The Sealed Dagger: A Ninja Guide

So then to keep this relevant for NIN...Kannagi's 70 AGI could be significant for monster TP gain, in a solo scenario this could be major with the rest of the Subtle Blow that a NIN acquires. But then if we can't even reach that window during general TP phase, it would be irrelevant...

Well. According to the aforementioned spreadsheet currently set;

A level 139 bat has 301 AGI
and a player (in the tp set, dont tell me it's not optimized I didnt make the damn set) has 111(race)((mithra)) + 175 AGI

Leaving 15+70 to cap the dAGI stat (which is conveniently pretty close to what kannagi offers)

There's a spreadsheet on page 1, it's available for everyone to check the "evidence". It's pretty good, even though no spreadsheet is perfect, and Langly put in a lot of great work on making it one of the better currently maintained FFXI job sheets. Plug stuff in and see for yourself, it's the best you're going to get other than getting both weapons and parsing them yourself. Don't forget to turn on Empy AM3 (50% ODT) for Kannagi (or hell, even AM1 30% ODT if you're going to say it's so hard to maintain AM3... it isn't though, especially with Sange when you need a quick TP burst).

Also... there ARE a couple issues with Kikoku 121 (Aug) in the current spreadsheet. The normal AG 121 version picks up the Relic AM fine, but the augmented version does not. I addressed that the quick-and-dirty way in my own tests by just sticking Atk+120 on the custom line for TP and WS (which is actually overestimating the Atk+, but I want to give Kikoku the benefit of the doubt as much as I can). I also don't believe Kikoku was getting the extra 20% Metsu damage from AG augment - not 100% sure whether that acts just like normal WSD, but I stuck an additional 20% WSD in the custom line for WS set to test.

EDIT: well, one way to fix the AM issue is just to rename the augmented Kikoku to "Kikoku 121" (where the heck are the cells that define the weapons that receive AM effects anyway?). That corrects the Relic AM toggle issue, though it doesn't address the seeming lack of Metsu Ws Damage (so I still did that via WSD in the custom line in WS set, which I think might not be right...)

For everyone who is too ***** lazy to go to the first page, download the spreadsheet, and play with gear for yourself... here's one with Kannagi and Kikoku side by side with good gear for each (identical high end TP sets, similar quality high end WS sets).

Note that I took Langly's currently posted sheet from page 1 of this thread, and made some minimal changes:
* Added Mache/Mache+1 earring
* Changed a couple GEO buff values so they all reflect a 900 skill Idris (non-Bolster/BoG)
* Used temporary Kikoku solution for the 121 augmented version not reflecting AM or WS damage boost, as discussed above
* Note that I think this sheet was last updated during the 'WSD applies to all hits of all WS' period in Sep-Nov, so may not be accurate for non-single hit WS like Shun. Shouldn't be relevant to the Ten/Metsu/Hi comparisons I've been doing.

For argument's sake, I defaulted this to buffs that people are claiming are a good situation for Kikoku/Metsu: i.e., severely uncapped. I took every freaking buff off except for Haste/Marches for capped delay reduction. No atk+/def- buffs/debuffs, no Berserk/Innin, just capped delay and good gear sets. You're getting every ounce of possible benefit from Kikoku's Atk+60 and Atk+10% AM. Kannagi still wins convincingly. Now, start adding your own buffs. Load em up, multiple GEO bubbles, 4 Rolls, 5 songs, whatever JAs. Kannagi wins consistently.

Aside from spreadsheet testing, I've personally compared myself using each of these two weapons (well, the non-augmented AG ones at least) in actual parsed situations on the same content. Several Ambuscades with same party/buff/strat setup with one weapon and several with the other, 60min of Apex mobs and same buffs with one weapon and 60min with the other, etc. I consistently get better DPS with Kannagi in real life situations. If you don't want to believe me, fine. But you better come with a more convincing argument than "I don't think Kannagi is good cuz Blade: Hi was lame when I tried using it on my [insert not-Kannagi weapon here]" or "I think Empy AM is bad despite not owning an Empy, and the spreadsheet must just be a total lie."

I'm just showing you a spreadsheet because it's the closest model I can provide, and I'm sure people are going to complain that the spreadsheet is wrong and inaccurate anyway, despite asking for evidence of what I'm talking about (and providing precious little of their own)...

Bonus!!!
I've included two other sheets comparing Heishi-Kikoku and Heishi-Kannagi, so you don't even have to use those tricky pull down menus and pick the gear and buffs yourself!

(1) Guess what, for Heishi-Kikoku I again started with a situation supposedly favoring Kikoku (very low buffs, getting all of that Atk) and... Heishi barely wins (practically a tie) even in that best case scenario for Kikoku. Heishi widens the gap substantially the more buffs it gets.

(2) I also get Kannagi beating Heishi pretty much across the board, whether lower buffs or heavy buffs. That was NOT the case prior to augments, you can easily see that in the spreadsheet by changing to the non-augmented versions of both and observing Heishi winning easily when getting a lot of buffs. (TBH, these post-augment findings were a little surprising and I wondered if something was overlooked on the spreadsheet... but I couldn't figure out anything that seemed off)

Sure seems to me like that supports the idea that Kannagi augments were indeed helpful, and that extra AGI/DEX+20 is meaningful. Yes, Empy AM is more annoying to deal with than Aeonic/Relic - but even when you bump it down to ODT+30% (1000tp WS) Kannagi is looking pretty good. And yes, Metsu SCs pretty nicely... but that doesn't make up for its other deficiencies.

I do not claim to be perfect, nor do I claim to know every issue or error in a spreadsheet which I didn't make (although I'm familiar with it and I am confident that Langly has consistently demonstrated he knows what he's doing). If there's some glaring error with the sheets, by all means please review and suggest corrections here. I've backed up my discussion with math (when nobody complaining about a lack of numbers was showing any to support their arguments). Please contribute to the community and discuss where there might be mathematical or modeling flaws if you see issues.

The Augment STACKS with the baked in boost. I'm not seeing it anywhere that Metsu has +30% from the ilvl version of Kikoku, but the augment gets +20% that ALSO stacks on the 30%, meaning that a fully augmented Kikoku does 56% more than a simple 80% DEX, 5.0 fTP WS Metsu.

It's a hidden effect and it's 40%.

Yeah, I meant I'm not seeing where in the spreadsheet it's being displayed. And if it's 40% that means that with a r15 Kikoku Metsu has +68% damage.

Gee, maybe they didn't plug it in correctly and is why they are seeing Kikoku perform so poorly?

It's in D254 and E254 in the data tab of the spreadsheet

Welp, problem solved. Pack it up bois~

That looks to be the 40%, should update it to 68% if aug 121.

ItemSet 363448

Check my math:

230dex gear + 100 dex base * 0.8 = 264

264 + 156 (r15 kikou) = 420

420 * 5.1 (fotia) * 1.44 (gear) * 1.4 (base bonus) * 1.2 (r15 bonus) * 3.575 (pDif including neck bonus) = 18525

Edit: pDIF cap is overstated (fixed), missing fSTR, how does the sub weapon factor in? (I normally only math SAM and DRK)

420*5.1*1.44*1.4*1.2*4.125 = 21375
412*1.23*1.4*1.2*4.125 = 3511

24886

Used Kannagi's base damage. Didn't mess with fSTR because it's target dependent. At best, it's an additional 1485 damage. At worst, you lose 1015 damage.

The spreadsheet put it at 22346 (with a different set) so I would say that's pretty damn close to accurate (on the sheet).

So heishi does still beat it by a significant margin. And that also means capuchin's numbers are right.

(can't see the filer dropper ***, don't have it, not getting it)

Use that set please, looks like BiS on paper, especially with all the WSdmg being multiplicative.

Data!D274/E274 and E275/D275

Apex Bat, dia4, frail/torpor Minuet 4 5 HM Madrigal Sam Chaos (sushi)

The set that was left up; 21367

Yours:21576

(Fotia changed to Windbuffet+1):21697 (I don't fully agree with this but that's what it states)

Kannaria TA vs Ochu Offhand is only 50dps in favor of Kannaria

Heshi blade tens: 24289

Don't even look at a Kikoku blade:ten you will only be disappointed lol

Fixed the Aug 121's lack of aftermath. (Data!65 + 127)

The relic bonus is listed around D274/275

It isn't said enough, thank you Capuchin for using the spreadsheet and telling me where to fix it.

IF the ODD/OTD tabs are set correctly (im not sure, just asked langley what they should be)

Kikoku; 5490
Kannagi; (maybe optimized Blade:hi set?) 5672
Heishi; (maybe optimal Blade:ten set?) 5873

(giving kikoku +70 agi because why not) raises dps by literally 1

For everyone's knowledge:

The ODD/OTD field only works for Kannagi. It will not do anything for any other weapon. Kikoku's triple damage contributions are found in Data!B256


No worries there, it was removed after the changes.

So a single SC per minute would make KannagiKikou beat Heishi?

Gravitation vs Frag/Darkness is a huge DPS difference that spreadsheets won’t show.

Kannagi BiS???

A spreadsheet cannot properly answer a question like that. A single skillchain (if applied equally across all 3 weapons) shouldn't change anything really. But with kannai's ability to Hi>Hi darkness... JUST based on using a single ws, kannagi is BiS including sc damage. If you include Metsu>Metsu it might make kikoku win, metsu does more damage than Hi.

If you want a specific answer to "if I'm soloing and using this ws > ws > ws combination" you just have to physically do it, or run extensive math.

Spreadsheet will only (give you an idea) of what is better with only one ws in mind and no change in buff/status/debuffs. (I know you know this)

I'm not qualified to answer the question "if kannagi uses hi>hi to make skillchains and kkk uses metsu>metsu to make skillchains, will the skillchain damage addition make kkk>kannagi"

if you're -soloing- for sc it's not just metsu metsu

shun metsu light metsu double light

Right, but then you're way out of the scope of what a spreadsheet can answer. Which I just said.


If we start using the "can I use any combination of ws to make the best possible skillchain and the highest possible DPS" I'm going to assume that Heishi takes that unquestionably.

Correct out of the scope of the spreadsheet but it’s 1 SC per minute with two WSs that offer great compatibility with other jobs/WSs. You should catch one a minute even in a Zerg obviously a lot more with less DPS in the group.

Pretty interesting discussion.

1) Re: "solo NIN DD" discussion... What the hell would anybody be doing with a NIN as your sole DD, where any of this really has any practical significance? If you want to make a Kikoku for that because you like its self SCs... OK? But the same target is going to melt with any RMEA (or good non-RMEA, really), and you're getting a weapon that is worse for high end content playing in an actual group.

2) If we're talking SCs, it's not as if Heishi is THAT bad with Shun/Radiance (including self-SC if solo for whatever reason). And despite Ten not having the most amazing SC properties, it does make Darkness with a lot of regularly used WS from other jobs: Fudo, Rudra, PK, CdC, Torcleaver, etc. Kannagi isn't god-awful either, Hi is pretty commonly a good SC WS (and unlike on other weapons, Kannagi Hi doesn't suck), and you don't lose access to Shun.

Kikoku/Metsu really isn't just walking away with things when you start adding SCs. I think this argument is people grasping at straws because they asked for math, I showed math, and they didn't like said math so now they're theorizing about stuff that doesn't show up on a spreadsheet.


Shouldn't need any program download! You CAN get their program (don't), but all you need is to insert the captcha letters and it's a direct download of my Excel sheet.


And thank YOU, as always, for doing a fantastic job keeping the thing updated!

What is this?

We took the data you helped make and broaden the discussion to what actually may occur. It was a logical discussion that can’t be answered with a spreadsheet.

If anything SE made these weapons extremely close at r15.

Ehhhh I still don't want to download from it! I'm curious what conditions you ran it at to put kannagi in first? What I saw was heishi edging it out (but thats with a complete set of geo/brd/cor buffs) (I know that you, like me, don't automatically assume 100% cap 100% of the time so really curious what you used)

Originally I assumed the agi would be at least in part of the reason so I assume you ran it with less buffs than it was at currently.

Also, spreadsheet has no mache earrings, I had to add them manually for taint. So much work.

All 3 actually are in the realm of "equal" except that 500 tpbonus is super broken comparatively. (when we start branching into the using more than one ws line of questions*)

That doesn't make double light, metsu is double darkness.

Ku > Retsu > shun > metsu.

Ten/Ku - > Metsu - > Shun

Ku - > Retsu - > Ten/ku - > Metsu

Metsu - > Metsu

Taking a look at the links Capuchin provided, they're entirely incorrect. I see that the lack of Kikoku R15 aug was addressed, but what wasn't addresssed is that in the spreadsheet links you provided, Blade: Hi is having its R15 aug being counted twice. Using the Kikoku - Kannagi link you provided, going to the Data tab, then scrolling down to cells D278 and E278 you can see errors screenshotted at these links. (The Kannagi set being on the left side and the Kikoku one being on the right)

https://i.imgur.com/2sYKo1m.png
https://i.imgur.com/5hGYTE4.png

You can see that there are two 110% multipliers for Blade:Hi in this. So the changes I made were to remove Aug Kannagi giving a non 100% value in cell D274, adjusting the formula in E278 to also include E275. After this, I removed the 20% custom wsd damage and adjusted the custom attack aug in WS set 2 in order to better simulate the increased attack of the WS set, changing it from the 120 provided to 137. The results of the changes can be found in the next two screenshots I took.

https://i.imgur.com/tvefYNr.png
https://i.imgur.com/Cm9srPf.png

So at this point it's showing 1817 vs 1714, which is definitely within spitting range considering a few other factors. This includes SC damage being higher with Kikoku since the average WS damage is higher. Also from what I can tell Empy AM3 is just a straight 50% multiplier to Melee DPS which entirely isn't realistic and has a higher dps cost to keep up compared to Relic AM.

In conclusion, a few factors of the weapons that cannot be easily simulated using a spreadsheet makes it pretty close to a wash to really say which one is better in that specific scenario. The weapons are (surprise!) better for different things. The SC value of metsu will change a lot based on factors such as enemy resistance, if solo or SCing with a partner, and the SC used. Any time you need fragmentation Kikoku pulls a huge advantage since you don't need to press Kamu. Aside from this, I also downloaded the Heishi - Kannagi spreadsheet which contained the same error, which is why Kannagi showed it was ahead in a situation it normally would not be.

Overall, if someone is requesting proof of Kannagi being better with evidence other than a spreadsheet, then you only give them a spreadsheet with critical errors in order to further your argument while throwing thinly-veiled insults, I'd really consider apologizing since the intent was clearly to get some discussion on the two weapons going and this discussion has lead to some fixes on the spreadsheet as well as providing more points of debate for other NIN player's to decide which REMA might be more suited to their needs.