dex/crit relation, WS crits, WS gorgets discussion

[Valyana]

Lesser Colibri VIT from mode
Lv63 VIT52
Lv64 VIT52
Lv65 VIT55

    A  B  C  D  E  F  G
63. 36 31 28 24 21 17 14
64. 36 32 29 25 21 17 14
65. 37 32 29 25 22 18 14

63. 24 + 12 + 17 = 53
64. 25 + 12 + 17 = 54
65. 25 + 12 + 18 = 55

Again the numbers are slightly off, so there must be an error somewhere or a variable missing, but its a pretty close match.

Looks fine to me; since fSTR is a step function a range of mob VIT will all give the same mode. So it's not surprising that 53 and 54 would give the same mode as 52.

[Icywave]

Very nice thread.


I would sugget most of the critical hits test run with total 100+ DEX (that's one of the most obvious tier, if you are wondering)

A lot of infomations in this thread, thank you^^
(Gotta read it again and again now lol)


and Hi BG ^^

[Icywave]

BTW donno if this is answered in this 12 pages of de-rail and re-rail and laughing on KI


But Ranged ATK's critical Hit is based on DEX, not AGI. ^^

[VZX]

But Ranged ATK's critical Hit is based on DEX, not AGI. ^^

Was there any specific thread testing that?

[Nagamaki]

I went and spent some time starting to gather data for the Gorget acc test, sample data is still really small and I intend to add some control tests, but I'm posting to report initial results since this has been brought up in other threads. I was hoping someone else would go ahead and test this if I waited long enough, but looks like nobody else cares enough to do it themselves

Here's the stats for the Aura Statues at Ulli pop, based on /c

Lv81-84

Lv.81
Def 349
Eva 332

Lv.82
Def 354
Eva 337

Lv.83
Def 359
Eva 342

Lv.84
Def 365
Eva 348


Using the gobli formulas mentionned in a previous post, here's their stats calculated from known data

VIT (race C, job D, sj D)
81. 36 + 31 + 15 = 82
82. 36 + 31 + 15 = 82
83. 36 + 31 + 15 = 82
84. 37 + 32 + 16 = 85

AGI (race D, job C, sj C)
81. 31 + 36 + 18 = 85
82. 31 + 36 + 18 = 85
83. 31 + 36 + 18 = 85
84. 32 + 37 + 18 = 87

Def (race C, defense bonus)
81. 290 + 8 + 41 + 10 = 349
82. 295 + 8 + 41 + 10 = 354
83. 300 + 8 + 41 + 10 = 359
84. 305 + 8 + 42 + 10 = 365

Eva (job C)
81. 290 + 42 = 332
82. 295 + 42 = 337
83. 300 + 42 = 342
84. 305 + 43 = 348

Calculated def and eva matches the /c, so pretty much confirms the accuracy of the formulas.


Anyway, for the test itself, in order to maximize the useful sample while minimizing the amount of killing time required, I decided to fight only Lv84 Auras. If there's nobody up there its easy enough to aggro a bunch and go circle the teleporter to depop them and repeat until a few repop lv84. Knowing which are lv84 is just a matter of finding which ITs /c as neutral eva with acc lowered just enough that lv83s are still low evasion.

Gear/food was exactly the same as in my old test on the Enkidus, only big difference is I never used focus in order to keep my base acc well away from cap.

Regular melee acc across this whole batch of testing was 65.68%.

Initial sample size (I'll increase this, its still far too small) is 20 uses of Asuran Fists, ie 160 hits.

Using Soil Gorget, total result was 103/160, 64.375%

Using Faith Torque, total result was 114/160, 71.25%


Repeating one more time, for an acc test this sample size is way too small, need at least a few hundred hits per set and control tests with Peacock Charm and no neckpiece at all before taking any conslusions. These results are just the initial batch, I'll add more over a few sessions based on when I have free time (and when I can get the area without competition for mobs).

[Volken]

Using Soil Gorget, total result was 103/160, 64.375%

Using Faith Torque, total result was 114/160, 71.25%

Fuck.

[Blackice]

Very interesting Naga, can't wait to see the results from a bigger batch.

[Zirk]

Thanks for even more testing, I really wanted to see what the results are, all I can say is... COMMMMMMOOOOON SOIL GORGET.

[solsovly]

Thanks for even more testing, I really wanted to see what the results are, all I can say is... COMMMMMMOOOOON SOIL GORGET.

Seriously, I don't have time to try for a Temperance Torque

[heavensword]

Bummer...the nice research on Effective accuracy, along with all the colibri stuff got nuked.

Anyone care to re-post?

Also, tremendous thanks to Nagamaki and the many others taking time to research. Very cool, guys. Very cool.

[Retrieving/C.D/Searain]

This is all I have on my HD, I'll dig up the DEX/Crits thread on the italian forum later, there might be more stuff saved in there since I'm sure I've quoted Naga and Aurik at least a few times while posting on those boards.

Retrieving/C.D/Searain : Here's some data my friends and I have collected in the past few weeks of merit'ing in two of the most popular camps:

Nyzul Isle (Greater Colibri)

90  DEX -> 11,49% Crits;
94  DEX -> 11,22% Crits;
94  DEX -> 12,52% Crits;
95  DEX -> 12,83% Crits;
95  DEX -> 13,14% Crits;
96  DEX -> 12,62% Crits;
96  DEX -> 12,42% Crits;
100 DEX -> 13,28% Crits;
101 DEX -> 13,72% Crits;
109 DEX -> 15,86% Crits;
110 DEX -> 20,62% Crits; (w/ Senji)
111 DEX -> 20,89% Crits;
112 DEX -> 21,07% Crits;
115 DEX -> 21,47% Crits;
117 DEX -> 23,03% Crits;
118 DEX -> 24,04% Crits;
118 DEX -> 23.38% Crits;
118 DEX -> 23,31% Crits;
118 DEX -> 24,41% Crits;

MMJ SP (Jade Sepulcher - mixed mobs)

74  DEX -> 15,20% Crits; (w/ Destroyers)
77  DEX -> 15,69% Crits; (w/ Destroyers + Pahluwaan Body)
83  DEX -> 13,46% Crits; (w/ Senji)
92  DEX -> 11,81% Crits;
97  DEX ->  8,96% Crits; (only #2 Crits merit upgrades)
99  DEX -> 17,91% Crits; (w/ Senji)
100 DEX -> 10,18% Crits;
100 DEX -> 10,35% Crits;
105 DEX -> 14,15% Crits;
105 DEX -> 14,36% Crits;
113 DEX -> 18,40% Crits; (w/ Senji)
116 DEX -> 17,94% Crits; (w/ Senji)
120 DEX -> 17,93% Crits;
121 DEX -> 18,01% Crits; (Skoffins camp: 20 Puks, 17 Skoffins, 7 Lurkers, 16 Infiltrators, 20 Mamools, 6 Pets)
121 DEX -> 18,85% Crits;
121 DEX -> 18,56% Crits;
121 DEX -> 18,23% Crits;
121 DEX -> 18,37% Crits;
121 DEX -> 20,31% Crits; (Central camp: 53 Puks, 14 Skoffins, 1 Lurker, 40 Mamools, 9 Pets)
121 DEX -> 19,48% Crits;
138 DEX -> 23,54% Crits;

I combined our data with Bastilla's results, previously posted in the first pages of this thread since they seemed to match perfectly the Criticals % variation observed in our tests.

The Critical Hit percentages listed above are including the +4% from merits unless explicitly mentioned otherwise, haven't been sweetened using Mighty Strikes or Rogue Roll (of course lol) and the sample size was generally fairly large (1000-2000 hits) since each data corresponds to a different party/another PT member perticipating to the same exp session keeping a costant amount of Dexterity on him while TP'ing, for testing purposes.

I can provide screenshots of the parsers if needed, in case anyone is either skeptical or questioning the reliability of the source, or simply curious.

EDIT: 11/6 updated.

Nagamaki :

As for weaponskills, each "Critical damage" weaponskill (as stated in their descriptions) seems to operate fairly simply. The base crit chance is your crit rate on regular hits against that target with your gear, and fTP is the amount of % added to the base amount.

The following fTP numbers are approximations from memory for demonstration, you could find the true value relatively easy with some patience and testing.

Backhand Blow (mainly tested at 100%, higher is semi guess)
100% Critical Hit Rate +40%
200% Critical Hit Rate +60%
300% Critical Hit Rate +80%

Rampage (tested reasonably much)
100% Critical Hit Rate +0%
200% Critical Hit Rate +20%
300% Critical Hit Rate +40%

So if you're a MNK with a 9% crit rate, 100TP BB will have a crit rate of 49%, 200TP will be 69%.. If instead your crit rate is capped from DEX at 24%, it'll be 100TP BB will be crit 64%, 200TP will be 84%..

With Rampage at 100% crit rate is same as your melee hits, and it goes up by approx 1% per 5 TP. Limited Jin tests seemed to show a similar pattern, I would guess its probably similar across the other multihit crit WS.

[heavensword]

Regarding the Greater Colibri stats, specifically AGI, how would assuming colibri were BLU/THF affect the stats? THF having B-rated AGI and potentially an evasion bonus. To me it makes sense, because they reflect spells and steal your food.

If accurate, that means they aren't WAR/WAR. I was kinda expecting that since they can be Aspired. Calculating their racial ranks on VIT and AGI will require figuring out their job first. Their evasion rank also might vary with their job.

Evasion: (f(Lv, Job Evasion Rank) + AGI/2 + Job Traits

Lv81 Eva 336

Rank A, lv81 = 308 AGI 56-57
Rank B, lv81 = 298 AGI 76-77
Rank C, lv81 = 290 AGI 92-93
Rank D, lv81 = 281 AGI 110-111
Rank E, lv81 = 265 AGI 142-143

Anything but B or C seems pretty unlikely.

Greater Colibris have MP, so they are probably a job with natural MP.

RDM D, BLM E, WHM E, PLD C, DRK C, SMN E, BLU C

PLD would throw off the VIT calc by adding defense bonus (not to mention rank A VIT), and they seriously don't seem DRK type to me. BLU seems to make sense since they can cast things back at you.

So assuming Colibri are BLU, BLU are D in every base attribute (not 100% on that one, I can't test it myself since BLU is my only lv1)

If they were BLU/THF with an evasion bonus of 10, that would make their AGI:

(f(Lv, Job Evasion Rank) + AGI/2 + Job Traits

Rank C, lv81 = 290
Evasion = 336
336 = 290 + AGI/2 + 10
AGI = 72

...which seems pretty reasonable. Sure, /THF get Evasion Bonus II (+30), but PLD mobs only get a bonus of 10 def when the Def Bonus II is 30, don't they?

On the other hand, BLU/THF AGI should be:

Lv?? C + D + B*0.5
Lv81 36 + 31 + 40*.5 = 87
Lv82 36 + 31 + 40*.5 = 87

...and assuming BLU are rank D evasion, with /THF evasion bonus:

Rank D, lv81 = 281

336 = 281 + 87/2 + Trait
Trait = 12 (close)

I hope I'm doing those right; It would be really nice to gear up for decent Acc/Att + Crit without food for Nyzule Isle merits.

EDIT:

Colibri data from Wiki:

# Job: RDM/RDM
# Defense rank: D
# Evasion rank: D (+15 racial Evasion Bonus trait)
# Agility rank: B
# Vitality rank: E

Rank D, lv81 = 281
Evasion = 336
336 = 281 + AGI/2 + 15
AGI = 80

RDM/RDM AGI should be:

Lv?? C + B + B*0.5
Lv81 36 + 40 + 40*.5 = 96
Lv82 36 + 40 + 40*.5 = 96

Something isn't right here...

[LLK]

God I couldn't read all of that with out my head exploding.

So summry prease mehbee?

Are we all stacking DEX now?

Torque > Gorgets for WS?

[Mojo]

Gorgets would appear to be the best for single hit weaponskills while Torques are best for multi hit, at least that is what everything at this point seems to indicate.

[Tulun]

Faith beat Gorget for Asur fists? Was the acc stacked elsewhere so high to overcome the gorget acc bonus, or is it more gimped than we thought >.>

I'd be curious if this holds up to Blade Jin, Guillotine, Rampage, etc...

[VZX]

I guess it's good to compare the finding from Studio's Gobli too...
If you look at their wiki page, specifically, this page, at the end of the page, they have their conclusion worded like this (pardon my slight mistranslation, at least I'm sure I didn't get the math wrong)

Here's the extended conclusion
DEX = Opponent's AGI +50 in order to cap the critical hit rate

Monster's AGI 	Lv81 	Lv82 	Lv83 	DEX Crit Upper cap 	Lv81 	Lv82 	Lv83
THF&NIN Mamool   96 	  96 	  96 				               146 	 146     146
BLM Mamool 	   90 	  90 	  90 				               140 	 140 	 140
DRG Mamool 	   78 	  82 	  82 				               128 	 132 	 132
WHM%BLU Mamool   76 	  76 	  76 				               126 	 126     126
BST Mamooll      73 	  69 	  69 				               123 	 119 	 119
Giant Colibri    67 	  67 					                    117 	 117

[Sidra]

I know Studio Gobli has the total Evasion of all the XP mobs (or at least the data needed to calculate the Evasion for all of the XP mobs), as the numbers were confirmed and used by Nagamaki in his accuracy calculations. If someone can provide them to me as well as the level ranges the mobs spawn at, I will put together a chart of total accuracy needed to hit various accuracy percentages at the various camps. I am thinking of taking the most evasive mob at each camp (or in some caes maybe list 2, like Lurkers and then next highest mob), and then showing how much total accuracy (listed as a whole number, eg. 350) you'd need to reach 80,85,90, and 95% on that specific mob. People can then build as they'd see fit. It'd be a nifty little chart, I think.

[solsovly]

Faith beat Gorget for Asur fists? Was the acc stacked elsewhere so high to overcome the gorget acc bonus, or is it more gimped than we thought >.>

I'd be curious if this holds up to Blade Jin, Guillotine, Rampage, etc...

Nagamaki did some testing that strongly stuggested that gorget accuracy only effected the first hit of a WS. If that's the case, then the accuracy wouldn't be hard to overcome at all. There is some other factors to think around the first hit of multihit ws. But for now, I'm probably going to switch to Peacock Charm instead of soil gorget for rampage (no torque) and see how that works.

[Valyana]

RDM/RDM AGI should be:

Lv?? C + B + B*0.5
Lv81 36 + 40 + 40*.5 = 96
Lv82 36 + 40 + 40*.5 = 96

Something isn't right here...

Yes, it should be B/E/e; B racial AGI, and RDM has E AGI. So 40+27+13 = 80.

Studio Gobli seems to have a different theory, though; they have 67 agi for the colibri, which would be E/E/e.

Sidra: I gathered def/eva numbers for merit Mamool Ja at http://wiki.ffxiclopedia.org/Category_talk:Mamool_Ja

[Tulun]

Nagamaki did some testing that strongly stuggested that gorget accuracy only effected the first hit of a WS. If that's the case, then the accuracy wouldn't be hard to overcome at all. There is some other factors to think around the first hit of multihit ws. But for now, I'm probably going to switch to Peacock Charm instead of soil gorget for rampage (no torque) and see how that works.

Well, if that's true, that's pretty sad. Seems like most Gorgets would be basically useless then...

I guess the next question is: if the acc on the gorgets does only effect 1 of (4-8 hits, go go multi-hit power...), is it worth to use for the base modifier of .1? I know for rampage, it starts at base .5, Guillotine .875 (etc). Would it still be better to use a Sea gorget for damage improvement, or to go with your Sea torque?

I suppose this would be assuming that you are hitting a satifsying accuracy amount, and want to pump damage...