Magic Accuracy and INT

[Pocketrockets]

Nashira seraweels have 5 skill 3 macc, compared to 10 int on Morrigan's. That's the main slot i can see it mattering.

<edit> Also, the best earrings for an accuracy set are the new ones that give 5 macc when you have both.

[BerenTebogo]

If I'm looking at this correctly, and others can confirm the testing procedure was not flawed, then it should indicate what gear choices are better for certain mobs give their INT. For instance, if the enemy's INT is 60, you know that after 75 INT, you're gaining less magic ACC from INT.

My question, however, is two fold and concerns the nature of the spells themselves. Are these spells affected by INT in any manner? Last I checked, I recall people saying INT did not affect the potency of said spells; so, does this mean the dINT test is only for magic ACC comparison purposes (i.e. determining whether Demon Helm or Elite Beret will provide more magic ACC) or is there a potency effect to calculate as well?

My understanding is that it has been most recently left as no potency. Obviously if there does turn out to be a potency multiplier then yes, this info skyrockets in value.

@Above: Yeah, spaced out on nashira but 3skill+2int earring, going on this maff, should be the same as the new earrings without the hassle and inventory loss (since they're probably gear you have already) but your point is still valid, I forget them

[Nevex]

All these tests fail to consider mob's elemental resistance. For what we know, test subjects might have a -20 ice resist and a +20 wind one, we just don't know it and there's no way to know unless it's a very obvious resistance like crabs-water, imps-dark and so on.

Before testing enfeebes people should have an idea of their elemental resistance values, even if they're small. Until then these results are "slightly accurate" at best. All we know is that INT and MND play a role on landing rates for their respective spells, but giving out numbers as if they're 100% correct is retarded.

[Yugl]

Unless Hpmede (SP) have elemental resistance adjustments for all elements (And that is possible), it's possible to flush out element resistance using the same test with different spells such as Gravity or Silence. Of course, there are a few elements such as Fire or Dark that will pose a problem in deducing all elemental resistances using this method, but it's closer to predicting whether it plays a heavy hand in all of this.

[Taruking]

There is no reason that this doesn't apply to HnM. And saying that carreer rdm only care about HnMs is the most stupid thing I've heard on this forum. Nothing worth your time drops there. Stop the bullshit, hnm=2004.

Idiotic as usual.

Any RDM who can't land enfeebling magic on any non HNM mob which are not immune to corresponding elements with current available gear should stop playing the job. We don't need your result to help us land spell on regular mobs. Landing enfeebling magic on HNM is a major challenge for career RDM, esp HNM introduced after TAU release. What you have proved has no impact on endgame battle because INT do not affect slow and paralyze.

Even for a solo RDM, landing gravity and bind is relatively easy with current enfeebling gear and it convers BOTH INT and MND based spells.

[pchan]

All these tests fail to consider mob's elemental resistance. For what we know, test subjects might have a -20 ice resist and a +20 wind one, we just don't know it and there's no way to know unless it's a very obvious resistance like crabs-water, imps-dark and so on.

Before testing enfeebes people should have an idea of their elemental resistance values, even if they're small. Until then these results are "slightly accurate" at best. All we know is that INT and MND play a role on landing rates for their respective spells, but giving out numbers as if they're 100% correct is retarded.

You don't knwo what "mob's elemental resistance" means. Each mob has a given magic evasion to a given spell. The magic evasion that I determined (243) is therefore the magic evasion to the spell "bind". This is the result of all the possible resistance that you put on the mobs : a magic evasion number, so you don't have to determine it.

[FailureMidgard]

Ok, feel free to correct any wrong information here, as I got it from wiki, but multiple people said the table was correct:

Target: Even Match Om'Hpemde for a lvl 75 char (used a mule).
Player: lvl 66 pld/blm with 50 int, no gear on at all
Spell: ES Blizzard1, Windsday.
Resulting Damage: 52

Calculation:
For dINT < 0: D = (V + dINT)*MAB
According to Wiki, V for Blizzard1 is 46. MAB for a 33blm is 1.24
IF INT=58:
( 46+50-58 )*1.24=38*1.24=47.12~=47
IF INT=54
( 46+50-54 )*1.24=42*1.24=52.08~=52


Feel free to point out where I erred, as I am assuming multiple things in these formulas, but I have seen nothing to contradict any of the formulas or numbers I used.

[Klutz]

My understanding is that it has been most recently left as no potency. Obviously if there does turn out to be a potency multiplier then yes, this info skyrockets in value.

@Above: Yeah, spaced out on nashira but 3skill+2int earring, going on this maff, should be the same as the new earrings without the hassle and inventory loss (since they're probably gear you have already) but your point is still valid, I forget them

I'm watching Rumble in the Bronx and have no idea what is going on.

[pchan]

Landing enfeebling magic on HNM is a major challenge for career RDM, esp HNM introduced after TAU release.

If you have RDM leveled and can't land slow on fafnir, you need to go to the AH and buy an HQ staff ; this test's purpose is not to determined if you can land or cannot land ; it's to compare various set of gears.

What you have proved has no impact on endgame battle because INT do not affect slow and paralyze.

Even for a solo RDM, landing gravity and bind is relatively easy with current enfeebling gear and it convers BOTH INT and MND based spells.

I was hoping that an idiot like you would understand that what works for INT based enfeebling spells (bind,gravity,sleep,poison) also works for MND enffeebling spells (slow,silence). I guess I put my hopes a little up. The relation int-mac is probably the same as mnd-mac for mnd spells, and there is no reason to doubt it.

Even for a solo RDM, landing gravity and bind is relatively easy with current enfeebling gear and it convers BOTH INT and MND based spells.

This doesn't make sense to me.

[BerenTebogo]

I'm watching Rumble in the Bronx and have no idea what is going on.

Celebrating 100 years of Jackie Chan!

[Kaeko]

First off,

Raw data converted to confidence intervals is correct (double checked his math). Assuming he did the basic z=+/-1.96 and standard error is sqrt(p(1-p)/n). So assuming raw data was collected in good faith the numbers are accurate. How you interpret the data is up to you.

*****

My question is how this data was collected? Was it some sort of bot + parse? I ask this because you used bind, which has a long recast of 45 base. With good haste gear let's say this is 35 seconds to recast Bind. If you count the number of total trials you conducted in this posted data set, it comes to 44,906 attempts. Multiply this by ~35 seconds (estimated minimum time to cast Bind again), this comes to:

1,571,710 seconds OR
26,195 minutes OR
436.6 hours OR
18.2 DAYS

This is the absolute minimum no delay cast right when the recast timer hits 0, ignoring casting time, MP, etc. This took at minimum, not ever stopping 18.2 days to conduct. Did you bot this? Because I would go crazy doing this in person. 44,906 trials is over double Lodeguy's total data set and he was using an ~8 second recast spell, not a 45 second recast spell. Just estimating, this data set would take at least 16x longer to conduct than Lodeguy's -entire- data set. What exactly was your methodology to collect data?

I don't want to be the asshole that cries faked data but any good statistician will hound over the data collection methodology. Considering the utter insanity of the trial count in addition to the choice of long recast spells I have to wonder. Then when considering your data and results, the math was fine, but the numbers are so incredibly perfect it's mind-boggling.

Of course there is always a chance, but running ~7500 trials (trial1) and hitting 49.99% is crazy. But let's just give that the benefit of the doubt. Weirder things have happened. Tests 4/5 (the ones that conflict with Lodeguy's data and are most relevant) hit what are essentially the most perfect value you can want given your ultimate prediction of dINT=15 is the cutoff, not dINT=10. The absolute perfect value for test 4 is T1Average + 18, or 68%. You got 68.6%. The absolute perfect value for test 5 is T4Average + 2.5, or 71.1%. You got 71.3%. I know the error bars are tight due to the large trial count, but the phrase "too perfect" comes to mind. We all know the "law of large numbers" tells us that eventually we will hit the true average with infinite trials, but it just sets off red flags for me.

I know I sound like a total asshole here because I'm essentially accusing you of faking data, but please understand why some of these things can bring up red flags. The near "too perfect" results coupled with just the insanity of conducting a trial of this scale... I don't know you outside of these boards and unlike a lot of people here I don't particularly dislike you, so please don't take huge personal offense to this. I would attempt to pick this to pieces if Rukenshin himself posted it. If you did a bot/parse I can conceivably see this, but manually I think this is almost inconceivable.

On a less objective note, calling people with legitimate questions idiots doesn't help your cause. Taruking's point on extrapolating data collected on normal mobs > HNMs is at the very least slightly valid. My answer to this would just be if you cannot make this "leap of faith" then the data is meaningless to you, but for most players, they generally believe that the MACC system is universal as a game mechanic and not mob specific. The same can be said about my enmity testing - everything on that is based on a LVL0 mandragora, yet people use this for HNMs.

[Taruking]

If you have RDM leveled and can't land slow on fafnir, you need to go to the AH and buy an HQ staff ; this test's purpose is not to determined if you can land or cannot land ; it's to compare various set of gears.

I can see why you said "HNM = 2004", apparently you are still living in the world of 2004 when the toughest HNM is still faf/nid, turtle, Behemoth.

I was hoping that an idiot like you would understand that what works for INT based enfeebling spells (bind,gravity,sleep,poison) also works for MND enffeebling spells (slow,silence). I guess I put my hopes a little up. The relation int-mac is probably the same as mnd-mac for mnd spells, and there is no reason to doubt it.

You are building your arguement based on assumption. You assume that what works on regular mob also works on HNM. You assume that INT affects magic accuracy in the way you just described will also apply to MND based spells. By default there is no reason for you to run those tests either because we can already assume it will affect enfeebling magic the same way INT affects elemental nuke accuracy. It's possible that SE could have implemented a slightly different formula for HNMs, for example a bigger dINT which explains why some HNM are so resistant to enfeebling magic.

In the world of statistics, if you want your results to mean something always run your tests on the real subject.

[pchan]

Ok, feel free to correct any wrong information here, as I got it from wiki, but multiple people said the table was correct:

Target: Even Match Om'Hpemde for a lvl 75 char (used a mule).
Player: lvl 66 pld/blm with 50 int, no gear on at all
Spell: ES Blizzard1, Windsday.
Resulting Damage: 52

Calculation:
For dINT < 0: D = (V + dINT)*MAB
According to Wiki, V for Blizzard1 is 46. MAB for a 33blm is 1.24
IF INT=58:
( 46+50-58 )*1.24=38*1.24=47.12~=47
IF INT=54
( 46+50-54 )*1.24=42*1.24=52.08~=52


Feel free to point out where I erred, as I am assuming multiple things in these formulas, but I have seen nothing to contradict any of the formulas or numbers I used.

No idea ... I just redid the test on rdm/whm, so 24 MAB.

Stone I :

INT 58 > 12 damage ;=floor ( (10+58-58 )*1.24 )
INT 59 > 13 damage ;=floor ( (10+59-58 )*1.24 )
INT 60 > 14 damage ;=floor ( (10+60-58 )*1.24 )

Blizzard I :

INT 58 > 57 damage ;=floor ( (46+58-58 )*1.24 )
INT 59 > 58 damage ;=floor ( (46+59-58 )*1.24 )
INT 60 > 59 damage ;=floor ( (46+60-58 )*1.24 )

[pchan]

It's possible that SE could have implemented a slightly different formula for HNMs, for example a bigger dINT which explains why some HNM are so resistant to enfeebling magic. .

Stop your crap already. The reason some HnMs are very resistant to slow is because SE gave them an absurd amount of magic evasion, to the point that you are at the lower 5% cap. This has nothing to do with the "fact" that they have high mnd (proof ?).

[pchan]

First off,

Raw data converted to confidence intervals is correct (double checked his math). Assuming he did the basic z=+/-1.96 and standard error is sqrt(p(1-p)/n). So assuming raw data was collected in good faith the numbers are accurate. How you interpret the data is up to you.

*****

My question is how this data was collected? Was it some sort of bot + parse? I ask this because you used bind, which has a long recast of 45 base. With good haste gear let's say this is 35 seconds to recast Bind. If you count the number of total trials you conducted in this posted data set, it comes to 44,906 attempts. Multiply this by ~35 seconds (estimated minimum time to cast Bind again), this comes to:

1,571,710 seconds OR
26,195 minutes OR
436.6 hours OR
18.2 DAYS

This is the absolute minimum no delay cast right when the recast timer hits 0, ignoring casting time, MP, etc. This took at minimum, not ever stopping 18.2 days to conduct. Did you bot this? Because I would go crazy doing this in person. 44,906 trials is over double Lodeguy's total data set and he was using an ~8 second recast spell, not a 45 second recast spell. Just estimating, this data set would take at least 16x longer to conduct than Lodeguy's -entire- data set. What exactly was your methodology to collect data?

I don't want to be the asshole that cries faked data but any good statistician will hound over the data collection methodology. Considering the utter insanity of the trial count in addition to the choice of long recast spells I have to wonder. Then when considering your data and results, the math was fine, but the numbers are so incredibly perfect it's mind-boggling.

Of course there is always a chance, but running ~7500 trials (trial1) and hitting 49.99% is crazy. But let's just give that the benefit of the doubt. Weirder things have happened. Tests 4/5 (the ones that conflict with Lodeguy's data and are most relevant) hit what are essentially the most perfect value you can want given your ultimate prediction of dINT=15 is the cutoff, not dINT=10. The absolute perfect value for test 4 is T1Average + 18, or 68%. You got 68.6%. The absolute perfect value for test 5 is T4Average + 2.5, or 71.1%. You got 71.3%. I know the error bars are tight due to the large trial count, but the phrase "too perfect" comes to mind. We all know the "law of large numbers" tells us that eventually we will hit the true average with infinite trials, but it just sets off red flags for me.

I know I sound like a total asshole here because I'm essentially accusing you of faking data, but please understand why some of these things can bring up red flags. The near "too perfect" results coupled with just the insanity of conducting a trial of this scale... I don't know you outside of these boards and unlike a lot of people here I don't particularly dislike you, so please don't take huge personal offense to this. I would attempt to pick this to pieces if Rukenshin himself posted it. If you did a bot/parse I can conceivably see this, but manually I think this is almost inconceivable.

On a less objective note, calling people with legitimate questions idiots doesn't help your cause. Taruking's point on extrapolating data collected on normal mobs > HNMs is at the very least slightly valid. My answer to this would just be if you cannot make this "leap of faith" then the data is meaningless to you, but for most players, they generally believe that the MACC system is universal as a game mechanic and not mob specific. The same can be said about my enmity testing - everything on that is based on a LVL0 mandragora, yet people use this for HNMs.

It's not the first time that you try to contest my tests, and I don't care. On PUP/RDM or BLM/RNG my recast was about 40 sec. This allows you do to 500-1000 binds in one night assuming you don't get d/c. Also the numbers are far from being perfect, the test on hit rate < 50% shows a little inconsistance when INT=81, I find 14% for what should be 13%.

[S N K]

It's not the first time that you try to contest my tests, and I don't care.

And that is why a lot of people don't listen to you when you post shit like this. You want people to take you seriously then pull the cock out of your ass and address what people are saying or just don't fucking post anymore.

Jesus H Christ, I know not everyone from France is like you thank god seeing as Yuffy is a socialable person to talk to here but you have a serious problem with lack of social skills which is why people are hating on your ass so much.

Grow up already.

[Kaeko]

I contest out of wanting clarity. Not to make you look bad. If I wanted to make you look bad I wouldn't have worded it in such a nice fashion. Saying "I don't care what you think" to what I feel are valid questions is just hurting your supposed cause.

If you wish to ignore my questions that is certainly your choice; however, I (and I feel others) will look at anything you post with a certain caution. Again, you will probably say you don't give a shit. I have my own thoughts, you have yours. So if you don't wish to engage in some sort of meaningful discussion and hide your methodology I won't interfere in this thread. People are smart enough to make their own judgments.

[Taruking]

Stop your crap already. The reason some HnMs are very resistant to slow is because SE gave them an absurd amount of magic evasion, to the point that you are at the lower 5% cap. This has nothing to do with the "fact" that they have high mnd (proof ?).

You are making assumption again, how the hell do you know they are resistant to slow because of the way you just described? They might have added a slightly different set of formula for different HNM. For example, players may need a smaller INT gap in order to maintain 1 INT = 1 m.acc. The only way to find out is to run test on HNMs themselves.

[Thunder]

I know I sound like a total asshole here because I'm essentially accusing you of faking data, but please understand why some of these things can bring up red flags. The near "too perfect" results coupled with just the insanity of conducting a trial of this scale... I don't know you outside of these boards and unlike a lot of people here I don't particularly dislike you, so please don't take huge personal offense to this. I would attempt to pick this to pieces if Rukenshin himself posted it. If you did a bot/parse I can conceivably see this, but manually I think this is almost inconceivable.

How can you possibly respond to that with:

It's not the first time that you try to contest my tests, and I don't care.

Did you want a meaningful discussion or just a place to 'brag' about figuring something out?

[Klutz]

Annnnnnnnnnnnnnnnnnnnd the thread turns to shit, lol.