Official Datamining Thread! (06/05/07)

[Typhus]

INT reduces the base damage of Black Magic spells and Ninjutsu.
Some people believe that INT also increases resistance to these spells, but this assertion is hotly contested by those who are certain that only MND, Elemental Skill, and Elemental Resistance affect this.

Sorry, should have specified that this section is under the topic resistance, in regards to what attributes of a target affect resistance. Basically, this line is saying that a mob with higher INT will take less damage than one with lower INT, but it is contested as to whether that same mob will have a higher resist rate.

Also, even if we accept that INT does affect resist rates, and using the supposed calculation that 2 INT = 1 magic acc, I don't think there would be much difference between the 2 as far as resists go. In that case, I still think I would go for the MAB over raw INT.

 

[Mai Ikari]

From my understanding, the whole INT-effects-resistance thing is much like STR and Attack effect damage. Two different checks, one of INT or MND vs INT or MND, and a second of MACC/skill vs MEVA (for lack of a better word).

Again, totally from my understanding of other people's tests, don't have a mage high enough to try anything myself.



And on a totally unrelated note... Found this over on MMO XOOPS, looks like it's the Valhalla armor (the damage reduction one)

 

[Gregedor]

And on a totally unrelated note... Found this over on MMO XOOPS, looks like it's the Valhalla armor (the damage reduction one)

lol, maybe the FFXI developers played some Valkyrie Profile recently.

 

[Mai Ikari]

lol, maybe the FFXI developers played some Valkyrie Profile recently.

Or maybe they're the same developers? >_>

 

[Uzor]

Or maybe they're the same developers? >_>

Well, might be. They're the same company, so they might have had contact at the very least

Valhalla better have hidden stats, 'cause the visible stats suck.

 

[pchan]

As with any time I see a formula I'm unfamiliar with ... source please. It makes a fair degree of logical sense due to similarity to melee accuracy, though you'd need to throw in the pre-200, post-200 bits.

But still, source please, I wasn't aware that this had actually been tested yet (to the point where it can just be thrown out there as an absolute formula like this), so I'm quite interested in seeing the tests, especially since I'd imagine good tests would've also tried out the whole INT and MND and CHR parts to see what - if anything - they did as well.

Sorry, I'm a numbers geek. XD

I have done the tests since noone has ever done it lol. What i did is trying to sleep 100 puddings consecutively. Those pudding have a high resistance to sleep. I use Sleep1 only and 4 different gears. First gear is naked with a Pluto's Staff. Second gear is gear 1 + AF body. Third gear is gear1 + HQ wise gear and the grammary cape. 4th gear is a standard INT/MND setup. It is important to make the tests on day that or not dark or light since you use sleep. Also Pudding have very high resist rate on fire day and very low resist rate on waters day. So test during other days of the week (wind, earth). The result are the following :

gear1 : 38% success
gear2 : 54% success
gear3 : 57% success

If you beleive that the games counts the total Mac / skill before throwing the resist roll

16 Mac= +50% success, while 15 skill=+42,11%
thus 1 magic accuracy= 1.11 skill

details of the test if you are interested

Mob : Flan in Mt Zayolm

Gear 1 : Macc 16 + Pluto Staff (280 enfeebling skill). INT 64+2 MND 70+2

0=fail, 1=succes

0100000011|0100101101|0000000110|1000101101|1101111011|1111011100|1111010000|0101111111|0110000101|1111111111| 57/100

Gear 2 : Naked + Pluto Staff

0011110010|1000011110|0010100000|0101000000|1101001010|1000000100|1001010001|0001110100|0110011100|0000011101| 38/100


Gear 3 : AF body + Pluto Staff

1101001000|1011011010|1111110010|0110111101|0001100100|1011010101|1110010011|1100011100|0001000011|1111101100| 54/100

I didn't write on my computer the result of the gear 4, but they showed no difference while casting sleep with int/mnd setup and casting naked. So mnd/sleep don't affect sleep. On a side note, i also tested anrin obi on darksday, and it didn't affect resist rate on sleep .

I'm assuming the ratio is the same for elemental, though i'm still testing this.

 

[Mai Ikari]

Well, might be. They're the same company, so they might have had contact at the very least

Valhalla better have hidden stats, 'cause the visible stats suck.

Half a D.ring on your body or head sucks? <.<;; (I'll admit the slow makes it less than the ultimate do-everything gear, but did PLD have a peice like that in the first place?)

 

[Gregedor]

Half a D.ring on your body or head sucks? <.<;; (I'll admit the slow makes it less than the ultimate do-everything gear, but did PLD have a peice like that in the first place?)

Nope, they sure didn't. Extra -10% damage is sexy. And they don't say only physical or only magical which is also nice.

 

[Gregedor]

Well, might be. They're the same company, so they might have had contact at the very least

I'm not so sure, the developers of the various FFXIII games have admitted they never talk to one another even though it's all supposed to be set in the same game world.

 

[Stouter]

100 sleeps on a mob isn't an accurate enough test to call it 1.1 over 1, especially when you're doing it against a mob that comes in 2 different levels.

INT will do quite a bit for your resist rate. Get your INT higher than the mob's, and your skill up to what would be around an A+ at the mob's level, and you will really start to notice it.

 

[pchan]

100 sleeps on a mob isn't an accurate enough test to call it 1.1 over 1, especially when you're doing it against a mob that comes in 2 different levels.

INT will do quite a bit for your resist rate. Get your INT higher than the mob's, and your skill up to what would be around an A+ at the mob's level, and you will really start to notice it.

100 sleep is a lot of spell. This test gives an idea about the ratio macc/skill. It's close to 1/1.

Now for int stuff, bring your blm with your highest INT and try to sleep them 100 times. I'll laugh very hard.

 

[Stouter]

Close to 1/1 is a fair thing to say, but it's not nearly enough of a sample to say something like it's 1.1 skill. I'll try it out with my blm after I finish up enfeebling merits and am actually close to an A skill for thier level, although I'd be down almost 30 INT from having to wear enfeebling gear to hit that point.

 

[pchan]

Close to 1/1 is a fair thing to say, but it's not nearly enough of a sample to say something like it's 1.1 skill. I'll try it out with my blm after I finish up enfeebling merits and am actually close to an A skill for thier level, although I'd be down almost 30 INT from having to wear enfeebling gear to hit that point.

I agree that the decimal might not be exact, but at least the ratio is not 0.5 or 1.5 like some ppl claimed. Also i made some test with my rdm nuking VT wamoura prince with different gears. It's easier done than enfeebling. The idea is to collect the Full dmg/ 1/2 resist / 1/4 resist / 1/8 resist frequencies. I cast tier 1 nukes on them and the summary is :


..........Full dmg.....>1/8 resist........> 1/4 resist
Naked..38%............74%..................63%
Macc...52%............86%..................68%
INT......37%............73%.................52%

Macc gear is done with 15 macc while int gear has 30 int boost. This is done to test the identity 2 int=1 mac. My base elem-skill with merit is 242.

First colon is the % of non resisted nukes, second is the % of nukes that were 1/4 resisted at worst and third colon those that were 1/2 resisted at worst.

As you see, macc always wins, and int doesn't do shit, just like for sleep. Once again i only did 100 nukes with each gear. But this gives an idea.

 

[Mattaru]

because BST shouldn't have been in the original E set to begin with.
WAR DRK are in a different class than pet jobs.
come to think of it; PLD shouldn't be in that list either.

Pet jobs do not have a specific armor set, in fact all of them are for the most part different. SMN, DRG, PUP, BST. So I don't even know what you are talking about.

E set is more chain than plate armor, which Valhalla seems to be.

 

[Sanoske]

Thanks for those test results. Interesting stuff. I was always told that Mag Acc was pretty worthless to a BLM... not sure why I was told that. But I never really had any Mag Acc stuff to test for myself.

 

[Bajaratt]

I agree that the decimal might not be exact, but at least the ratio is not 0.5 or 1.5 like some ppl claimed. Also i made some test with my rdm nuking VT wamoura prince with different gears. It's easier done than enfeebling. The idea is to collect the Full dmg/ 1/2 resist / 1/4 resist / 1/8 resist frequencies. I cast tier 1 nukes on them and the summary is :


..........Full dmg.....>1/8 resist........> 1/4 resist
Naked..38%............74%..................63%
Macc...52%............86%..................68%
INT......37%............73%.................52%

Macc gear is done with 15 macc while int gear has 30 int boost. This is done to test the identity 2 int=1 mac. My base elem-skill with merit is 242.

First colon is the % of non resisted nukes, second is the % of nukes that were 1/4 resisted at worst and third colon those that were 1/2 resisted at worst.

As you see, macc always wins, and int doesn't do shit, just like for sleep. Once again i only did 100 nukes with each gear. But this gives an idea.

Why do your tests add up to 175%, 206%, and 162% respectively? It throws off the value of the tests when you reach impossible outcomes.

 

[pchan]

Thanks for those test results. Interesting stuff. I was always told that Mag Acc was pretty worthless to a BLM... not sure why I was told that. But I never really had any Mag Acc stuff to test for myself.

Blm don't have Magic accuracy stuff, or very little actually. There is this Aureole dropped from AV, Grammary cape (+1) and diabolos earring (+2 on darkweather).
So it is quite hard to test Macc on BLM. But with this new gear they get macc now.

 

[pchan]

Why do your tests add up to 175%, 206%, and 162% respectively? It throws off the value of the tests when you reach impossible outcomes.

Because you didn't understand what i said. I didn't list the % for each type of resist, but the % of nukes that were 1/8 resisted or more, 1/4 resisted or more etc.. I did that to be coherent with the tests of sleep. When you cast sleep, you only know if you are full resisted or if spell lands. You don't measure the intermediate resist rates.

Let me rewrite the % for each type of resist :

..........Full dmg.....1/2 resist........1/4 resist......1/8 resist
Naked..38%............25%..................11%........26%
Macc...52%............16%..................18%........14%
INT......37%............15%.................21%........27%

 

[Stouter]

i made some test with my rdm nuking VT wamoura prince with different gears.

..

Macc gear is done with 15 macc while int gear has 30 int boost. This is done to test the identity 2 int=1 mac. My base elem-skill with merit is 242.

...

As you see, macc always wins, and int doesn't do shit, just like for sleep. Once again i only did 100 nukes with each gear. But this gives an idea.

You did somewhat prove that 2 INT doesn't give 1 Macc, but I don't think anyone was debating that. You just proved that INT doesn't do shit when your skill is horribly under the mob's.

INT will do quite a bit for your resist rate. Get your INT higher than the mob's, and your skill up to what would be around an A+ at the mob's level, and you will really start to notice it.

You were still below the mob's A+ level on skill, and I'm not sure you got up over to the mob's level with that Elvaan Rdm INT.

 

[pchan]

You did somewhat prove that 2 INT doesn't give 1 Macc, but I don't think anyone was debating that. You just proved that INT doesn't do shit when your skill is horribly under the mob's.

You were still below the mob's A+ level on skill, and I'm not sure you got up over to the mob's level with that Elvaan Rdm INT.

If i bring my skill to 290 i know what will happen lol. Basically i would get almost no resists, w/e my int.