Blue Mage: Best thread ever

[Kegsay]

Oh, I thought you were the OP.

In that case, I'll have to re-do the testing, with less averaging and a different method for obtaining the conclusion (rather than average damage which needs hundreds of trials, using min/max damage only needs 1 point of damage to disprove). Testing will simply be to disprove that STR contributes to blue attack, by obtaining a really low (impossibly low if STR is contributing to PDIF) value on head butt. If I can't prove that, then I'll be happy that STR does contribute something more than just fSTR and WSC improvements.

[Nightfyre]

I'm online and not doing anything right now, I'll gather a few data points before dinner. After will need to wait, have events.

EDIT: Or not, dinner's ready

[LD]

Oh, I thought you were the OP.

OP? There are children younger than this thread.

[Sho]

Sho, can you explain how VZX determined the base dmg for blu spells D please ?

and also testing ways ?

Ahh ok, this is the base formula for calculating Blue Magic Skill:

floor(Blue Magic Skill * 0.11) * 2 + 3

For example, my current Blue Magic Skill is 331, so you plug it in the formula:

floor(331 * 0.11) * 2 + 3

floor(36.41) * 2 + 3

"Floor" is a mathematical function that truncates (removes) the decimal place on whatever is in the braces. So now it looks like this:

(36) * 2 + 3 = 75 <- Base D

So for 331 Skill, the base D is 75. If I recall correctly, it was the folks at Studio Globi that figured these maths out forever ago in 2006~2007. The are also the reason for many of the findings in WS calculations that we know now.

http://www32.atwiki.jp/studiogobli/pages/90.html -- Studo Globi website.

As for testing, you can use this formula:

D Value＝[D+SV+WS Correction]×Multiplier

Physical Blue Magic functions almost exactly as Weapon Skills, so you can run checks vs a mobs defenses with fSTR and fSTR2 (ranged spells, not Cannonball) to see if the D Value is working correctly. Alternatively you can eyeball the changes on certain spells (Frenetic Rip, Dimensional Death) as they begin to reach the cap on their damage.

http://www.freewebs.com/vzx-01/blumag.htm -- VZX Blue Magic Pages

Speaking of the Blue Magic Attack stat, I always had that idea that it existed... but I never tested on it at all, even after Ebur/Concordia got those +Attack evolith. Maybe I'll start doing that now after I finish COR and BST to 75. Also in particular Boost (/MNK) and Souleater (/DRK) directly modifies that stat.

Hopes this helps a lil.

[Nightfyre]

OP? There are children younger than this thread.

The thread I linked to.

Sho, are you saying Attack Bonus affects spells as well?

[Sho]

The thread I linked to.

Sho, are you saying Attack Bonus affects spells as well?

Oh, I mean there was a Hidden "Blue Magic Attack" stat, that's not so hidden anymore. An example would be this, on my sword:

http://img695.imageshack.us/img695/9453/concordia.png

Between this sword (bonus can be 4~5 I think it was), and Ebur Gear for BLU, around 40~50 "Blue Attack" is possible.

[Katas]

please clarify or explain in detail /interested

[Masamune]

Good morning all,

Totally siding with Kegsay, i personnally always understood Studio Gobli clearly separated the dmg formula into 2 parts: one "static" containing base dmg D, and fSTR (+ WSC in the case of WSs and blu spells, and probably also avatar Physical BPs?), and 2nd part of equation being the randomized part so called "pDIF" which includes everything related to attack/Def ratio.

Observed DMG = RANDOMmultiplier x BaseDMG
Observed DMG = pDIF(=f(cRatio)) x [D + fSTR(2) (+WSC)]
with:
cRatio=Player_Attack/Mob_Defense - LVLcorrection

That means essentially, on topic of blu spells, that logically blu magic skill would acts like DD's weapon skills (ie Sword skill, Dagger skill etc...) in the form of 1 skill = 2Attack+2Acccuracy. Direct result of this reasoning says that increasing/decreasing blu magic skill would NOT affect AT ALL the Base Damage of the dmg formula.

That's why i asked this question Sho, Kegsay apparently had also same suspicions.

Nonetheless thank you for your (pretty clear) answer Sho XD (for info i "dissecated" Studio Gobli and VZX's sites since soon 2 years for damage formulas verification to end up in conclusion they are off, also their equations are as old as 2005, not 2006 & they don't even have anything changed post-2H update).

My second question was the most important, i'll try to keep your answer in mind when i'll have free time to test this (too busy atm with playing and updating FFXIMDCv4.5).

[Kegsay]

I think more research needs to be done in the area, namely to re-confirm Studio Gobli's formulas for the calculation of D+fSTR and then new tests to try to figure out how pDIF works with BLU spells, which Gobli has not covered.

We are, after all, assuming that BLU physical spells follow normal player pDIF calculations, which I know from experience is not a good assumption to make (read: Made that assumption on avatars, ended up working out new pDIF formulas for them). Even docterugh's testing is based under that assumption, which I have not seen any evidence to prove (relatively simple on Lv0 monsters, abusing max/min pDIF and crit min/max pDIF and see if they work with Gobli's DMG calculations).

[Kyte]

I would say while most physical spells operate on the formula for BLU attack that was found, some get an attack bonus. Most obvious examples are spells with Attack as their fTP, such as Vertical Cleave; however, Death Scissors, and to a lesser extent, Frenetic Rip seem to get some kind of hidden bonus as well.

However, I would say Disseverment adheres to that formula (which is more or less the same as the one for melee attack) quite well.

Also in particular Boost (/MNK) and Souleater (/DRK) directly modifies that stat.

Are you sure about Boost? Obviously Souleater boosts damage through HP consumption, but I have never seen or observed anything definitive on boost. I once tried to use Boost x ~10 and Asuran Claws (a spell that seems to maybe even get an attack penalty) to see if there was any kind of increase, and observed none. However, thinking back on it, Boost may only affect the first hit (has this ever been proven???)- therefore maybe I should go ahead and try a single hit spell versus a target where PDIF wouldn't be capped. Multiple boosts+spell should cause a pretty obvious result.

[Sho]

Ooooh, that's a very interesting find! Now that I think about it, you're Kegsay from Allakhazam?

Observed DMG = RANDOMmultiplier x BaseDMG almost seems like it accounts for Critical Hit rate on spells. I know every one has said again and again that physical blue magic cannot crit hit, but a factor like that simply doesn't make sense otherwise. Multipliers would seem to be static on Blue Magic, no?

cRatio=Player_Attack/Mob_Defense - LVLcorrection would also confirm Blue Attack as well.

It feels as if one should start a BLU mule from lvl1 to test accurately.

[Sho]

I would say while most physical spells operate on the formula for BLU attack that was found, some get an attack bonus. Most obvious examples are spells with Attack as their fTP, such as Vertical Cleave; however, Death Scissors, and to a lesser extent, Frenetic Rip seem to get some kind of hidden bonus as well.

However, I would say Disseverment adheres to that formula (which is more or less the same as the one for melee attack) quite well.



Are you sure about Boost? Obviously Souleater boosts damage through HP consumption, but I have never seen or observed anything definitive on boost. I once tried to use Boost x ~10 and Asuran Claws (a spell that seems to maybe even get an attack penalty) to see if there was any kind of increase, and observed none. However, thinking back on it, Boost may only affect the first hit (has this ever been proven???)- therefore maybe I should go ahead and try a single hit spell versus a target where PDIF wouldn't be capped. Multiple boosts+spell should cause a pretty obvious result.

I'm sad to say I only tested immediately after I got Jet Stream years ago. ; ; It definitely affected at least the first hit of the spell at least. Who knows how much BLU was changed since then.

[Rena]

That means essentially, on topic of blu spells, that logically blu magic skill would acts like DD's weapon skills (ie Sword skill, Dagger skill etc...) in the form of 1 skill = 2Attack+2Acccuracy. Direct result of this reasoning says that increasing/decreasing blu magic skill would NOT affect AT ALL the Base Damage of the dmg formula.

Well, first of all, 1 skill for combat skills is 1 attack and 1 acc (and above 200, 0.9 acc only per skill point), second of all, there ARE combat skills that affect damage rating - h2h does this. It isn't without precedent, and everything that has been tested strongly indicates this is the case for blue magic skill. (And for the record, they both have similar formulae too, H2H damage formula is floor(0.11*skill)+3, whereas blue magic skill is floor(0.11*skill)*2+3.)

We are, after all, assuming that BLU physical spells follow normal player pDIF calculations, which I know from experience is not a good assumption to make (read: Made that assumption on avatars, ended up working out new pDIF formulas for them).

Mobs use different pDIF formulae from PCs (main notable difference is that cap is 4.0. This was what was known pre-2h patch however, and looking at your blog, you claim the cap is 4.2, which would work if they also received the 0-5% modifier, 4*1.05=4.2), so it wouldn't be much of a surprise if avatars were held to the same rules, and comparing the two doesn't quite work. Regardless, everything that has been tested about blue magic strongly suggests it uses PC calculations, more specifically 1h pDIF, with at the very least the caps being exactly the same, which is part of the reason we could determine base damage.

And yes, I realize using one statement some here consider unproven to back up another will be considered bad logic by the same people, but that'd be one huge coincidence if it just happened to be that a) calculations are different, and b) the results, when tested, happened to match up with expected values if calculations were identical. It is far more likely and far simpler that it just happens to use the same calculations.

If there is anything that needs to be tested still, it is not base damage, but comparing skill to attack (and figuring out attack bonuses for various spells), as well as STR's effect on blue attack, if just to verify what was found so far, but the testing that has been done is hardly anything unusual or unexpected, and given what we know, it is the most likely explanation. More tests to verify that would be good, but a lot of what you two are doing feels more like reinventing the wheel than anything.

Observed DMG = RANDOMmultiplier x BaseDMG almost seems like it accounts for Critical Hit rate on spells. I know every one has said again and again that physical blue magic cannot crit hit, but a factor like that simply doesn't make sense otherwise. Multipliers would seem to be static on Blue Magic, no?

Ranged blue magic can crit, but how does this imply any other ones can?

[Kegsay]

reinventing the wheel than anything

That irks me. All I'm suggesting is that Gobli's DMG formulas are verified (I don't think they will be wrong, simple Lv0 verification tests are hardly a mammoth test) with 1-hand pDIF formulas. If they match, awesome, let's test Blue Skill/STR and attack correlation. If not, then the groundwork for testing Skill/STR and Attack is not solid, which would make those tests utterly pointless. The few (only 1 that I've been linked to on Alla) tests which looked at pDIF could've been stronger, both in samples but more importantly, the methodology. Averaging out damage is a poor way to see if pDIF has improved; using min/max is a far better way, which wasn't done. There's no data sets that he gave to work with either, so new tests will need to be done; hardly re-inventing the wheel.

[Kyte]

Another quick set of observations from Boost+Vertical Cleave didn't really suggest that it affects physical blue magic, despite the effect being consumed. It was a really small test though, so I suppose I shouldn't completely rule it out.

[Kegsay]

Another quick set of observations from Boost+Vertical Cleave didn't really suggest that it affects physical blue magic, despite the effect being consumed.

That's good to hear, since we know raw Attack doesn't affect spells, it would be a huge mess if Boost's attack bonus did.

Ooooh, that's a very interesting find! Now that I think about it, you're Kegsay from Allakhazam?

Observed DMG = RANDOMmultiplier x BaseDMG almost seems like it accounts for Critical Hit rate on spells. I know every one has said again and again that physical blue magic cannot crit hit, but a factor like that simply doesn't make sense otherwise. Multipliers would seem to be static on Blue Magic, no?

cRatio=Player_Attack/Mob_Defense - LVLcorrection would also confirm Blue Attack as well.

Yeah I am. As for critical on spells, I don't think they can happen. I'm doing Chigoe for the 500 Vermin trial for OAThrice sword, and have not once had the Chigoe 'fall to the ground' from a critical. When I finish the trial, or am near to, I will go /thf, pop Sneak Attack and see if I can force it to fall then. If so, it will strongly support that physical spells can't critical (naturally) but can be forced.

cRatio is something I'd like to know. Namely, why do the Big Three go to shit when we fight HNMs? Is it purely due to their DEF? In which case, can it be offset using tons of skill+evoliths? Is it due to Level Correction which is unavoidable? Combination of both?

[Rena]

That irks me. All I'm suggesting is that Gobli's DMG formulas are verified (I don't think they will be wrong, simple Lv0 verification tests are hardly a mammoth test) with 1-hand pDIF formulas. If they match, awesome, let's test Blue Skill/STR and attack correlation. If not, then the groundwork for testing Skill/STR and Attack is not solid, which would make those tests utterly pointless. The few (only 1 that I've been linked to on Alla) tests which looked at pDIF could've been stronger, both in samples but more importantly, the methodology. Averaging out damage is a poor way to see if pDIF has improved; using min/max is a far better way, which wasn't done. There's no data sets that he gave to work with either, so new tests will need to be done; hardly re-inventing the wheel.

Yes, those tests were insufficient, and further testing to verify exactly how blue magic skill relates to blue attack? Yes. Suggesting the possibility of using a different way of calculating pDIF, when tests at capped blue attack are completely in line with expected values for 1h pDIF? No. Reevaluating base damage, when no tests ever suggested what we know is incorrect, and every test done to verify damage tiers has supported current knowledge? No. If you really want to, of course, go ahead, and if you get different results, then there is something to discuss, but talking about it as if it is uncertain and unproven, when there is no reason to do so, and plenty of reasons to think the opposite, is silly.

cRatio is something I'd like to know. Namely, why do the Big Three go to shit when we fight HNMs? Is it purely due to their DEF? In which case, can it be offset using tons of skill+evoliths? Is it due to Level Correction which is unavoidable? Combination of both?

No actual tests on this, but it's pretty obvious to me it's a combination of both, but mainly level correction. HNMs aren't known to get massive boosts to defense, and even with plenty of defense reducing effects (dia, angon, steps, etc.), you still get pretty horrid results. You could partially offset this with blue attack, since you're probably not anywhere near cap, but with how little blue attack there is available, results will still be lacking, just like any melee without buffs (or a WS with a significant attack bonus) would be.

[Kegsay]

when tests at capped blue attack are completely in line with expected values for 1h pDIF?
[...]
every test done to verify damage tiers has supported current knowledge

Please link me to some of these tests. If some basic tests do indeed match up to the formulas, then awesome, I don't need to do tests on it. I have yet to see any basic testing which does this though, perhaps through lack of searching? I do not expect these formulas to be wrong, but they should be verified first before using them as a basis for more models. If no such tests exist, then I'll do them. It won't take long, and it will just re-confirm what we know. I don't want to get to the end of testing on BLU pDIF only to find 'whoops, Gobli forgot to carry a 1 and now the tests need to be re-thought out.'

[Rena]

Sho claimed to have done some tests just recently, ask for his results. There was also testing done around the time relic head came out to verify we could reach a new D rating, and it worked as expected. Results are somewhere in this thread, around... september 2008? I don't remember exactly when it was added, and given how huge the thread is, it might be hard to find, but I'll see if I can find it.

[Kegsay]

Sho claimed to have done some tests just recently, ask for his results. There was also testing done around the time relic head came out to verify we could reach a new D rating, and it worked as expected. Results are somewhere in this thread, around... september 2008? I don't remember exactly when it was added, and given how huge the thread is, it might be hard to find, but I'll see if I can find it.

Fantastic, thank you. Tried searching Allakhazam for new testing, but it just seems to be a regurgitation of the Studio Gobli stuff cropping up over and over.