Blue Mage: Best thread ever

[doctorugh]

Can you add the data here please? Also, this was a while back, but I could use the pDIF data as well (I think Proth said he had this?).

Target EM Goblin Plunderer (THF - was testing -acc on HS also). Def 430, capped acc acc at 560 (determined through "check")

STR 222 (fSTR capped)
DEX 192
INT 206
blue skill 440 (D=99)
expected blue att = 559 (cratio 1.3)

Roughly 50-60 casts on each spell tested, mob was never finished with multi-hit.

Disseverment = 1240 (at 95% acc expected 1257) (Used as baseline)

VD = 1129 (with no +att would be 803) +40% bonus

HS = 50% acc ( so -90 acc not enough cast to pin down acc exactly, but likely -80 to -100)

GR = 1519 (at 95% acc expected 1092) +39% bonus (likely40%)

QS = 1408 (at 95% acc expected 1111) ( I determined no crit chance unless CA or Efflux in another test) +26% bonus (likely 25%)

AS = 1326 (at 95% acc expected 1179) +12.5% (Prothescar if you have data on efflux testing on this I would appreciate, wonder if it get + attbonus +ftp bonus, I might test later but efflux testing is.....tedious)

BR (w/ Efflux) = 1353 (at 95% acc expected 1060) +27.5% (not enough casts to prove anything specific on this other than to generally say it does get a +att bonus)



All of this assumes blue attack = skill + 50%STR + 8base. None of this testing would indicate any +blue attack from potency merits, and I still believe its acc, I plan on testing acc of these multihits, now that Ive got the +att bonus close enough to ballpark a +acc

[fantasticdan]

Does the attack bonus on these spells apply to all hits?

[doctorugh]

yes

***I may need to use a ranged attack and keep cratio btw .9 and 1.1 to determine if there is any validity to potency = +att, but i thought someone else was already doing this

[Nightfyre]

yes

***I may need to use a ranged attack and keep cratio btw .9 and 1.1 to determine if there is any validity to potency = +att, but i thought someone else was already doing this

Two pages back. There is absolutely an attack bonus, I tested with and without potency merits to verify.

[doctorugh]

Can I get the data on that (either here or PM), my plan was to find def of crab and add blue attack until cratio > 1.1 (where # would change) to determine exact bonus. If it was +10 the att values I have for other might move down by 2.5% (ie GR @ 37.5% bonus, AS @ 10% bonus), but if it was +25....that would....mess things up.

[Prothescar]

Need some more in depth testing on the attack bonuses, sample sizes still being quite small and the possibility of capped ratio being there, particularly with the damage of the spells being only slightly higher than expected. I'm almost certain that I had Vanity Dive pegged with a 66% or so attack bonus. Discrepencies in your findings, particularly Goblin Rush @ 40% instead of 37.5%, indicates some sort of tiny attack boost, i.e. potency merits. Without bigger sample sizes and more exact data, however, it'd be impossible to say.


Potential +10 attack would be so tiny that there'd be much better ways to test for it. Test Server with an artificially limited skill would be best, and was how I was testing it prior to being caught up in other things and looking at ranged pDIF on blue magic (which was looking very strange).

[doctorugh]

Need some more in depth testing on the attack bonuses, sample sizes still being quite small and the possibility of capped ratio being there, particularly with the damage of the spells being only slightly higher than expected. I'm almost certain that I had Vanity Dive pegged with a 66% or so attack bonus. Discrepencies in your findings, particularly Goblin Rush @ 40% instead of 37.5%, indicates some sort of tiny attack boost, i.e. potency merits. Without bigger sample sizes and more exact data, however, it'd be impossible to say.


Potential +10 attack would be so tiny that there'd be much better ways to test for it. Test Server with an artificially limited skill would be best, and was how I was testing it prior to being caught up in other things and looking at ranged pDIF on blue magic (which was looking very strange).

No chance of cap 1.3 x 1.4 = 1.82. A +10% difference would not be statistically possible. When I get a chance ill just do the ranged attack on crabs to pin down that exact bonus, shouldnt take more than 30 min. +10 skill seems possible and 2.5% would be within reason of the smaller sample. I've tested VD and GR before a come up with the same results, but I had lost that data a while back. Mostly was testing for QS, which upon eyeballing while screwing around seemed larger than it should have been and AS which I've always been suspect about the magnitude of its attack bonus.

[Yugl]

I haven't worked with physical spells for a while (Or even BLU lol), so look through this math carefully (Though the pDIF should be right):

Spell = VD
Damage = 99
fSTR Cap = (99/9) + 8 = 19
WSC = FLOOR(FLOOR(192/2)*(0.85)) = 81

Total base damage = (99 + 19 + 81) = 199

fTP = 3

Attack = 559
Mob Defense = 430

Ratio = (559/430) = 1.3
cRatio = (1.3 - 0) = 1.3

Using Motenten's PDIF Model,

wRatio = (1.3 + 0) = 1.3
qRatio UL = (1.3 * 0.25) + 1.3 = 1.625
qRatio LL = 1

UL Max Random = (1.625 * 1.05) = 1.70625
UL Min Random = (1 * 1) = 1

pDIF Range: 1-1.70625

FLOOR[(199)*(3)*(1.70625)] = 1018
FLOOR[(199)*(3)*(1)] = 597

So, there is some sort of bonus going on if your high or avg (Didn't specify which) is 1129.

If the bonus is 40% to ratio:

1.3*1.4 = 1.82

qRatio UL = 1.82 + 0.375 = 2.175
qRatio LL = (1.82 * (1176/1024)) - (755/1024) = 1.35285

UL Max Random = (2.175 * 1.05) = 2.28375
LL Min Random = 1.35285

pDIF Range: 1.35285-2.28375

FLOOR[(199)*(3)*(2.28375)] = 1363
FLOOR[(199)*(3)*(1.35285)] = 807

Did you see any samples up to 1363 or less than 807?

[Prothescar]

There's something weird going on with ranged pDIF or base attack. Currently I have the breaking point in damage consistency on a monster with 70DEF @ 221 -> 222 attack, which would put ratio at 3.17~. The possibility presents itself of ranged attacks not having a base attack of 8, so that could be throwing things off a bit. Can't draw any hypotheses until I bother doing some more ranged pDIF and base attack tests though. I have enough here to be able to get a close or exact value for Physical Potency merits at least, should have some results soon.

[doctorugh]

I haven't worked with physical spells for a while (Or even BLU lol), so look through this math carefully (Though the pDIF should be right):

Spell = VD
Damage = 99
fSTR Cap = (99/9) + 8 = 19
WSC = FLOOR(FLOOR(192/2)*(0.85)) = 81

Total base damage = (99 + 19 + 81) = 199

fTP = 3

Attack = 559
Mob Defense = 430

Ratio = (559/430) = 1.3
cRatio = (1.3 - 0) = 1.3

Using Motenten's PDIF Model,

wRatio = (1.3 + 0) = 1.3
qRatio UL = (1.3 * 0.25) + 1.3 = 1.625
qRatio LL = 1

UL Max Random = (1.625 * 1.05) = 1.70625
UL Min Random = (1 * 1) = 1

pDIF Range: 1-1.70625

FLOOR[(199)*(3)*(1.70625)] = 1018
FLOOR[(199)*(3)*(1)] = 597

So, there is some sort of bonus going on if your high or avg (Didn't specify which) is 1129.

If the bonus is 40% to ratio:

1.3*1.4 = 1.82

qRatio UL = 1.82 + 0.375 = 2.175
qRatio LL = (1.82 * (1176/1024)) - (755/1024) = 1.35285

UL Max Random = (2.175 * 1.05) = 2.28375
LL Min Random = 1.35285

pDIF Range: 1.35285-2.28375

FLOOR[(199)*(3)*(2.28375)] = 1363
FLOOR[(199)*(3)*(1.35285)] = 807

Did you see any samples up to 1363 or less than 807?

Low was 982, High 1385 (so the high would put it over 40% on this model, but on pchans model it could go up to 1420, if im looking at that right)

{(1.82-0.75)x(10/9) + 1}x1.05 = 2.298 x 3 x (206) = 1420

I remember looking at Motenten's model and seeing something wrong with it before, so i dont really use it. The min on pchans would be 876, which would put it alot closer to the min I had.

[Helel]

Is there some kind of hidden effect on sanus ensis? I did bhaflau remnants for a while on bst, and the empathetic flans typically go into spikey head mode almost immediately, and never change back. I switched to going on blu for the extra TH, and I've noticed it takes them a long ass time to go into spike mode using almace + sanus ensis, infact a couple times I've had to switch to my 1 dmg cooking dagger + dream bell to avoid killing them prematurely. I didn't think much of it until today, when I had the rampart not pop after killing all 4 after going into spike mode. I looked at the log, and judging by my melee damage, one of them switched out of spike mode right before it died. I wasn't using requiescat/skillchains/magical spells/etc. I had this happen once before, but I had a thf with me, and I figured it was a fluke skillchain or something (despite it not being in either of our logs).

The only other thing I can think of is physical spells are counting for their 'magic damage' counter.

Not sure if this is your problem since you used the term "magical spells," but physical spells will prevent them from going into spike mode as well. They can't go into spike mode while casting a spell/using a tp move either. The two flans in the side rooms have much lower defense; it's definitely possible to one-shot them with a WS before they go into spike mode.

[Prothescar]

Looking like we have our own set of defender merits here guys. will leave it at that until I finish my redundant tests

[fantasticdan]

Not sure if this is your problem since you used the term "magical spells," but physical spells will prevent them from going into spike mode as well. They can't go into spike mode while casting a spell/using a tp move either. The two flans in the side rooms have much lower defense; it's definitely possible to one-shot them with a WS before they go into spike mode.

Thanks for the confirmation on that. In both cases for me, all 4 flans had entered spike mode at some point, so apparently they can knock them back out of spike mode as well.

[Prothescar]

k, comfortable with what I have for now. Will have to do tests with non ranged spells next time I'm feeling masochistic to be safe, but...

Worked out that saplings in Jugner Forest have two levels, and their DEF values are 70 and 74. I isolated the 70DEF targets and came to the conclusion that, to cap ranged pDIF on Feather Storm for these, you need 222 Blue Magic Attack. (Inconsistency popped up at 221, no inconsistency at 222 after it became statistically improbable that an inconsistency would occur).

I dropped 10 attack and put 5 merits into Physical Potency at this point. I planned on using the 74DEF monsters to rule in or out a larger attack bonus, possibly percentage based, and it worked just fine.

@ -10 attack from cap: Inconsistencies popped up within three casts on both the 70DEF targets and the 74DEF targets.

So I added 5 more attack via Mavi Tathlum.

@ -5 attack from cap: No inconsistencies presented themselves in a period of what I believe to be statistically sound in terms of the improbability of an inconsistent sample popping up.

Spoiler: show

70: 522 522 522 522 522 652 522 522 522 522 522 522 652 522 522 522 652 522 522 652 652 522 522 522 522 522

74: 516 516 516 516 516 645 516 516 516 516 516 516 516 516 516 516 516 516 516 516 516 645 645 516 516

Current Conclusion: 5/5 Physical Potency = +5 Blue Magic Attack, based on the precedent set by my tests as I was searching for pDIF cap on the target and my tests after adding merits. Probability of each merit providing 1 attack is likely.

Issues I have: There's a possibility that these merits don't affect ranged attacks, however the fact that I was able to cap attack 5 points earlier, or at least get very convincingly close, suggests that this isn't the case. Therefor, it's possible that, in our currently limited understanding on how ranged pDIF works and how it works with blue magic, a margin of error is possible, however the probability of an attack bonus of more than 2 or so attack per merit is incredibly unlikely, and the possibility of more than 1 attack per merit is at this point unproven.

[Yugl]

Low was 982, High 1385 (so the high would put it over 40% on this model, but on pchans model it could go up to 1420, if im looking at that right)

{(1.82-0.75)x(10/9) + 1}x1.05 = 2.298 x 3 x (206) = 1420

I remember looking at Motenten's model and seeing something wrong with it before, so i dont really use it. The min on pchans would be 876, which would put it alot closer to the min I had.

Where is the 206 from?

Spell = VD
Damage = 99
fSTR Cap = (99/9) + 8 = 19
WSC = FLOOR(FLOOR(192/2)*(0.85)) = 81

Total base damage = (99 + 19 + 81) = 199

If you use 199 with pChan's model, you only get 1371, so the damage would exceed that model's prediction as well.

Edit: I see the fSTR cap set to 26 on the calculation page.

[Yugl]

With 206 base damage using Motenten's pDIF:

836-1411 damage, so both values are within range.

[doctorugh]

Current Conclusion: 5/5 Physical Potency = +5 Blue Magic Attack, based on the precedent set by my tests as I was searching for pDIF cap on the target and my tests after adding merits. Probability of each merit providing 1 attack is likely.

Issues I have: There's a possibility that these merits don't affect ranged attacks, however the fact that I was able to cap attack 5 points earlier, or at least get very convincingly close, suggests that this isn't the case. Therefor, it's possible that, in our currently limited understanding on how ranged pDIF works and how it works with blue magic, a margin of error is possible, however the probability of an attack bonus of more than 2 or so attack per merit is incredibly unlikely, and the possibility of more than 1 attack per merit is at this point unproven.

So mostly you are saying, some jobs get like triple attack +1%, where as blue mage gets 0.2% damage to blue physical spells. Yeah that seems balanced. This wouldnt really upset any testing I've done with limited samples, being as its so small.

[Yugl]

The description makes no sense then lol.

[dasva]

The description makes no sense then lol.

Sounds like ffxi to me lol

[doctorugh]

2 = 1, working as intended.