT 100, M.Dmg 0, Skill 0:
50 to 100 MAB: 521 + 243 = 764
100 to 150 MAB: 764 + 192 = 956

(if it were constant return, as might be expected, the numbers in bold would be the same accounting for rounding

INT 100, M.Dmg 0, Skill 424:
50 to 100 MAB: 892 + 416 = 1308
100 to 150 MAB: 1308 + 327 = 1635

While MAB may not be independent of the base damage calculation (I could check for MAB/MDB = 1 some other time), it still appears to function as a simple multiplier (ratio of damage at MAB 150 to that at MAB 100 is ~1.25 for all four tested combinations of M.Dmg and skill, given INT 100).

Incidentally this observed ratio is consistent with 2.5/2.0 = 1.25...

[CDF]

This data collected given 0 Magic Damage (MDMG) and 0 elemental skill (skill).


The proposed base damage multiplier appears to be a function of Magic Attack Bonus (MAB). It appears to be approximately 3 + 500/1024 between MAB 50 and MAB 100 (this is from the prior testing on the test server when Meteor came out), and ~3.83 between MAB 100 and MAB 200.

A very rough estimate of Meteor damage (given the data so far) is

floor((1+MAB/100)floor((multiplier)(INT + MDMG(750/1024) + Skill/6))

where the multiplier is 3 + 500/1024 if MAB < 100 (only consistent with MAB 50 and MAB 75 at the moment as that's all I checked) or 3.83 if MAB >= 100 (up to 200).

Other things to check would be to get data points for INT past 200 and other values of Magic Damage and elemental skill.

Data if you want to try to obtain a numerical solution for the multipliers in the above equation (if it is even close to correct):

MAB	INT	Mdmg	Skill	Damage
50	100	0	0	521
50	100	0	424	892
100	100	0	0	764
100	100	0	424	1308
100	100	86	0	1248
100	100	86	424	1740
150	100	0	0	956
150	100	0	424	1635
150	100	86	0	1558
150	100	86	424	2176
200	100	0	0	1148
50	150	0	0	787
50	200	0	0	1046
75	100	0	0	608
75	150	0	0	918
75	200	0	0	1219
100	150	0	0	1152
100	200	0	0	1532
150	150	0	0	1443
150	200	0	0	1916
200	150	0	0	1731
200	200	0	0	2300
250	100	0	0	1340

[Motenten]

Based on CDF's data, it looks like 424 skill is equivalent to ~97 Magic Damage. That works at both the 100 MAB and 150 MAB levels.

In the case of 150 MAB, the damage delta between 0 Skill/0 MDmg and 86 MDmg (602) is exactly 86 * 7. The damage delta vs one with 424 skill (679) is exactly 97 * 7.

In the case of 100 MAB, the damage delta between 0 Skill/0 MDmg and 86 MDmg (484) is 86 * 5.628. The damage delta vs one with 424 skill (544) is 96.66 * 5.628.

7/5.628 = 1.244, which is fairly close to the 2.5/2.0 ratio one would expect from 150 MAB vs 100 MAB.

With the 7x multiplier on the 150 MAB tier, if we divide that by the MAB factor we start with a 2.8 base multiplier.

Assuming 424 skill == 97 base damage

Damage delta for 424 skill @ 100 MAB: 544
Damage delta for 424 skill @ 150 MAB: 679

97 * 2.8 = 271.6

271.6 * 2.0 (100 MAB) = 543.2
271.6 * 2.5 (150 MAB) = 679

The other portion of the base damage will have to contribute the remaining 0.8 for the 100 MAB total.

Does it still seem to work for the combined skill+MDmg results?

@100 MAB:
delta +skill = 544
delta +MDmg = 484
544 + 484 = 1,028
delta +skill + MDmg = 976

@150 MAB:
delta +skill = 679
delta +MDmg = 602
679 + 602 = 1,281
delta +skill + MDmg = 1,220

So no, they do not match. The total when adding delta skill and delta mdmg independently is higher than when they're used together.

And can't even attribute it to a change in scaling due to changes in Int or MAB, since those are constant in these comparisons. So.... Dunno. Need to give it more thought.

[Motenten]

Ok, took the latest data from CDF. Looking only at the samples that have 0 MDmg/0 Skill. Looked at the absolute increase in damage from adding 50 int (50 to 100, 100 to 150), and then the ratio of that increase to the current MAB multiplier (100+MAB).

MAB     INT     Mdmg    Skill   Damage          dInt 50    dInt 50 / MAB                 dInt 100    dInt 100/MAB
50      100     0       0       521
50      150     0       0       787             266        1.7733
50      200     0       0       1046            259        1.7267             0.9737        525           3.5000
75      100     0       0       608
75      150     0       0       918             310        1.7714
75      200     0       0       1219            301        1.7200             0.9710        611           3.4914
100     100     0       0       764
100     150     0       0       1152            388        1.9400
100     200     0       0       1532            380        1.9000             0.9794        768           3.8400
150     100     0       0       956
150     150     0       0       1443            487        1.9480
150     200     0       0       1916            473        1.8920             0.9713        960           3.8400
200     100     0       0       1148
200     150     0       0       1731            583        1.9433
200     200     0       0       2300            569        1.8967             0.9760        1152          3.8400

Going from 150 to 200 Int always generates a lower delta value than going from 100 to 150 Int. The difference is not large, but it's consistent.

There does appear to be two tiers of scaling relative to the current MAB value. Below 100 MAB, it's at 1.77 for the 50 to 100 Int jump, while at 100 MAB and higher it's at 1.94-1.95.

The ratio of the dInt50/MAB ratio at 150-200 Int vs 100-150 Int is approximately 0.975 for all MAB tiers, even though the ratios themselves change at different MAB tiers.

Extended the comparison to dInt 100 (100 to 200). Now, the dInt 100/MAB ratio firms up into very consistent values: 3.5 for under 100 MAB, and 3.84 for over 100 MAB.

Then I can try to adjust the ratios to the ratio increase per int (so /50 for the dInt 50 set, /100 for the dInt 100 set). That gives me:

Ratio/Int (50)   Ratio/Int (100)

0.03547
0.03453          0.03500

0.03543
0.03440          0.03491

0.03880
0.03800          0.03840

0.03896
0.03784          0.03840

0.03887
0.03793          0.03840

Each of those ratios can be considered as representing the midpoint of their ratio ranges. So:

for MAB < 100:
125 = 0.0355
150 = 0.0350
175 = 0.0345

for MAB > 100:
125 = 0.0389
150 = 0.0384
175 = 0.0379

In each case, it's decreasing by 0.0005 for every 25 Int, or 0.0001 for every 5 Int. Since I used 100 as the base value for MAB instead of 1.0, we can multiply those by 100 for easier use and have it as -0.01 multiplier per 5 Int.

However that only tells us the value of that specific point of Int. We actually need the average value of all the Ints from 0 to N.

That allows us to baseline the multipliers in use:

For MAB < 100, bMM (base Meteor Multiplier) = 3.80
For MAB > 100, bMM (base Meteor Multiplier) = 4.14

MM (Meteor Multiplier) = bMM - Int/500

That gives us the MM value for a specific Int. However we then need to work out the average multiplier for all Ints up to that point: Average(bMM, MM).

Edit: It's actually simpler just to use MM = bMM - Int/1000, and skip the averaging step.

I put that into the spreadsheet and did some fiddling. Found that bMM for MAB < 100 fits better at 3.79 than 3.80.

However, while the slopes are flat per MAB, it's also apparent that there's an additional offset, and that the offset depends on MAB. From experimenting, it appears to be (for MAB < 100): -(22 + MAB/5), and (for MAB >= 100): -(24 + MAB/5)

The results aren't perfect, but they are very close (exact on most samples, 1 point off on a few, 2 points off on one).

bMM = base Meteor Multiplier (for Int = 0)
MM = Meteor Multiplier (for a given Int)
bMO = base Meteor Offset
MO = Meteor Offset


For MAB < 100:
bMM = 3.79
bMO = 22

For MAB >= 100:
bMM = 4.14
bMO = 24

MM = bMM - Int/1000
MO = bMO + MAB/5

Damage = (Int * MM) * (100+MAB)/100 - MO


That concludes the formula without accounting for skill or Magic Damage+.

[Motenten]

Continuing on with the Skill data. Will have to assume that there's no additional sliding scale for now, since we only have a few data points.

MAB     INT     Mdmg    Skill   Damage   Int-Only Dmg     Diff   MAB Ratio   Skill Ratio
50      100     0       424     892               521      371     247.333       0.58333
100     100     0       424     1308              764      544     272.000       0.64151
150     100     0       424     1635              956      679     271.600       0.64057

Here we see that, factoring out MAB from the difference between the 0 Skill and 424 Skill results (assuming MAB affects skill), we get two different ratio values: 247 for under 100 MAB, and 272 for over 100 MAB.

We then look at the ratios between skill and the MAB-factored ratio. For MAB < 100, the ratio is exactly 7/12. This corresponds to the wiki formula that uses (skill/6)*3.5. The 3.5 is close to the MM value used on Int, but not exact, and MM varies. It's easy to see where it would look like it was part of the same overall multiplier, though.

For MAB >= 100, there doesn't appear to be a clean ratio, but 41/64 comes within 1 point on one of the samples and exact on the other. Revised formulas is:


bMM = base Meteor Multiplier (for Int = 0)
MM = Meteor Multiplier (for a given Int)
bMO = base Meteor Offset
MO = Meteor Offset
MSM = Meteor Skill Multiplier


For MAB < 100:
bMM = 3.79
bMO = 22
MSM = 7/12

For MAB >= 100:
bMM = 4.14
bMO = 24
MSM = 41/64

MM = bMM - Int/1000
MO = bMO + MAB/5

Damage = (Int * MM + Skill * MSM) * (100+MAB)/100 - MO

[Motenten]

For +Magic Damage:

MAB     INT     Mdmg    Skill   Damage  bMM     MM @ dInt    MO      +Skill  Damage  Diff    MAB Ratio     MDmg Ratio
100     100     86      0       1248    4.14    4.04         44      0.000   764     484     242.000            2.814
150     100     86      0       1558    4.14    4.04         54      0.000   956     602     240.800            2.800

After factoring out MAB, it appears we get about 242 damage from +86 Magic Damage, or about a 2.8 multiplier. It's probably different for MAB<100, but no data on that, so skipping that part.

However we have another set of data with +Magic Damage, the samples with 424 skill on them:

MAB     INT     Mdmg    Skill   Damage  bMM     MM @ dInt    MO      +Skill  Damage  Diff    MAB Ratio     MDmg Ratio
100     100     86      424     1740    4.14    4.04         44      271.625 1307    433     216.500            2.517
150     100     86      424     2176    4.14    4.04         54      271.625 1635    541     216.400            2.516

And here the final ratios are not the same. It would appear that the multiplier for Magic Damage+ varies with Skill. Unfortunately we have no data to indicate -how- it scales. It could be hitting a damage cap, or it could be getting a reduced multiplier the same way Int is reduced (except based on Skill).

Since I have nothing else to go on, I'll make a speculative declining multiplier based on Skill. Base = 2.8, reduction = 0.67*Skill/1000.


bMIM = base Meteor Int Multiplier (for Int = 0)
MIM = Meteor Multiplier (for a given Int)
bMO = base Meteor Offset
MO = Meteor Offset
MSM = Meteor Skill Multiplier
bMSSM = base Meteor Skill Scale Multiplier
MSSM = Meteor Skill Scale Multiplier
MMDM = Meteor Magic Damage Multiplier


For MAB < 100:
bMIM = 3.79
bMO = 22
MSM = 7/12
bMSSM = 2.8

For MAB >= 100:
bMIM = 4.14
bMO = 24
MSM = 41/64
bMSSM = 2.8

MIM = bMIM - Int/1000
MO = bMO + MAB/5
MSSM = 0.67*Skill/1000
MMDM = 2.8 - MSSM

Damage = (Int * MIM + Skill * MSM + MDmg * MMDM) * (100+MAB)/100 - MO