dex/crit relation, WS crits, WS gorgets discussion

[genome]

This is from half a year ago, if anybody needs more crit% vs DEX data:
http://genomeffxi.livejournal.com/11649.html
I had +30DEX ~= +10% crit on misareaux diatryma (lvl 50)

[Kaelan]

Gives your DRK's something to do besides Stunning or Krakening, too. Keep tapping AGI off the mob and see if it doesn't up everyone's crit rates. (Obviously, BLM debuff for good measure- that's a pretty huge swing while it's active). I'm a n00b for DRK- how do ABS-xxx spells stack?

If doable, ABS DEX + ABS AGI might well = critfest for DRK.

They're all independant of each other, you can have as many up as your mp allows.

The biggest problem I see with this is that absorb spells don't just give you a fixed bonus, it gives an initial +18 bonus, then slowly decreases by 1 point every few seconds until it wears off. Although you still might be able to get some decent data on wether or not a DEX vs AGI relationship exists with more tests using absorbs, thanks to the way they work it's unlikely you'll be able use them to get some hard values on exactly what that relationship is like

[Bagel]

Slight siderail: Does AGI increase ranged crit rate? I just did a relatively small sample which seemed to indicate that DEX did not, which means that if AGI doesn't do anything there's either no base stat to effect ranged crit rate or it's something really goofy like CHR.

[VZX]

Slight siderail: Does AGI increase ranged crit rate? I just did a relatively small sample which seemed to indicate that DEX did not, which means that if AGI doesn't do anything there's either no base stat to effect ranged crit rate or it's something really goofy like CHR.

Analogically, it should. But so far I haven't seen conclusive test regarding that.
I'd probably do it sometimes around late april or late may, together with rapid shot testing(stat-to-JA/JT relation)

[Nagamaki]

Here's some results on my damage tests comparing Soil Gorget and Faith Torque for Asuran Fists, and also comparing those numbers with the numbers I would get through direct use of the damage formula. Originally my goal was to test every possible effect (mainly the pdif and accuracy) but I quickly realized accuracy tests wouldn't work in that setting, my acc was just too high for any meaningful differences to be detectable. I tried using the same mobs while killing off my acc on purpose, but that made the fights slow and annoying, and getting good accuracy data takes a really big sample, which I was clearly not going to have the patience to get, so I stuck to the test with capped acc. I have some ideas for acc tests, it'll just take awhile. In other words, none of the results here say anything at all about what gorgets do for acc on multihit WS, which is obviously the key to their usefulness with them.

Targets

Enkidu

Lv.77 (Tough 160 Exp)
Defense 327
VIT 77
292 hth + 3 Acc = Low Evasion

Lv.78 (Tough 200 Exp)
Defense 332
VIT 77
292 hth + 8 Acc = Low Evasion

Lv.79 (very Tough 240 Exp)
Defense 339
VIT 80
292 hth + 14 Acc = Low Evasion

Lv.80 (Very Tough 280 Exp)
Defense 344
VIT 80
292 hth + 19 Acc = Low Evasion

VIT is calculated using the fSTR formula (fSTR itself is known because of the mode) and then substracting the delta strength from my STR. It is an approximation, but it probably is reasonably accurate.

Mode against 77-78 was 64.
Mode against 79-80 was 63.

Stats with TPing gear/food (mithkabobs)

STR78+21
DEX68+20
Accuracy +26 (+30 Focus)
Hand-to-Hand Skill +7

Attack 495

Stats with WS gear

STR78+34
DEX68+20
VIT75+19
Accuracy +23 (+30 Focus)

With Torque, Boost
Attack 561

With Gorget, Boost
Attack 552


Test data results

Most field meanings are pretty self-evident, "F High" means "Full High", the fields labelled F just use the data from WS that landed all 8 hits (mainly just to have a more meaningfull WS Low)

Soil Gorget

Lv77
Hits 95/96
Acc 98.958%
Total 12
High 1099
Avg 1020.416
Low 869
F Total 11
F High 1099
F Avg 1034.182
F Low 908
Avg/h 128.895

Lv78
Hits 198/208
Acc 95.192%
Total 26
High 1057
Avg 945.038
Low 786
F Total 17
F High 1057
F Avg 984
F Low 911
Avg/h 124.096

Lv79
Hits 129/136
Acc 94.853%
Total 17
High 992
Avg 841.059
Low 670
F Total 11
F High 992
F Avg 889.455
F Low 810
Avg/h 110.837

Lv80
Hits 121/128
Acc 94.531%
Total 16
High 960
Avg 789
Low 634
F Total 10
F High 960
F Avg 835.4
F Low 752
Avg/h 104.331

Faith Torque

Lv77
Hits 136/144
Acc 94.444%
Total 18
High 1100
Avg 963.056
Low 831
F Total 10
F High 1100
F Avg 1026.2
F Low 929
Avg/h 127.463

Lv78
Hits 74/80
Acc 92.5%
Total 10
High 1058
Avg 920
Low 674
F Total 6
F High 1058
F Avg 1000.167
F Low 923
Avg/h 124.324

Lv79
Hits 80/80
Acc 100%
Total 10
High 1011
Avg 892.3
Low 793
F Total 10
F High 1011
F Avg 892.3
F Low 793
Avg/h 111.538

Lv80
Hits 153/160
Acc 95.625%
Total 20
High 937
Avg 819.05
Low 690
F Total 13
F High 937
F Avg 861.538
F Low 757
Avg/h 107.065


Calculated averages

Some time after completing those I decided to try and see how those results would compare with the number I'd get through direct use of the damage formula.

D 58
fSTR 9
WSC 15
WD 82
1.1*WD 90

Using the VIT I calculated, with WS gear fSTR should have been 12 in all cases, so I went with that number.

MaxPdif/MinPdif are the pdif values calculated using wiki formula. MaxHit/MinHit are the values for normal (fTP1) hits, Max1st/Min1st are for fTP 1.1 1st hits, MaxTot/MinTot are totals (MaxHit*8 or MaxHit*7+Max1st)

Mid is the midrange value (MaxTot+MinTot/2), and MidHit is Mid/8, for the average per hit. pdif distribution within the bounds of Min/Max isn't perfectly even, but I still figured it would give a rough idea.

Lv.77
Defense 327

Torque
Ratio 561 / 327 = 1.715
cRatio 1.715 - 0.1 = 1.615
MaxPdif 1.938
MinPdif 1.138
MaxHit 158.916
MinHit 93.316
MaxTot 1271
MinTot 746
Mid 1008.5
MidHit 126.062

Gorget
Ratio 552 / 327 = 1.688
cRatio 1.688 - 0.1 = 1.588
MaxPdif 1.906
MinPdif 1.105
MaxHit 156.292
MinHit 90.61
Max1st 171.54
Min1st 99.45
MaxTot 1265
MinTot 733
Mid 999
MidHit 124.875


Lv.78
Defense 332

Torque
Ratio 561 / 332 = 1.689
cRatio 1.689 - 0.15 = 1.539
MaxPdif 1.847
MinPdif 1.0468
MaxHit 151.454
MinHit 85.837
MaxTot 1211
MinTot 686
Mid 948.5
MidHit 118.562

Gorget
Ratio 552 / 332 = 1.662
cRatio 1.662 - 0.15 = 1.512
MaxPdif 1.814
MinPdif 1.014
MaxHit 148.748
MinHit 83.148
Max1st 163.26
Min1st 91.26
MaxTot 1204
MinTot 673
Mid 938.5
MidHit 117.312


Lv.79
Defense 339

Torque
Ratio 561 / 339 = 1.654
cRatio 1.654 - 0.2 = 1.454
MaxPdif 1.745
MinPdif 1
MaxHit 143.09
MinHit 82
MaxTot 1144
MinTot 656
Mid 900
MidHit 112.5

Gorget
Ratio 552 / 339 = 1.628
cRatio 1.628 - 0.2 = 1.428
MaxPdif 1.714
MinPdif 1
MaxHit 140.548
MinHit 82
Max1st 154.26
Min1st 90
MaxTot 1138
MinTot 664
Mid 901
MidHit 112.625


VT80
Defense 344

Torque
Ratio 561 / 344 = 1.630
cRatio 1.630 - 0.25 = 1.380
MaxPdif 1.656
MinPdif 1
MaxHit 135.792
MinHit 82
MaxTot 1086
MinTot 656
Mid 871
MidHit 108.875

Gorget
Ratio 552 / 344 = 1.604
cRatio 1.604 - 0.25 = 1.345
MaxPdif 1.614
MinPdif 1
MaxHit 132.348
MinHit 82
Max1st 145.26
Min1st 90
MaxTot 1071
MinTot 664
Mid 867.5
MidHit 108.437



Overall, the numbers aren't quite exact matches with my test results, but they're pretty damn close. If gorgets had some effect on pdif, I believe the numbers with them should have been fairly significantly higher.

If anyone notices any mistakes in there please let me know, I applied everything properly as far as I know.



edit: I forgot to mention that I used Focus before each pull (forced myself to wait for timer between mobs) and no WS was used without Focus being on. Also, I didn't calculate any WS I did after a golem used Crystal Shield (which was actually pretty rare, they usually died before getting a 2nd WS off and they have a fairly big list to pick from). AF was always used as close to 100% as possible unless mob HP was too low to take all 8 hits.

late edit: fixed the VIT data in there, calculated it wrong originally and I forgot to update when I re-calculated. It doesn't change the value of the damage calcs themselves, just changes the VIT value for the mobs

[Dirian]

Nothing meaningful to add, just want to give massive kudos to Nagamaki for all the tests he's done in this thread.


Carry on.

[ringthree]

69 DEX (Elv MNK/THF base)
9.3%

89 DEX (Byakko's, Rajas)
21.2%

105 DEX (Byakko's, Rajas, Osode, Squid)
23.9%

100 DEX (Byakko's, Rajas, Thunder, Squid)
23.7%

This is against the TW Lufaise mobs I used for early cap tests.

It's pretty clear that crit rate isn't DEX/10+crit bonus items/merit, or something like those rates would never happen.

Neither would the rate hikes observed on around 120 DEX.

What needs to be defined is the exact DEX vs mob stat relation, but it is 100% clear that crit rate is not some flat value.


About ranged crits, I don't doubt that rogue's roll, and most likely crit merits affect ranged crit rate, its just clear that Thunder Staff doesn't. I only tested that to see if if Sniper Shot would be a good test, but Backhand Blow turned out to be a good test subject for WS crits.

Finally, I decided to do practical Gorget vs Torque tests with AF, using Enkidus (cause they're stupid easy, reasonable solo xp and drop Diorite). I want to get a really good size data sample so I won't be posting results for awhile more, but what I got so far is rather extremely underwhelming when it comes to the gorgets.

Just concerning crits, when I first saw this my first question was that, if this test was against TW mobs, then your DEX should dominate over their AGI (assuming of course this is the calculation) but the numbers appear to be able to be replicated on XP mobs, or at least reasonably close, where their AGI should be much higher. I guess what I am asking is this: Is the difference between a lvl 50 TW mob and a level 75 XP mob's AGI not that great, or is there some kind of prior mentioned tier of DEX that just in general increases crits?

[VZX]

umm... I'm a bit confused with your naming scheme, Nagamaki.
Probably you could explain a bit why you put LAvg, LHAvg, MinH, MaxH

[Nagamaki]

I should have made it LAvgH for consistency, but H stands for hit, L is for Linear (cause I calculate it by doing Min+Max/2). Pdif distribution between min/max isn't perfectly linear, at least not between .5 and 1.5 cRatio because of the mode, that's why I added the L (to indicate its not necessarily the actual average).

A more appropriate term would have been Mid and MidH for midrange, just didn't think of it when I calculated those.

I explained all the labels I used in the post though, so whole the meaning might have been confusing, what the value represents was clearly stated.

I'll go back and edit the post for more appropriate labels.

[Retrieving/C.D/Searain]

69 DEX (Elv MNK/THF base)
9.3%

89 DEX (Byakko's, Rajas)
21.2%

105 DEX (Byakko's, Rajas, Osode, Squid)
23.9%

100 DEX (Byakko's, Rajas, Thunder, Squid)
23.7%

This is against the TW Lufaise mobs I used for early cap tests.

It's pretty clear that crit rate isn't DEX/10+crit bonus items/merit, or something like those rates would never happen.

Neither would the rate hikes observed on around 120 DEX.

What needs to be defined is the exact DEX vs mob stat relation, but it is 100% clear that crit rate is not some flat value.

Just concerning crits, when I first saw this my first question was that, if this test was against TW mobs, then your DEX should dominate over their AGI (assuming of course this is the calculation) but the numbers appear to be able to be replicated on XP mobs, or at least reasonably close, where their AGI should be much higher. I guess what I am asking is this: Is the difference between a lvl 50 TW mob and a level 75 XP mob's AGI not that great, or is there some kind of prior mentioned tier of DEX that just in general increases crits?

@ Jade Sepulcher, with +4 Crits from merits:

A. (add +6 from Sole Sushi +1) 100 DEX -> 10% Crits (rather common War/Nin TP gear)

B. 121 DEX -> 15-16% Crits

C. 138 DEX -> 22-23% Crits

http://img168.imageshack.us/img168/1...itsgearib6.jpg

...and Vorpal Blade average damage went up accordingly to that chart (the Axe was just there randomly while taking the SS, never really used it), however my Crits rate varied greatly depending on the mob; it seemed reaching some sort of cap on Mamool Ja Wyverns (Raptors seemed to have much higher AGI), Sea Puks and Blusterer/Philosophers where I could score 8-9 Crits in a row giving a sort of "Mighty Strikes" feeling with just 121 total DEX whereas I had relatively hard time crit'ting with some regularity Mamool Ja Thieves even with 69+69.

This is a relatively old test though and I wasn't the person parsing so I couldn't show you the actual numbers but feel free to replicate the same gear set-ups I used and go test them by yourself, our samples size was fairly large so I doubt other people's results could be that different, I hope it helps to figure out how much Dexterity is needed for that particular camp if you're planning on using a DEX heavy build.

Also notice how the B. set-up would be perfectly viable for a War/Nin since the AF2 mask could have been swapped out for Turban if I had full DEX merits, leaving basically untouched my classic Haste+ gearset; my Accuracy with that gear and Sole Sushi+1 was around 90%, Flashga included.

If you're curious to know how similar amount of DEX affect your Crits hit rate on weaker mobs just check this thread:

http://ffxi.allakhazam.com/db/jobs.h...;num=37;page=1

Bye.

[suji]

If you're curious to know how similar amount of DEX affect your Crits hit rate on weaker mobs just check this thread:

http://ffxi.allakhazam.com/db/jobs.h...;num=37;page=1

I mentioned this above but apparently no one saw it. The info is pretty interesting and currently suggests a DEX-born critical hit bonus cap of 20%, though that's a matter of interpretation given the current margins of error.

[Zirk]

If you're curious to know how similar amount of DEX affect your Crits hit rate on weaker mobs just check this thread:

http://ffxi.allakhazam.com/db/jobs.h...;num=37;page=1

I mentioned this above but apparently no one saw it. The info is pretty interesting and currently suggests a DEX-born critical hit bonus cap of 20%, though that's a matter of interpretation given the current margins of error.

Nagamaki showed earlier in this thread with his own tests that the base crit cap from stacked dex is around 20% as well.

[suji]

Nagamaki showed earlier in this thread with his own tests that the base crit cap from stacked dex is around 20% as well.

Yeah, but the same size was much smaller. I'm psyched to see the numbers line up, though I'd still say it's a theory and not a fact. The alla test still needs a lot of smoothing in the high sample ranges. Margin of error is still pretty big even with 1000 samples per DEX.

Also, it doesn't really look like there are any 'tiers' like people have been saying here, other than the final plateau which I guess you could call a tier.

[Terrance]

Those are with no +crit items. That has me wondering if +crit items act weirdly then because last time I tried crabs back when NPC buddies came out I was getting way WAY above 30.5% on crabs. Xarc demons weren't nearly as crit happy, but the only dex I had was byakko and warwolf.

[VZX]

Nice, I can read it better now.

I guess Studio Gobli's pDIF min/max function is still not good enough to produce tight bound on non-capped cRatio.
And yeah... I don't notice any pDIF bonus from there. Maybe I missed something too. gotta read it again later some time tonight

[RKenshin]

Nothing meaningful to add, just want to give massive kudos to Nagamaki for all the tests he's done in this thread.


Carry on.

Seconded, really great job on the tests Naga. I regret I couldn't be much of a help in providing more testing in this thread but time hasn't been on my side, I still hope to do a little more in the future if at all possible.

On the topic of gorgets though, is this all official then? Based on the plethora of test results Naga provided in this thread, it would seem the sea gorgets do almost nothing worthwhile for multi-hit weapon skills. Albeit the difference is incredibly small in the long run, but the difference is still there... And evidence is in favor of it being a mediocre neck choice at best for multi-hits, at least if I'm understanding it all correctly.

Also on the DEX topic, I'm a little unclear with a few specifics of Sea's test.

@ Jade Sepulcher, with +4 Crits from merits:

A. (add +6 from Sole Sushi +1) 100 DEX -> 10% Crits (rather common War/Nin TP gear)

B. 121 DEX -> 15-16% Crits

C. 138 DEX -> 22-23% Critsp

First thing being I've never PTed in or even been to Jade Sepulcher, and I really have no idea what the mobs are like there. From reading your post I'm assuming it's mostly puks and mamool, but is there anything different about them or are they basically just carbon copies of the ones in Thickets? For example, puks in Thickets and puks in Mamook aren't exactly the same (puks in Mamool have a type of dispelga wind attack; Ill wind I believe).

I'm going to assume you ended up fighting a decent amount of puks, and if that's true then things actually line up pretty well. In my test a few pages earlier with 121 DEX my critical hit-rate came out to be almost the same as yours, except about 1-2% lower. I'm pretty sure this is mostly because of the fact I was at the north camp and fighting mamool/skoffins over 90% of the time. In editting out the fights with mamool/skoffins in that parse my critical hit-rate was over 16% (just counting fights with puks), so the mamool definitely bring it down.

In addition, I'm also curious about this line since it was very vague:

...and Vorpal Blade average damage went up accordingly to that chart

I realize that you don't have any actual parsed numbers (I think?) but exactly what kind of an improvement did you recieve? And if possible what was the WS average around, and what is it normally around? It's also a little unclear what chart you're referring to (I have no idea). I'm curious how old the parse is if you remember as well, it doesn't look too old based on the gear I see.

As far as the viability of that set of gear in a practical situation, did you really feel wearing all the DEX gear in equipment build B is worth the 2-5% increase in critical hit-rate? TPing in basically the same gear set you have I see most WARs parse at around 10-13% critical hit-rate. And of course this is all assuming the whole theorized critical tier can't be achieved with just a few more points of DEX on top of 121.

My results on SAM were also far from being in favor of the DEX build, although it was only one parse so it's hard to say for sure. I love hitting critical hits as much as the next guy but, my overall damage from the loss in attack/acc/haste suffered more than I was comfortable with for a small increase in the critical hit-rate.

[Sylvrdragon]

First thing being I've never PTed in or even been to Jade Sepulcher, and I really have no idea what the mobs are like there. From reading your post I'm assuming it's mostly puks and mamool, but is there anything different about them or are they basically just carbon copies of the ones in Thickets? For example, puks in Thickets and puks in Mamook aren't exactly the same (puks in Mamool have a type of dispelga wind attack; Ill wind I believe).

I don't think he meant IN Jade Sepulcher. More likely he meant the Bhaflau Thickets camp at the zone-line TO Jade Sepulcher (The back of the Mamool Ja Staging Point camp, past the Skoffins or whatever the fuck they're called these days)

[Sidanoir]

By Jade Sepulcher he means the camp just outside the entrance to Jade Sepulcher from the Mamool Ja Staging Point...

Its imho the most demanding camp for melees in thickets, unless you want to fight skoffin only...

2 Infiltrator, Lurker are usually fought during chain....there is puks, a Pikeman, can pull a stabler or skoffin sometimes...and Blusterer + the other mage one can't think of name atm drawing a blank...

its a great camp imho, my fav if colibri isn't open

Honestly from naga's test correct me if I am wrong but...he was testing gorget vs Even Match to low end Very Tough Mob and a WS that is known for its horrendous modifiers? Asuran Fists never seemed to fluctuate much to me....probably one of the most consistent WS and it always seemed hard to raise its damage much....

Would not a better test of a gorget's usefulness for mult-hit be with a lower accuracy build for 1 to see if it really does help accuracy?
And then a decent accuracy set to see what it does for damage? I would think a test vs IT mobs would be more valuable or high end VT...to determine what is best in merit situation....

Honestly also it might be worthwhile to test with charm? And then maybe a Chiv. Chain type neck item?

I thought people had always felt Torque > Gorget for WS....

But I would hope Gorget > Charm?

Anyways, its cool to see all this data and stuff but for me it opens up more questions than answers! I only wish I had time to do some testing myself.....instead of just "feel"/ eyeball stuff and relying on others

[Nagamaki]

Test was on 77-80 mobs, so mid T to mid VT. Greater Colibris are 81-82.

I did make it pretty clear there were no real conclusions about accuracy, since my acc was already capped on those mobs just from basic gear and focus.

Asuran has weak modifiers, and a low modifier means the effect of the fTP on first hit is lessened (since the based damage is lower). I was actually hoping to get some accuracy results before I started, but realized I was grossly overestimating their evasion. I actually tested a bit without using focus and with a Behemoth Ring, and my acc was still barely dipping below 90%.

The flaw with dropping acc like mad for tests is that the mobs actually become a pain to fight solo, and then doing enough fights to get a big enough sample on accuracy becomes a huge chore. AF is 8 hits per WS, but even Enkidus have a range of 4 levels, with 2 separate tests (torque and gorget).. That means a gigantic amount of mobs required in order to even reach a tiny sample size of around 300 hits, and while they are easy since their delay is long enough to get Ichi off without any debuff, it does get annoying pretty quick.

Testing against IT would be great, but then you need to find people (at least a healer) to kill with willing to stick to it long enough. Fighting those things as MNK is relatively fun, but being the haste/healer probably isn't exactly a blast of entertainment. Ruling out mobs with eva buffs or bad WS, probably the best test subject would be Ulli room Aura Statues, kind of a logical next step after Enks. Any RDM volunteers for going up there on Seraph? They're slow enough to solo as MNK without bad luck, but not at a productive pace.

Also as a sidenote, SE needs to expand merits again so bad, I'm down to spending merits on ninjutsu effects and san spells, wtf... Sure I'll get a 5th job category when I finish BST, but then what the hell am I gonna do with BST merits.

[Akroma_Seraph]

@ Naga: If You and I share some free time from our various LS/Social events I would be happy to join you up in sky for your testing. I personally enjoy reading every one of your tests so anything I could do to help, just ask :D