Shield Asstery

[Wrish]

6.49% parry @ 95 AGI certainly doesn't look notably high on an EM steelshell... definitely not a recognized cap to me. For AGI testing I'd expect more of a control with respect to skill & other stats, plus a bigger sample size.

Practically, rather than all-out spamming +40 agility, adding/levelling parry skill might help more if you just want to proc parries. But for PLD tanking you'd definitely want to squeeze better stats out of those slots - shield skill, haste, VIT/defense.

With 292-296 parry and agility of 60-65, I took over 2100 hits exclusively from even match crabs and parsed 14.0% parry. I have no data with a high AGI build @ similar parry skill because I haven't bothered with one... yet. Can't find a practical application because this same skill level of parry procs less than 8% on merit-level mobs.

Realize that with extremely low parry skill you'll still parry at a 4-5% base rate. It is troubling to gauge parry improvements when you start at that baseline... you don't know how much of the "improvement" is sapped by starting below the baseline. E.g., moving from 15 to 150 in any combat skill would do zilch for an EM crab, but 150 to 250 skill is a different story.

[willriker]

He was at either 220 or 230 parry in testing. PLD cap Parry rating is 210. so at very least, its somewhat of an aproximation of what a PLD would see for pary results if he/she stacked AGI. Do I think that its worth it for a PLD to stack these skills? Not really, there are very few fights where PLD would see decent results form this stat selection.

Another thing i am satified with, is that Brutala confirmed that as an enemies level decreases you will see more shield block activations while keeping the shield skill constant. This was only conjecture before, unless i missed a test. The only other test i have seen on this determined that it takes 300 shield skill to reach the activation rate cap on EM-DC enemies with a size 3 shield. That tester left it to conjecture that lower level enemies would require less shield skill to reach this cap. By the same rational, higher level enemies would require more shield skill to reach the same cap.

Brutala, thank you for clearing up that issue. Most of us knew this to be true, but noone really went out to see how much of a diferance it actualy is. While this test did not provide a defined amount of shield skill to reach the % cap per enemy level, it did verify the general trend that was just a geustimation before.

Though that bit in your post about the day of the week/moon phase having an effect on shield block activation rates is kind of interesting. Thats a whole differant issue though... and while i agree that something weird might be giong on with this, it would only be proven to be true with a sample size large enough to eliminate the enherent dificulty that arizes when dealing with programmed random occurances.

[Mojo]

Evasion procs before Parrying. Those numbers do not represent the 'true' growth of parrying because of this.

[Masamune]

He was at either 220 or 230 parry in testing. PLD cap Parry rating is 210. so at very least, its somewhat of an aproximation of what a PLD would see for pary results if he/she stacked AGI. Do I think that its worth it for a PLD to stack these skills? Not really, there are very few fights where PLD would see decent results form this stat selection.

Another thing i am satified with, is that Brutala confirmed that as an enemies level decreases you will see more shield block activations while keeping the shield skill constant. This was only conjecture before, unless i missed a test. The only other test i have seen on this determined that it takes 300 shield skill to reach the activation rate cap on EM-DC enemies with a size 3 shield. That tester left it to conjecture that lower level enemies would require less shield skill to reach this cap. By the same rational, higher level enemies would require more shield skill to reach the same cap.

Brutala, thank you for clearing up that issue. Most of us knew this to be true, but noone really went out to see how much of a diferance it actualy is. While this test did not provide a defined amount of shield skill to reach the % cap per enemy level, it did verify the general trend that was just a geustimation before.

Though that bit in your post about the day of the week/moon phase having an effect on shield block activation rates is kind of interesting. Thats a whole differant issue though... and while i agree that something weird might be giong on with this, it would only be proven to be true with a sample size large enough to eliminate the enherent dificulty that arizes when dealing with programmed random occurances.

I second this too, great job Brut !

[Byrthnoth]

First off, I don't think it's smart to combine the two Parrying samples at the same AGI values, because he did one set at 220 Parrying Skill and the other set at 230. They aren't statistically different, but that raises the question of why Parrying Skill isn't affecting the results. If you look at them as four populations, corrected for Evasion:

(1) 55AGI, 230 Skill: n=436; 4.23+/-1.87% (89STR 62DEX 92VIT 55AGI 52INT 72CHR 263 Shield 230 Parrying)
(2) 55AGI, 220 Skill: n=324; 5.56+/-2.49% (94STR 82DEX 78VIT 55AGI 52INT 62CHR 220 Parrying)

(3) 95AGI, 230 Skill: n=339; 8.85+/-3.02% (71STR 63DEX 88VIT 95AGI 57INT 72CHR 230 Parrying)
(4) 95AGI, 220 Skill: n=370; 6.76+/-2.56% (76STR 68DEX 88VIT 95AGI 57INT 62CHR 220 Parrying)

Also, it's not cool to pretend that these were the only stats that varied. As Agility increased in the gearsets:
* STR dropped
* INT increased

Everything but MND changed randomly, but STR and INT were the only things to change consistently with AGI.

[willriker]

You dont think that +40 AGI (~+70%) would out weigh a +10 differential from 220 parry skill (~+5%) enough that a general trend can be seen? I did admit that it wouldnt be exact, the numbers could not be used to produce any sort of equation for exact results. But, i would think that it is good enough to see a trend where AGI effects things. It wouldnt be exact, just a general behavior pattern.

I think so.

Oh, somehow i doubt STR or INT would be modifiers for Parry. I could be wrong, but im willin to put money on it. And im a cheep sone of a gun. LOL

[Mojo]

There are huge flaws in how those tests were done in accordance to AGI, parrying, and evasion. The sample size is also incredibly low for the occurrence rate that is being measured. 10 skill can account for a lot depending on where your at in regards to evasion, it's likely that the same would occur for parrying. You can't conclusively say anything with much confidence based on this data set.

[Brutala]

Parry Differances:
55 AGI testing 18/457 + 18/342 = 36/799 = .0451 = 4.51% parry rate
61 AGI testing 31/522 = .0594 = 5.94% parry rate
95 AGI testing 30/424 + 25/423 = 55/847 = .0649 = 6.49% parry rate

Evasion Differances:
55 AGI testing 21/475 + 19/342 = 40/799 = .0501 = 5.01% evasion rate
61 AGI testing 42/522 = .0805 = 8.05% evasion rate
95 AGI testing 85/424 + 53/423 = 138/847 = .1629 = 16.29% evasion rate

No. For example, the 457 in the first sample is only the total blocked and unblocked. 18 parries are outside of that making the total 18/(457+18 ) if you want parrying rate. Then if you want evasion rate in that sample you'd have 21/(457+18+21). That's assuming intimidation procs before eva, I dunno \oO/. If it by some freak chance doesn't, then you'd have 21/(457+18+21+27). Rinse and repeat for other samples. Sorry I didn't make it clear what I meant by totals. Also, low sample size etc. I'm posting this not because I think I found something but because I quit and wanted to contribute to the sample sizes of any future testers and warn them to pay attention to moon or w/e else I'm missing.

If you're insane enough to care about eva and parry in such a small sample this is how the rates should have been calculated (and setups shouldn't have been lumped together just for having the same total agi):

55AGI, 230 parry, 235 eva skill, -10 eva in gear
18/(457+18 ) 0.0379 parrying rate
21/(457+18+21) 0.0423 evade rate

55AGI, 220 parry, 225 eva skill, -10 eva in gear
18/(342+18 ) 0.0500 parrying rate
18/(342+18+18 ) 0.0476 evade rate

95AGI, 230 parry, 235 eva skill, +5 eva in gear
30/(424+30) 0.0661 parrying rate
85/(424+30+85) 0.1577 evade rate

95AGI, 220 parry, 225 eva skill, +10 eva in gear
25/(423+25) 0.0558 parrying rate
53/(423+25+53) 0.1058 evade rate

What sucks is that based on this I can't say AGI had an effect on Anything (though for the same parrying skill a 40 AGI boost did give a minimal increase in parries, but it's negligable given the sample sizes) since the raw evasion stat difference alone can account for the differences in evade rates. I'll edit out that AGI statement. In any case, it did nothing for shield which was the point of the 4 tests.

[willriker]

oof haha thanks man.

[Byrthnoth]

10 Skill doesn't show a significant effect comparing the samples we have here. That should be either a testament to how floored the Parrying rate is or how small the sample size is.

I agree it'd be stupid for Parrying rate to be inversely proportional to Strength. I also think it's stupid that +40 AGI resulted in a ~3% increase in Parrying rate though. +5 INT? >.> Mage stats lol. OR MAYBE THAT'S WHY (*insert piece of melee AF with INT+ on it here*) has INT~!!! (jkjk)

If we were really interested in Parrying rate, we'd repeat the test with a SAM/DNC with capped Parrying jumping only their AGI instead of a WAR. Higher Parrying skill, 2H weapon Parrying boost.

I'd be hesitant to conclude much of anything from these tests except that AGI doesn't seem to affect Shield proc rate. Actually on that topic, did anyone do a correction for the Evaded/Parried attacks when calculating Shield proc rate?


Edit: orz, picked a bad time to post. Well, disregard my last post then.

[Masamune]

There are huge flaws in how those tests were done in accordance to AGI, parrying, and evasion. The sample size is also incredibly low for the occurrence rate that is being measured. 10 skill can account for a lot depending on where your at in regards to evasion, it's likely that the same would occur for parrying. You can't conclusively say anything with much confidence based on this data set.

^ This. Brutala did a great job ..... for actually initiating those tests, not for giving results leading to actual conclusions, baring maybe the AGI vs. shield procrate, but that's it.

If you want to conclude anything, you need either complementary tests and/or continue Brut's tests to increase their sample sizes.

[Niiro]

2H weapon Parrying boost.

Do we even know if such a thing exists?

I don't think I've ever seen anything to show or even imply that 2h weapons parry more than 1h ones.

[Byrthnoth]

I dunno. It might come from the same time as the AGI = Parrying/Shield rate+ lore.

[deathdealer]

lol, Reizvoll is giving PLD advice.

It could have been worse... it could have been you giving PLD advice...