New Monk Weapon skills

[xbobx]

question. I see a lot of testing in abyessa. Isn't it an obvious thing to conclude that the crit ws with all the crit bonus in abyessa will do more for this new WS?

what about outside, wouldnt the new ws be better because it doesn't rely on crits to be very effective? Since the game is moving away from abyessa.

[Greatguardian]

They're not using Abyssea to test "How good it is". They're using Abyssea to narrow down fTP hit distribution and possible Attack penalties that we're seeing in other WS.

For those who just want a "Here's what you use, bro" summary:

95 Vereth w/Smite = 95 Spharai w/Spiral > 90 Vereth w/Smite > 85 Vereth w/Smite > 95 STR Taipans w/Spiral > 95 WoE w/Smite

Throw in whatever AH/R/EX weapons that aren't magians around the Taipan/WoE mark because I don't even bother following those.

[MarkovChain]

Shijin Spiral has an attack boost effect, so it can effectively deal damage to stronger opponents. Thus, we decided that there was no need for adjustment. Also, please keep in mind that its modifiers are DEX and not STR.

mewwwwwwwwwwwwww, going to guestimated this boost after dumbamis.

[Monchat]

so whats better here 99 relic using new ws or a 99 vere with smite no clear anwsers are shown lol

they seem so farm to be on par ( lv 99 relic versus empy). verthragna users have no reason to use shijin from what we know ( because of how impetus boosts VS), but its the best WS for people w/o verethragna. Thats the perfect situation that makes everyone happy if you ask me.

[Motenten]

As a dev has explicitly stated that Shijin gets an attack boost, this needs to be retested.


Elvaan (per Markov's avatar icon) mnk/dnc: 92 str (no merits)

Weapon:
Lvl 95 Vere (17 str)

Total str: 92+17 = 109. fStr cap at 111
Add 2 str merits to get it to 111

No other str or att gear during weaponskill. Dex is fine.

Skill @99: 424 (no merits)

Base attack: 424 + 8 + 111/2 = 487

Greater Colibri lvl 81: 322 def
Greater Colibri lvl 82: 327 def

cRatio at attack bonus of:

00% -- 1.512 / 1.489
10% -- 1.661 / 1.636
20% -- 1.814 / 1.786
25% -- 1.888 / 1.859
30% -- 1.966 / 1.936

If the buff is 30% or higher, need to retest with more defensive mobs to get accurate values.

Weaponskill the mob with around 5% HP left to only land one hit.

pDif Damage range (including 5% damage spread):

00% -- 1.014-1.945 / 1.000-1.915
10% -- 1.183-2.132 / 1.154-2.104
20% -- 1.355-2.292 / 1.324-2.263
25% -- 1.439-2.370 / 1.406-2.340
30% -- 1.527-2.452 / 1.493-2.420


Example:
Base damage @424 = 49
fStr: 12
Weapon: 39

Dex: 88
Byakko's Haidate: 15
(most other dex gear I can think of offhand has str on it, so avoiding; ensure TP gear with str/att on it is not used for ws)
Total: 103

WSC: 87

Base damage: 187
fTP (with gorget): 1.1602

Final base damage: 216

Damage range:

00% -- 219-420 / 216-413
10% -- 255-460 / 249-454
20% -- 292-495 / 285-488
25% -- 310-511 / 303-505
30% -- 329-529 / 322-522


Note that these are still just approximations of the range of values, not exact predictions. A couple dozen weaponskills should get a fairly decent idea of the attack bonus, though.

Edit: fixed errors

[MarkovChain]

PUP/WAR 99/49
591 att => using defender so 443 att
STR 178, DEX 216, H2H 432 D+39

Target= decent challenge sand sweeper for 451 def (stops checking low def @564 att)

shijin=

[1522,1614,1384,1859,1240,1755,1236,1877,1221,1515,1193,1650,1419,1427,1140,
1382,1394,1659,1201,1369,1579,1544,1076,1774,1138,1455,1398,1799,1102,1687,
1453,1540,1574,1415,1763,1347,1355,1140,1487,1372,1909,1493,1479,1438,1396,
1596,1531,1840,1577,1764,1364,1481,1409,1592,1455,1468,1363,1085,1372,1427,
1810,1604,1542,1498,1478,1694,1359,1359,1508,1389,1470,1457,1312,1461] (74 WSs)

mean : M=1473.44
max=1909 / min=1076 / midle=1492 (data is about centered)
std.var=74
CI=+/- 44

my ftp averaged with acc would be
ftp:=0.95*1.0625*(floor(432*0.11)+3+12+39+floor(0.85*216))+4*0.95*(floor(432*0.11)+3+12+39+floor(0.85*216))=1365.86

The % increase of pdif ranges therefore between
[(M-CI)/(ftp*443/451),(M+CI)/(ftp*443/451)]=[1.065451220, 1.131042864]

So there is a boost that is between 6.5% and 13.1% .. make it 10% then, no too bad.

[Motenten]

Interesting. Adding it to the spreadsheet.

Treating it like Berserk and Boost, in that it's an addition based on a multiplier on the base attack, not a multiplier on the cumulative total attack. That will lower its relative value when combined with other attack boosts.

Note that a percentage boost to attack is more valuable the higher the level correction. So the mention of being better against higher level mobs makes sense.

Quick check on spreadsheet damage: with Impetus up, Smite goes from a 20% lead (on Bukhis target) to just a 10% lead. With Impetus down, Shijin goes from tied to having a 9% lead. Overall, generally closes the gap on DPS vs Smite; Smite still has a decent edge due to ODD, though.

[MarkovChain]

err small correction, I made a logical error in my last post but the conclusion will be the same : i have 443/451 cratio (that's not the pdif average) so according to masa's work
pdif average=-0.12311+(443/451)*1.129179=.986039217

so the actual "average pdif" increase lies in the interval (with 95% likelihood as usual with stats)

[(M-CI)/(ftp*pdifavg),(M+CI)/(ftp*pdifavg)]=[1.061369412, 1.126709770]
It still appears to be 10%, and is still good news. It should make Shijin always superior beside high crit rates situations.

edit: this is the increase of average pdif only. If we want the increase of cratio we have to backsolve the average pdif equation : cratio during WS lies in
[1.035851591,1.092909113]
so the cratio increase lies between 5.4% and 11.2%

Also note that I never TPed exactly at 100 TP so if the boost varies with TP it's going to be harder to guess correctly.

[MarkovChain]

Ok I tried it out in a simultation

I compare spharai(99)+ shijin spiral VS verethragna(99)+victory smite

Assumption for spharai:

usual TP gear, capped accuracy

melee attack (base with food and gear) 730
WS attack 730
10% crit rate,3% triple attack rate,21% double attack rate, 150 str on TP and WS
DEX on WS : 200 I know I'm targetting high end setups (consider that 2x march will give you 10 dex also with marches)
TP skill:=426 WS skill=412

mob's def:=500*0.90 (dia) 1 level difference
no boosting
693/1024 haste (maxed)

Kick rate rate:=0.25*1.1
regain:=1/tick
tripleproc on relic 12%
no gorget
counterrate:=80%


Assumption for vere:

usual TP gear, capped accuracy

melee attack (base with food and gear) 700
WS attack 700
10% crit rate [with impetus and VS's natural crit it averages to 39% crit rate on ws, using rancor collar : 15 base 15 boost and rest comes from impetus up/down],3% triple attack rate,21% double attack rate, 160 str on TP and 200 on WS (pushing it near the high end too)

TP skill:=426 WS skill=412

mob's def:=500*0.90 (dia) 1 level difference
no boosting
691/1024 haste (maxed)

Kick rate rate:=0.25*1.1
regain:=1/tick
ODD rate 15%
no gorget
counterrate:=65%

For mob's stats :
110 vit, 500 def + dia2, +1 level difference
10k HP, 4 sec delay between each fight, I don't use any ws under 1k HP remaining, mobs have 4 second delay (for counter calculations)

Here are the results with current knowledge

"DPS spharai,vere, spharai/vere"
[388.9669757, 379.0248486, 1.026230806]
"acc average"
[0.9561073074, 0.9563744842]
"wsacc average"
[0.9498607832, 0.9492771651]
"critical%-melee"
[0.2059280258, 0.2092076548]
"critical%-ws"
[0., 0.3891860806]
"critical%-global"
[0.1623425645, 0.2409357812]
"melee%"
[0.6504023593, 0.6575677767]
"melee dmg p/s"
[268.4073412, 262.8751178]
"WSdmg p/s"
[144.2715757, 136.8937383]
"number of rounds per WS avg"
[6.428989197, 6.809890266]
"avg Ws dmg"
[shijin=2489.513089, victory=2273.579329]
"avg ws damage divided by non crit pdif(include crit dmg+ and critrate)"
[1491.588933, 1590.139342]
"avg punch incl. KA and DA"
[236.8036157, 217.2948220]
"counter damage%"
[0.03030805889, 0.02172873248]
"counter rate"
[0.7591093767, 0.6173330208]
"distribution of the # of rounds to get 100% TPs"
[0., 0., 0.003440445882, 1.153237460, 17.77059107, 40.76515516,25.11387876, 9.461914264, 4.188398816, 1.345902429,0.1857840776, 0.01100942682, 0.0006880891764, 0., 0., 0., 0.,0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]
,[0., 0., 0., .1816027408, 6.874500914, 35.80278715, 37.43592386, 12.12488086, 4.603822673, 2.238479174, .6562169960, .7792174339e-1, .3863888102e-2, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]

With lower haste spharai will win even more due to counter. SS is decently higher in damage (9.5%) and spharai's ws damage per second is slightly higher than vere (5.4%) since vere is still a bit less delay. Spharai's stronger base damage and 3x proc helps it surpass vere's dots even though it has 15% ODD rate : spharai melee hits are 9% stronger but it's slower so in the end it's 2.1% better melee damage per second only. Also in practice Shijin has much less variance than vere so it's more reliable in a sense. Note that the counter+ on spharai only slightly helps damage output due to capped haste.

We know have to wait what the afterglow is.

Spharai is ~2.6% better than vere for pure damage not counting the tanking capability (-5/tick TP for 30 second on shijin and more counter).

[MarkovChain]

I continued the data posted in #76 of this thread :

PUP99,432skill,D+39
No multiattack, no gorget
219 DEX / 578 att / 173 str

I did easy preys sand sweepers ; only registered WS that didn't kill them. I stopped when they started to repop decent challenge.

shijin=[2128,2206,1963,2124,1676,1878,1943,2033,1912,2073,1950,1853,2166,2240,1104,2200,2124,1289,1425,1968,1882,2015,2237,2395,2068,2261,2265,2207,1843,2175,1672,2230,1886,2320,2422,1845,1985,2039,1969,1794,1879,2244,2148,2184,2223,2466,2099,1819,2214,1955,2235,1782,1494,1809,1818,2530,2090,1772,2070,2123,1921,1938,2318,1640,2215,2073,1675,2335,2032] 69 WSs

Mean M=2012.550725
IC : +/- 62

def=392
cratio=578/392=1.4744
in this range of cratio the formula for the average pdif is 0.7486*x^2-1.5814*x+2.1349 with x=cratio
The interval in which the avg pdif on WS lies is therefore
[(M-r)/(ftp1),(M+r)/(ftp1)] with ftp1=0.95*1.0625*(floor(432*0.11)+3+12+39+floor(0.85*219))+4*0.95*(floor(432*0.11)+3+12+39+floor(0.85*219))
so [1.413144949, 1.502981102]
Backsolving the pdif equation : cratio lies in
[1.445467775,1.577299846] so the pdif "increase" is between
-2% and 7%... In this case the attack boost is not to obvious.