Dual-Wielding KC and Ridill(or M.kris)

[Xanthu]

Also, i thought Ridill distribution is 30:50:20, not 20:50:30 ?

Mistake on the write-up, these are the numbers I went with, just got flipped when typing.

Math with an assumed 30:60:10 on M.kris:

0:1.5:7.5:17:24:22:14.5:13.5

For a new average of 5.55

Assuming you have a Suppa:
In order for the KC/M.Kris to be better it must have an average of 5.57 swings, or 46% better. In order for the Ridill/KC to be better it must have an average of 6.11, or 60%, better.

(If you have a Nuevo-augmented with Dualwield it becomes 5.04 swings for kris-combo, 32%/5.57 swings, 45% for Ridill)

So now it's basically only worth it if you get a nuevo AND a suppa, but I'm not seeing that really worthwhile.

[Masamune]

Mistake on the write-up, these are the numbers I went with, just got filpped.

I'll re-do the math with an assumed 30:60:10 on M.kris

You shouldnot.

Saying this because i'm really wondering how Ridill, KC and Bahamut Zaghnal have been tested, what were the methodology ?

Also, i'm not sure at all if the slow18% i used to parse is enough to avoid KParser from mixing rounds between them, with a delay 196 dagger. Motenten can maybe answer that question tough?

[CDF]

Without a parser, the only methodology is to count manually the number of attacks per round, as was done with Ridill, but you would have to ask Searain if he didn't use any haste equipment. (~4 seconds between rounds seems enough not to observe overlap.)

As for Kraken Club and the others, I've never seen any count data to support any of the claims I've seen about them.

[Xanthu]

You shouldnot.

Too late!

Saying this because i'm really wondering how Ridill, KC and Bahamut Zaghnal have been tested, what were the methodology ?

Also, i'm not sure at all if the slow18% i used to parse is enough to avoid KParser from mixing rounds between them, with a delay 196 dagger. Motenten can maybe answer that question tough?

I'm running with what I have right now. Something I saw proposed to get KC percentages was to find the relic-weapon of choice and go KC/relic. Adding the 999 delay means you have a lot more time to count the KC/Mkris rounds, then 2-4 rounds becomes 1-3. The only flaw here is no way to find out the 7/8 exact percentage, though it may not matter

[Masamune]

Something I saw proposed to get KC percentages was to find the relic-weapon of choice and go KC/relic. Adding the 999 delay means you have a lot more time to count the KC/Mkris rounds, then 2-4 rounds becomes 1-3. The only flaw here is no way to find out the 7/8 exact percentage, though it may not matter

Good idea, but i can't do it for MKrys since i don't own a relic 1H weapon ;;

[Masamune]

Without a parser, the only methodology is to count manually the number of attacks per round, as was done with Ridill, but you would have to ask Searain if he didn't use any haste equipment. (~4 seconds between rounds seems enough not to observe overlap.)

Indeed, with my slow18% for example, i get 3.76secs between rounds, continuing same parse but added heca feet for an additional slow4% -> 3.89s, hope that's still enough. Just from eyeballing, a triple attacks round ends very closely to next round starting hit :S

As for Kraken Club and the others, I've never seen any count data to support any of the claims I've seen about them.

That's exactly my wondering... so i'm like wtf at all those pseudo calculators like xanthu thinking can use such pseudo results out of nowhere... Don't take offense from that comment, but even myself with my current calculator build, i'm not at all confident with results obtained to make a release, hence all the recent testings on very basic stuff like crit%, crit% WSs, monster stats, or multihit weapons...

[Xanthu]

That's exactly my wondering... so i'm like wtf at all those pseudo calculators like xanthu thinking can use such pseudo results out of nowhere... Don't take offense from that comment, but even myself with my current calculator build, i'm not at all confident with results obtained to make a release, hence all the recent testings on very basic stuff like crit%, crit% WSs, monster stats, or multihit weapons...

Naw it's cool, I know that this is something that's really an unfounded set, it's just the only thing out there to be seen so far. It does seem to match the sorta 30-45% area of the multi-hit-dial.

I think regardless, with rough numbers (yes we still need a better KC set) it shows that the dual-wield isn't reeeeally worth it.

[Masamune]

I think regardless, with rough numbers (yes we still need a better KC set) it shows that the dual-wield isn't reeeeally worth it.

Formulated like that, ok.

[Byrthnoth]

If you have the money and are on DRK/NIN already due to AoEs/what-have-you, then it seems it might be worth it. Of course, this depends most strongly on the K-club distribution. 7% increase in attacks wouldn't be anything to sneeze at!

That said, two DRKs (one with a Merc Kris and one with a K-club) would do more damage than one DRK dual-wielding them.

[Streak]

So what's with the procs on Mercurial Sword then?

Occasional attacks 2 to 3 times... yet it also does 4 swings. I'm not sure all the distributions with this mod can be assumed to be equal with that wrench thrown in.

[Byrthnoth]

Yeah, supposedly Mercurial Pole attacks one more time than it's supposed to as well.

[Rufox]

So, What about Kraken Club / Octave Club as DRK 72 for zerg?

[Xanthu]

So, What about Kraken Club / Octave Club as DRK 72 for zerg?

Sure why not.

KC Delay 264
Dual-wield delays Base:448.8 w/Suppa:422.4 w/Suppa and Nuevo:406.56

Average number to beat is respectively 6.482424... 6.1062121... 5.8828

The swing numbers breaks down to:

0*:0.25:1.5:4.75:10:15.25:18:50.25**

**8 swings breaks down into 17.05:13.6:9.35:5.5:2.9:1.2:0.49:0.12:0.04 (lots of extra unused swings)


Average becomes 6.935

KC/Octave is the winner without even a suppa.

Once again this doesn't account for loss of power from /drg, from actually having the last 3 levels of DRK (if you had a KC why wouldn't you) or anything else. Note a 72 DRK cannot use the Januwiyah, so no 30 HP there. There's the Terror Shield with 20 HP (million+ gil though, PLDs want it) and a couple of options for an HP+10

[Byrthnoth]

So wait, now I'm kind of confused. You just showed that Main: KC, Sub: Octave Club is better than KC at 75?

1 Attack from main hand -- 5:15:25:25:15:10:5 = 4.8 Attacks per round
2 Attacks from main hand -- 5:15:25:25:15:15 = 5.75 Attacks per round
3 Attacks from main hand -- 5:15:25:25:30 = 6.6 Attacks per round
4 Attacks from main hand -- 5:15:25:55 = 7.3 Attacks per round
5 Attacks from main hand -- 5:15:80 = 7.75 Attacks per round.
6 Attacks from main hand -- 5:95 = 7.95 Attacks per round.
7 Attacks from main hand -- 100 = 8 Attacks per round.
8 Attacks from main hand -- 0 = 8 Attacks per round.

6.935 Attacks per round.
(264+264)*.8/60 = 7.03 seconds per attack round pre-Haste
.986 Attacks per second

That's 6% more than Merc Kris/K-club, or 13.5% more than k-club alone at 75.

I rather doubt that /DRG will not trounce it at 75 though.

[Retrieving/C.D/Searain]

Without a parser, the only methodology is to count manually the number of attacks per round, as was done with Ridill, but you would have to ask Searain if he didn't use any haste equipment. (~4 seconds between rounds seems enough not to observe overlap.)

That's basically what I did: I had my window minimized the whole time and I would type the number of swings into notepad as my character kept attacking the robber crab and I made a point to keep my character naked the whole time, while single-wielding Ridill (didn't use food or buffs either, aside from the signet boost).

This is what I got during the first test:

1322212312222113322312332212322121122212223222231222221222231321213222223212312221222313233223122321 2112212223222223131221223231222112131213212112223331312332223222222211311221213221122212211221221213 1232332221223222213221121331221321222222211322223222212233223212213231322223133212122222131222122232 1122123112211232122222212211221213232322321121122122222233222121221112123321132221222332322132321233 2223211232122121123222312221323121132211131231222132212312313221323221122222223231122212122132122223 1221231312312221222222112122231322122221221111222233122212331131222121221111231221112222222212311113 3222232221122223213222212223112222222123221212322322211123321231333231212123232121113113133133132321 2133123222332223222123212123223221222221222212222313223312122221312312212222331323122312222221222221 3332122323231322221212131131132323322222121222333222322213221321222321132212332121222121322332222222 2222121112122123222222221313323211311312223211213221331222132332112212212111121221123231212331311221 32322123223212212221

1020 rounds of attacks, the proc rate turned out to be → 27% single attacks, 53% DA, 20% TA.

I then repeated the test while wearing Brutal Earring, Fighter's Calligae and Warrior's Cuisses and got these numbers:

2212131322322232221212222131122122222112231211213222322221122322212122112211213123122323222122122222211123322122221231311332313 1132112121212222122211232121122223212122222222232122211222132123123122222321222121123122322223221222323222222211222331212222222 3112131212222223212112321321322212222232222213323112213112213221112221132122222222222233122122212222212222221221112122213221123 2211232212222322131222122122212121223221222222122222123322212233222221213223122222221222123222222222133121123222233221322221331 2221223212123322221212322221222222222121332322223113121222221122222231222223221322221222212221332221232212222123222111221211222 2223222331223223222211222113222222211322212222323131322111222132133211232212212212331222232123321231223123322112122311323222111 2122122222232121112222211222221122212222222211222232332132212121322322312222121222122223312212112112222221333122123333322322122 3222331322223222321233222211221222222222121222122221211122121331211112122323233322232221223221212221222222222221322123111221321 111222

1021 rounds of attacks → 25% SA, 60% DA, 15% TA.


I don't think haste gear matters much though; during my last test, the one which involved a 2box'ed COR mule spamming Fighter's roll, I used Scorpion Harness + Haste gear but the swings weren't overlapping anyway - for those who're curious the results of this test were the following (please note that Winning Streak and Snake Eye weren't in the game back then, besides, my mule was just LV 50 which resulted in a "nerfed" Fighter's Roll effect):

21222212232111212331212122212232322112332211222232221211122221123131223312221222232212222131112222232322112 31222131312221113221212222222222121121223122132122221212122322312211231112122222221222322212312112212222222 22132122211212212221233123222232222322V12211222112222222231233212232223222221222222232311112223212222221222 12131211112223131323322122222212122212122112323222232222222221312212312212211221211233211212222112122222223 22212222122322221221122222122322321332222232223212212121213232333122322322123222221232321233113333121122221 22312122223222231221111221231321122322132112221212322212212221112122132212223221212111121223121312212222313 22112222221232212222222212132211213211211122112311213321223311113121122222322231122221222323122211312222122 22222222323222222223211123332222111112223211211222211222223222122122222313212321121321222212131223212211222 23222332121121312223222222121221222133221231332321212222321322232121222222333121122122221123232122212222112 31221222312222222313113212212333212222112221222211212222221121111222213222322122222122212222233223211112222 32323222212232222211221221232212223222223222322323212132213131212332122122221122222222123121123223232121113 31212222312212221222312221222221322222322132132221231122111232322222213232322132122312222213122223222322221 222212222222222213122212121221112222123222122221322231222322212122311221312133222122112232321

the COR rolled:

6, 9, 8, 4, 4, 7, 7, 5, 5, 7, 6, 8, 6, 7, 8, 8, 5, 10, 8, 5, 8, 6, 8, 6, 7, 7, 5, 6;

Ridill proc rate → 26% SA, 59% DA, 15% TA.


Still, as I tested all of the above as War/Nin you must take into account the influence of the inherent Warrior DA trait as well (didn't have DA merits at the time though).

[Masamune]

Still, as I tested all of the above as War/Nin you must take into account the influence of the inherent Warrior DA trait as well (didn't have DA merits at the time though).

Dang... was happy to see your results till that last sentence made me like "orz...".

Like it has been discussed in that thread http://www.bluegartr.com/forum/76198...ml#post2794225 with CDF, having a single DA trait or gear is enough to skew a distribution testing. In that case of testing "base" distribution as war means both singles and TAs % got reduced, in favor of DAs.

So that means that what's has been thought for so many years 30:50:20 ridill distrib were actually ..... slightly wrong ;; /em shot himself...

Nonetheless, Thank you Searain for pointing us to actually redo Ridill testing.




EDIT: with probabilities model developped from mentioned thread above, i get following distribs:
- as war (10% DA not merited), ridill equipped only, 100% accu: 27%/55%/18%
- as drk (0% DA) same conditions : 30%/50%/20%

... and that is with having base Ridill distrib in my database entered as 30/50/20. Made that edit just to show how 10% DA trait can skew a distrib (if model is correct).

Also, after verifying Searain's 1st test data into excel i obtain these results:
http://www.bluegartr.com/forum/attac...1&d=1259331637
... that shows again that, considering the sample size, can't know how much % the war 10% DA trait skewed distrib. But that's still consistent with the 2% skew in both singles/TAs from model.

EDIT2: with probabilities model developped from mentioned thread above, i get following distribs:
- as war (10% DA not merited), ridill+AF2legs+Brutal+AF1feet equipped only, 100% accu: 24.9%/58.5%/16.6%
... still with same base Ridill distrib in my database entered as 30/50/20, and AF2legs=AF1feet=1% DA.
Then, after verifying Searain's 2nd test data into excel i obtain these results:
http://www.bluegartr.com/forum/attac...1&d=1259335074
... again considering relatively wide confidence intervals, both model's results and 30/50/20 asumption for ridill's distrib are consistent :/

I won't bother with 3rd test, f..k cor's roll randomness @ tests.

Conclusion: still need more samples :s (to get below 1% error to be exact), and/or redo testing with another job than war.

[Byrthnoth]

Well, if we knew how much DA Warrior's Cuisses, Brutal Earring, and Fighter's Calligae give, can't we use the change in DA rate and the change in Ridill Proc rate to tell both how DA affects Ridill and to predict what the DA-less distribution is?

[TheMysteriousX]

Well, if we knew how much DA Warrior's Cuisses, Brutal Earring, and Fighter's Calligae give, can't we use the change in DA rate and the change in Ridill Proc rate to tell both how DA affects Ridill and to predict what the DA-less distribution is?

Or y'know, use something other than WAR to parse it. Someone here must have a BST or DRK or SAM or RNG.

Its easy to get a rough figure anyway. Take the number of single hits, divide it by known DA%, then remove that number of hits from column 2, and stick them in column one.

That will give you an approximated, adjusted ridill proc rate.

[Masamune]

Its easy to get a rough figure anyway. Take the number of single hits, divide it by known DA%, then remove that number of hits from column 2, and stick them in column one.

That will give you an approximated, adjusted ridill proc rate.

Wrong, since DA job trait or gears reduces BOTH single hits AND triple hits rounds.

[TheMysteriousX]

Wrong, since DA job trait or gears reduces BOTH single hits AND triple hits rounds.

Forgot the triple attack part.

The problem is indeterminate then.