Rehauled Weapon Skills: tier lists

[Fival]

Alright so, in abyssea misareaux on coastal colibri, exactly 2000 TP.
101+103 DEX (assumed dDex is giving me +15% crit rate), 5% base, 5% merits, 5% ranc mantle, 3% nef collar, 2% yetshila +1, 20% dark depths.
That comes out to 55% crit hit rate.

Spoiler: show

Ukko's Fury
+ 2256: 1
+ 4580: 1
+ 4582: 1
+ 4602: 1
+ 4649: 1
+^ 4658: 1
+ 4661: 1
+ 4680: 1
+ 4701: 1
+ 4743: 1
+ 6914: 1 < Mob HP was too high, extra hits landed.

The non crit immediately rules out a 10% increase to the old value of 35%, which likely leaves a 5% increase like the 3000 tp value, because they wouldn't really do something ridiculous like 38%, right? right??

[Motenten]

@dasva -- It's very difficult to ascertain the values given the data provided. Ideally, you need to double the number of samples (for 20 total), plus do another set of samples while using gorget+belt (if possible; only one is acceptable if both aren't available).

I can say that current analysis indicates a 73% WSC with a 204/256 (0.796875) fTP as the best fit. However the range of valid WSC values is very narrow, and easily shifts with a little uncertainty in the base damage value. fTP in general looks like it's between 0.77 and 0.82. It's very difficult to get a WSC above 75%.

[Byrthnoth]

Fival, are you sure that the 2256 was actually a non-crit and not a miss of the first swing? How much STR did you have?

[ccl]

drk/thf 221 str 5/5 reso 264 base dmg wpn

2000tp: 2428 2378 2419 2408 2476 2441

3000tp: 3662 3596 3564 3649

[Viar]

Meanhwile, data from Studio Gobli.
If someone like Kincard or anybody else wants to translate... they have some data, might be cool to compare it with the data we gathered on our own and doublecheck things.
At first glance, we found out more stuff than they did, but they worked more on magic WSs

My somewhat inaccurate translation:

魔法WSダメージ＝[[(魔法WS基本D値＋補正)x倍率] + 魔法ダメージ + 系統 ]x魔法攻撃力やら

Magical WS damage=( ((Magical WS basic D value + Mod)xfTP)+ magic damage+ formula((something like AGI- enemy INT)x2 for marksmanship WS).)xMAB or additional bonus.

Magical WS basic D value:

ILV 113 weapon = 186
ILV117= 196
ILV119=201

[dasva]

@dasva -- It's very difficult to ascertain the values given the data provided. Ideally, you need to double the number of samples (for 20 total), plus do another set of samples while using gorget+belt (if possible; only one is acceptable if both aren't available).

I can say that current analysis indicates a 73% WSC with a 204/256 (0.796875) fTP as the best fit. However the range of valid WSC values is very narrow, and easily shifts with a little uncertainty in the base damage value. fTP in general looks like it's between 0.77 and 0.82. It's very difficult to get a WSC above 75%.

Hmmm well this was just on a mule so might be awhile before I feel like getting it a gorget/belt. Have we not determine any of the merit ws yet?

[Viar]

Current summary of what's still needed:

1) Solid new formula for magic weaponskills.

2) fTP for magic weaponskills:

Spoiler: show

Hand-to-hand
Dagger
Sword
Great Sword
Axe
Great Axe
Sythe
Polearm
Katana
Great Katana
Club
Staff
Bow
Marksmanship

Data from Studio Gobli:

Sword:

Red Lotus Blade:
TP 1000:1.00 TP 2000:*2.40(2.00+98/256) TP 3000:3.75
Burning Blade:
TP 1000:1.00 TP 2000:*2.10(2.00+25/256) TP 3000:*3.40(3.00+102/256)
Shining Blade:
TP 1000:1.125 TP 2000: *2.23(2.00+57/256) TP 3000:*3.53(3.00+134/256)
Spirits within:
TP 1000: HPx0.125 TP 2000: HPx0.50 TP 3000:HPx1.00
Seraph Blade:
TP 1000:1.125 TP 2000: 2.625 TP 3000:4.125

[Sechs]

Hmm that's interesting, thanks for translating that stuff.

[Fival]

Fival, are you sure that the 2256 was actually a non-crit and not a miss of the first swing? How much STR did you have?

I'm pretty sure, i was using seigan/third eye and full timing 1 gear set to land directly on 2000 tp, and always have 200 tp return if only the first hit landed, im fairly certain that was a 200 tp return, but if it doesn't add up i can try again.

STR: 106+163
DEX: 101+103
Atma: Dark depths, Sanguine scythe, Mounted champion

[doctorugh]

Ruinator:
STR 251
fSTR 26
Base D 165

No gorget/belt

SA LVL1 mobs: 1346,1350,1354,1364

Looks like they removed alpha (use 3.0 and 5% randomizer to determine right?)

Side note: Was trying to test for +att (which has been listed at 35%), but vs crabs in foret (im guess fSTR cap but not sure) with 660 def (determined by /check), with 993 att, avg hits (determined by watching fTP for # of hits landed) was at 828 (w/ gorget). Sample size of 48 hits (not WS's).

Without any attack bonus number should have been on avg 798, which is only 3.7% below my small sample test and within reason. So.....I'm saying no attack bonus (maybe there never was one or someone tested for +att before they realized gorget stuff was transferred to all hits?)

[Prothescar]

it has an attack bonus, but it isn't 35%, not sure who originally came up with that number tbh. it's only 5%


edit: mah b, got numbers mixed up. it's 10%

[JSHidaka]

Alright, i'm going to start Ukko at 2000 TP now, this will likely take awhile.

Edit: I just realized how much of a pain its going to be to Ukko at exactly 200 TP in abyssea, getting hit puts me at 2019 tp, this may take longer then expected.

Isnt there a gorget that give you TP by loosing hp?...

Edit; http://www.ffxiah.com/item/15530/berserkers-torque

[dasva]

Isnt there a gorget that give you TP by loosing hp?...

Edit; http://www.ffxiah.com/item/15530/berserkers-torque

The hard part is doing this while holding a mob at low enough to ensure one shotting though. Almost think the best bet is to abyssea it up with tons of regain atmas and sleep it when it gets low

[Fival]

Seigan/Third eye worked just fine, with ilvl gear they miss often enough theres very few instanced where they hit more then third eye can prevent, and when it happens i just reset my tp by re-equiping grip. That combined with fulltiming a set that gave 200 tp per hit(including first and only hit landing of the ws), it was working out nicely.

[Byrthnoth]

I strongly recommend setting it up so you can distinguish a non-crit from a first-swing miss.

I'm using Light Gorget/Belt and Phorcys/Ogier's to do this.

[Motenten]

@dasva:

As it happens, I also have Realmrazer at 1/5. So, took some time today to expand the sample space.

Mnd: 85+9 to match your 94 mnd sample, and 85+136 to match your 221 mnd sample.

First did more of the non-gorget tests to add to the ones you had.

94 mnd: (don't really need more samples)
221 mnd: 618, 642, 650, 653, 653, 657, 662

Base dmg: 137 @94 mnd, 211 @221 mnd

Then did tests with gorget+belt.

94 mnd: 515, 533, 536 (almost perfect min-max range in just a few hits)
221 mnd: 789, 801, 801, 803, 804, 804, 807, 812, 814, 814, 824, 826, 826

Base dmg: 171 @94 mnd, 263 @221 mnd

Procedure explanation for determining fTP and WSC:

First, chart out what the WSC would be at even % intervals between 60% and 100%. I am assuming for these purposes that there are no fractional WSC percentages.

For each WSC percentage, determine the WSC value. Divide that value by the determined base damage value to see what the fTP would have to be to get that result. Give a range of [BaseDamage to (BaseDamage+0.999)].

Do that for both the low and high mnd samples.

Compare the valid ranges between the low and high mnd samples, and mark those that overlap. That is, what fTP values would give the observed result in both the low mnd and high mnd cases.

Repeat that procedure for the tests that use the gorget+belt.

For those fTP ranges that were valid (at a given WSC %) for both the non-gorget and the gorget tests, look at the difference between the range values. Look specifically for those with a difference that can match the 50/256 addition from the gorget+belt.

Determine an fTP value in /256 units that works for the non-gorget case, and see if it can also match the results of the gorget+belt case.

Summarize the resulting WSC+fTP values that give valid results.


From non-gorget test:
fTP: 205/256 @ 72% WSC
fTP: 204/256 @ 73% WSC
fTP: 202/256 @ 74% WSC
fTP: 201/256 @ 75% WSC

From gorget+belt test (only testing +50/256 fTP values):
fTP: 254/256 @ 73% WSC
fTP: 252/256 @ 74% WSC

The last two are the two possible valid values. 73% is notable because of the way the game displays information on merit weaponskills (+3% per merit), and the known maximum WSC value before the update (85%). If 85% is 5/5, and each 1/5 is 3%, then subtracting 4/5 is subtracting 12%, which leaves you at 73%.


Conclusion:

Realmrazer has an fTP of 204/256 (0.796875), with a 73% WSC at 1/5.

Supplementary:

Merit weaponskills can be considered to have WSC values of 70% + 3% per merit level. No alpha value needs to be applied.

[Byrthnoth]

So . . . ~7 fTP with gorget/belt. Looks like about 5800~6000 damage average with Mjollnir, capped Ratio, and a decent set against targets that aren't incredibly strong? Not bad for a 1H WS.

[dasva]

Sounds like it might be better than cdc for pld if you can get ahold of a Sharur and you are still capping acc

[doctorugh]

Supplementary:

Merit weaponskills can be considered to have WSC values of 70% + 3% per merit level. No alpha value needs to be applied.

Hrmm, check this again, I was pretty confident the ruinator (new ws) was at 100% no alpha

[Motenten]

Hrmm, check this again, I was pretty confident the ruinator (new ws) was at 100% no alpha

All the data is right there in the post (or a few samples in dasva's earlier post), if you wish to dispute it. The methodology for determining the value is also explained. Granted this is only for 1/5 merits. The scaling may have changed for higher merit tiers. Your sample does seem to indicate that 5/5 may be 100% WSC, but other weaponskill tests have shown other results at 5/5. You do not provide nearly enough information to do a solid validation of WSC vs fTP, nor do we have enough information to guarantee that all merit weaponskills still behave the same.