Item Search
     
BG-Wiki Search
+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 20 of 32

Thread: Playstyle Thread     submit to reddit submit to twitter

  1. #1
    E. Body
    Join Date
    Dec 2008
    Posts
    2,408
    BG Level
    7
    FFXI Server
    Quetzalcoatl

    Playstyle Thread

    So there is a wide collection of daggers dancer can use, and there's also a lot of speculation over what to use and when.

    So I'm going to go ahead and take it on to myself to math all this out.

    So to start, here's the gearset I'll use for comparison: link

    Haste: (256/1024) 25%
    Dual Wield: (30% Trait + 3% neck + 5% Suppa + 10% body) = 48% DW
    Double Attack: (3% Ring + 5% Brutal + 2% Belt + 3% Back) = 13% DA
    Haste Samba: 10% Haste
    Store TP: 21

    Equation I'll use for DPS is DMG * Swings/min

    Daggers

    Phurba:
    DMG: 44 Delay: 190

    Twilight Knife:
    DMG: 40 Delay: 190 "Quadruple Attack" +3%

    Auric Dagger:
    DMG: 39 Delay: 201 Dual Wield +5%

    Parazonium +2
    DMG: 22 Delay: 211 Occasionally Attacks 2-4 times (Averages 2.07 per round) source

    Parazonium +2
    DMG: 45 Delay: 211 Double Attack+10(%)

    I will use 3 values for Double attack based on Saber Dance, these are 13% from gear, then 30%/50% Double attack+ from Saber dance Max/Min based on average merits. Thus to calculate true double attack rate I will use:

    SD%(0~30~50) + (100-SD%)*GearDA

    Normal:
    0%: 13% Double Attack
    30%: 30+(1-0.3)*13 = 39.1% Double Attack
    50%: 50+(1+0.5)*13 = 56.5% Double Attack

    Magian Double Attack+ Parazonium:
    0%: 23% Double Attack
    30% 30+(1-0.3)*23 = 46.1% Double Attack
    50% 50+(1-0.5)*23 = 61.5% Double Attack


    Lets start.

    Dagger1+Dagger2
    Dual Wield: -48/53%
    Haste: -35%
    Combined DMG: Dmg1+Dmg2
    Combined Delay: Delay1+Delay2
    Dual Wield Delay: -(cDelay*DW) = (cDelay-dwDelay)
    Individual Delay: dwDelay/2 = iDelay
    TP/Swing = 5.0 + [(iDelay - 180) * 1.5 / 180] = TP/S
    Hasted Delay: -(0.35*iDelay) = (iDelay -hastedelay)
    Swings/min = 3600/hDelay = S/M
    Double Attack1+ = 0.13*S/M = +DA1
    Double Attack2+ = 0.391*S/M = +DA2
    Double Attack3+ = 0.565*S/M = +DA3
    Quadruple Attack+ = (0.03%procrate*quadhits)*S/M = +QA
    Total Swings/min = 60.3+QA+(DA1~DA2~DA3) = 67.5+(DA1~DA2~DA3)
    = TS/M1~TS/M2~TS/M3
    TPM = TP/S(TS/M1~TS/M2~TS/M3) = TPM1~TPM2~TPM3
    DPS = cDMG(TS/M1~TS/M2~TS/M3) = DPS1~DPS2~DPS3

    ==================================================

    Auric+Twilight
    Dual Wield: -53%
    Haste: -35%
    Combined DMG: 79
    Combined Delay: 391
    Dual Wield Delay: -207.23 = 183.7
    Individual Delay: 183.77/2 = 91.8
    TP/Swing = 5.0 + [(91.8 - 180) * 1.5 / 180] = 4.2
    Hasted Delay: -32.1 = 59.7
    Swings/min = 3600/59.7 = 60.3
    Double Attack1+ = 0.13*60.3 = +7.8
    Double Attack2+ = 0.391*60.3 = +23.6
    Double Attack3+ = 0.565*60.3 = +34.1
    Quadruple Attack+ = (0.03*4)*60.3 = +7.2
    Total Swings/min = 60.3+7.2+(7.8~23.6~34.1) = 67.5+(7.8~23.6~34.1)
    = 75.3~91.1~101.6
    TPM = 4.2(75.3~91.1~101.6) = 316.26~382.62~426.72
    DPS = 79(75.3~91.1~101.6) = 5948.7~7196.9~8026.4

    ============================================
    Phurba+Auric
    Dual Wield: -53%
    Haste: -35%
    Combined DMG: 83
    Combined Delay: 391
    Dual Wield Delay: -207.23 = 183.7
    Individual Delay: 183.77/2 = 91.8
    TP/Swing = 5.0 + [(91.8 - 180) * 1.5 / 180] = 4.2
    Hasted Delay: -32.1 = 59.7
    Swings/min = 3600/59.7 = 60.3
    Double Attack1+ = 0.13*60.3 = +7.8
    Double Attack2+ = 0.391*60.3 = +23.6
    Double Attack3+ = 0.565*60.3 = +34.1
    Total Swings/min = 60.3+(7.8~23.6~34.1) = 68.1~83.9~94.4
    TPM = 4.2(68.1~83.9~94.4) = 286.02~352.38~396.48
    DPS = 83(68.1~83.9~94.4) = 5652.3~6963.7~7835.2

    ============================================

    Auric+OA2-4 Parazonium+2
    Dual Wield: -53%
    Haste: -35%
    Combined Delay: 412
    Dual Wield Delay: -218.3 = 193.7
    Individual Delay: 193.7/2 = 96.9
    TP/Swing = 5.0 + [(96.9 - 180) * 1.5 / 180] = 4.3
    Hasted Delay: -33.9 = 63.0
    Swings/min = (3600/63.0)/2 = 28.6 Each Hand
    Double Attack1+ = 0.13*28.6 = +3.7
    Double Attack2+ = 0.391*28.6 = +11.2
    Double Attack3+ = 0.565*28.6 = +16.2

    MAINHAND(Auric)
    Total Swings: 28.6+(3.7~11.2~16.2) = 32.3~39.8~44.8
    DPS: 39(32.3~39.8~44.8) = 1259.7~1552.2~1747.2
    TP/M: 4.3(32.3~39.8~44.8) = 138.5~171.1~192.6

    OFFHAND(Parazonium) (OA2-4 doesnt proc on swing that procced DA)
    Proccable swings:
    Double Attack1: -3.7 = 24.9
    Double Attack2: -11.2 = 17.4
    Double Attack3: -16.2 = 12.4

    Average hits = Successful proc hits *2.07 + double attack procs*2
    Double Attack1: 24.9*2.07 + 3.7*2 = 51.5+7.4 = 58.9
    Double Attack2: 17.4*2.07 + 11.2*2 = 36.0+22.4 = 58.4
    Double Attack3: 12.4*2.07 + 16.2*2 = 25.7+32.4 = 58.1

    DPS: 22(58.9~58.4~58.1) = 1295.8~1284.8~1278.2
    TP/M: 4.3(58.9~58.4~58.1) = 253.3~251.1~249.8

    Totals:
    DPS: (1259.7+1295.8)~(1552.2+1284.8)~(1747.2+1278.2) =
    2554.8~2837~3025.4
    TP/M: (138.5+253.3)~(171.1+251.1)~(192.6+249.8) =
    391.8~422.2~442.4

    ============================================

    Auric+DA Parazonium+2
    Dual Wield: -53%
    Haste: -35%
    Combined DMG: 84
    Combined Delay: 412
    Dual Wield Delay: -218.3 = 193.7
    Individual Delay: 193.7/2 = 96.9
    TP/Swing = 5.0 + [(96.9 - 180) * 1.5 / 180] = 4.3
    Hasted Delay: -33.9 = 63.0
    Swings/min = (3600/63.0) = 57.1
    Double Attack1+ = 57.1*0.23 = 13.1
    Double Attack2+ = 57.1*0.461 = 26.3
    Double Attack3+ = 57.1*0.615 = 35.1

    Total Swings/min = 57.1+(13.1~26.3~35.1) = 70.2~83.4~92.2
    TPM = 4.3(70.2~83.4~92.2) = 301.86~358.62~396.46
    DPS = 84(70.2~83.4~92.2) = 5896.8~7005.6~7744.8

    ============================================

    Phurba+Twilight
    Dual Wield: -48%
    Haste: -35%
    Combined DMG: 84
    Combined Delay: 380
    Dual Wield Delay: -182.4 = 197.6
    Individual Delay: 197.6/2 = 98.8
    TP/Swing = 5.0 + [(98.8 - 180) * 1.5 / 180] = 4.3
    Hasted Delay: -34.5 = 64.3
    Swings/min = 3600/64.3 = 56.0
    Double Attack1+ = 0.13*56.0 = +7.3
    Double Attack2+ = 0.391*56.0 = +21.9
    Double Attack3+ = 0.565*56.0 = +31.6
    Quadruple Attack+ = (0.03*4)*56.0 = +6.7
    Total Swings/min = 60.3+6.7+(7.3~21.9~31.6) = 67+(7.3~21.9~31.6)
    =74.3~88.9~98.6
    TPM = 4.3(74.3~88.9~98.6) = 319.49~382.27~423.98
    DPS = 84(74.3~88.9~98.6) = 6241.2~7467.6~8282.4

    ============================================

    Parazonium(DA) + Twilight
    Dual Wield: -48%
    Haste: -35%
    Combined DMG: 85
    Combined Delay: 401
    Dual Wield Delay: -192.4 = 208.6
    Individual Delay: 208.6/2 = 104.3
    TP/Swing = 5.0 + [(104.3 - 180) * 1.5 / 180] = 4.3
    Hasted Delay: -36.5 = 67.8
    Swings/min = 3600/67.8 = 53.1
    Double Attack1+ = 53.1*0.23 = 12.2
    Double Attack2+ = 53.1*0.461 = 24.5
    Double Attack3+ = 53.1*0.615 = 32.7
    (I am going to assume Quad Attack and Double Attack can't both proc, right?)
    Quadruple Attack1+ = (0.03*4)*(53.1-12.2) = 0.12*40.9 = +4.9
    Quadruple Attack2+ = (0.03*4)*(53.1-24.5) = 0.12*28.6 = +3.4
    Quadruple Attack3+ = (0.03*4)*(53.1-32.7) = 0.12*20.4 = +2.5
    Total Swings/min1 = 53.1+4.9(quad)+12.2(DA) = 70.2
    Total Swings/min2 = 53.1+3.4(quad)+24.5(DA) = 81
    Total Swings/min3 = 53.1+2.5(quad)+32.7(DA) = 88.3
    = 70.2~81~88.3
    TPM = 4.3(70.2~81~88.3) = 301.86~348.3~379.69
    DPS = 85(70.2~81~88.3) = 5967~6885~7505.5

    ===============================================

    Auric+Twilight
    TPM = 316.26~382.62~426.72
    DPS = 5948.7~7196.9~8026.4

    Phurba+Auric
    TPM = 286.02~352.38~396.48
    DPS = 5652.3~6963.7~7835.2

    Auric+OA2-4 Parazonium+2
    TPM: 391.8~422.2~442.4
    DPS: 2554.8~2837~3025.4

    Auric+DA Parazonium+2
    TPM = 301.86~358.62~396.46
    DPS = 5896.8~7005.6~7744.8

    Phurba+Twilight
    TPM = 319.49~382.27~423.98
    DPS = 6241.2~7467.6~8282.4

    Parazonium(DA) + Twilight
    TPM = 301.86~348.3~379.69
    DPS = 5967~6885~7505.5

    Dagger Combos By DPS, Highest -> Lowest

    Phurba+Twilight
    v
    Auric+Twilight
    v
    Phurba+Auric
    v
    Auric+DA Parazonium+2
    v
    Parazonium(DA) + Twilight
    v
    Auric+OA2-4 Parazonium+2


    Dagger Combos by TP/M, Highest -> Lowest
    Auric+OA2-4 Parazonium+2
    v
    Phurba+Twilight
    v
    Auric+Twilight
    v
    Phurba+Auric
    v
    Auric+DA Parazonium+2
    v
    Parazonium(DA) + Twilight

    Synopsis (AKA: TL;DR)

    SO we have, for highest TP gain, Auric +OA2-4 dagger offhanded. The TP gain of this combo soars above the rest with fan dance up, however once you put up saber dance its TP rate actually gets caught up to by all the other sets quite quickly. This makes sense as you'll often double attack on OA2-4 swings instead of having the effect proc, slowing TP gain a bit.

    The next best for TP gain is Phurba+Twilight, which also happens to have significantly the highest DPS. With Saber Dance up this pair is only 20 TP/M lower than Auric+OA2-4, yet it has nearly triple the DPS.

    Auric makes a decent substitute in for twilight or phurba in this deadly combo, ranking in at second and third place for DPS and 3rd/4th in TPM.

    Tl;Dr Any combination of Auric, Twilight, and Phurba is a recipe for win. Note: all you dancers with +2 AF3 gear, we're fairly certain Twilight's proc effect cancels out or +2 set bonus so avoid mainhanding Twilight. Tops you lose 1 base DMG in the auric+twilight combo, which seems like a good trade off imo for a "just in case" measure.

    If you have any questions feel free to ask. Also if there's an issue in anything I calculated please point it out!

    If you have a request for another dagger combo to see how it compares against the rest, feel free to post, though I'd appreciate it if you tried to calculate it yourself, I even posted the equations up at the top. It's not hard, just time consuming. Most of this is grade 7 math.

    And if you find a rather notable dagger combo in DPS or TP/M, do post it for us to see!

    Finally:
    I spent over 4 hours doing all this number crunching, so please, if you're going to post all of this somewhere else (And I will see it, I browse nearly every big FFXI forum out there) take the 15 seconds to give me credit, ok?

    ~Urat (of Quetz)

  2. #2
    Masamune
    Guest

    I think you forgeting something fundamental in equipement comparison: mobs stats.

    example coming to my mind is accuracy affecting what you call "swings/min".
    another example would be mob's def vs yur attack + weapon(s) base dmg.

    a lot of comparisons i read bit everywhere involves capped situations, which actually happens rarely in "interesting" situations (NMs)...

  3. #3
    E. Body
    Join Date
    Dec 2008
    Posts
    2,408
    BG Level
    7
    FFXI Server
    Quetzalcoatl

    Yeah I know I didn't cover situations where you have lower acc, in which case the dex kila would pretty assuredly rise to be the champion.

    However in few situations where you're up vs something that you have uncapped acc on would you be using any of these daggers anyways. This would be some big HNM or powerful T3+ NM in abyssea. Maybe a boss of an endgame run etc etc etc.

    In those cases on dancer you're doing 1 of 2 things.

    A: Soloing it like a boss. In which case obviously you're using Agi kilas

    B: Back stepping to lower stats for whoever is actually tanking/fighting it, probably a war doing a red proc or a mnk doing blu proc (shinryu etc) though even then the majority of the time you're not even doing that due to AoE, which is prevalent on big NMs. In this case you're probably rocking Eva kila's if the AoE is physical based, or dex kilas to land steps with more reliability, maybe a phurba in to make your cures a little better if you're assisting a monk tank.

    Or finally C: Tanking it for a party, in which case I'd say either a combo of Agi kila + PDT kila for a physical monster, or MDB + Phurba for a magic using monster. Really only parses can tell you whats best from that point on, in which you look at what your lacking and swap to the dagger that fortifies your weakness.

    tl;dr in cases where you don't have capped accuracy you probably have more important things to worry about than your DPS, and higher tp rates are bad due to TP spamming said monster. Ergo: I assumed close to capped accuracy for this.

    Sorry, should have mentioned that in OP =/

  4. #4
    Melee Summoner
    Join Date
    Apr 2010
    Posts
    42
    BG Level
    1
    FFXI Server
    Ragnarok

    One thing that I think a lot of dancers fail to consider is the Samba Set Bonus of +2 gear.
    It's my understanding that samba will not proc on twilight knife, should the drain effect proc from the knife itself. This, coupled with the extremely high proc rate on the knife would cut the effectiveness of your +2 set bonus nearly in half. Some people may scoff at the set bonus which considering dagger choices, but when the difference between twilight and a few other choices is as close as it is, wouldn't an extra ~5% damage push the others into the lead?

  5. #5
    E. Body
    Join Date
    Dec 2008
    Posts
    2,408
    BG Level
    7
    FFXI Server
    Quetzalcoatl

    Afaik there's evidence that the set effect of our AF3+2 comes from having the daze on the opponent, and has nothing to do with the additional effect, therefor Twilight knife does indeed still work with the +2 set, however you're other dagger has to be a normal dagger so you can land the daze with 1/2 of your swings, which is all you need.

    Thus the debate and possibility that dancer's set effect applies to everyone in your party, not just you. There's no evidence either way for THAT one though.

  6. #6
    Melee Summoner
    Join Date
    Apr 2010
    Posts
    42
    BG Level
    1
    FFXI Server
    Ragnarok

    Quote Originally Posted by urat View Post
    Afaik there's evidence that the set effect of our AF3+2 comes from having the daze on the opponent, and has nothing to do with the additional effect, therefor Twilight knife does indeed still work with the +2 set, however you're other dagger has to be a normal dagger so you can land the daze with 1/2 of your swings, which is all you need.

    Thus the debate and possibility that dancer's set effect applies to everyone in your party, not just you. There's no evidence either way for THAT one though.
    I suppose the best way to test this would be to have 2 DNCs with set bonus available, have one with samba on and one without, and try all the possible combinations of who has +2 set on.

    This could eliminate all possibilities but the right one:

    1: Does samba set bonus require the samba user to be wearing the gear?
    2: Can other PT members receive samba set bonus?
    2b: Must DNC +2 gear be worn to receive samba set bonus by other PT members who aren't using sambas themselves?
    3: Does samba set bonus proc off the daze or off the samba effect?

  7. #7
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    OAT TP Overage losses are a myth and I'm sick of hearing about them.

    This is something floating about with all the talk and comparisons to OAT weapons with Merit WS.

    With a hypothetical weapon and WS and 6-hit and using numbers to give a useful TP:WS ratio that is skewed towards WS output:

    100 damage hits, 1000 damage WS.

    Clean cycle: 500 TP damage, 1000 WS damage. 1500 damage per cycle.

    Overage cycle: 600 TP damage, 1010 WS damage*. 1610 damage per cycle.

    *That's a 0.02 fTP per hit bump from 16.7 TP overage with Gorget and Belt on Resolution. Just to show it's not a huge boon to WS damage to go a hit over.

    I saw someone use the following method elsewhere a long time ago when the OAT magians were first introduced:

    It takes five overage occurences to have 'lost' one clean cycle's WS:

    1500 * 6 = 9000
    1610 * 5 = 8050
    I contended that if this only happens 50% of the time, OAT+DA gear for 50% DA rate, you get

    12*1500 vs 6*1500 + 5*1610
    18000 to 17050

    So one might think that's a 5.5% loss to overage overall from a 50% DA rate with a 33:66 TP/WS ratio. This is what I thought for a while, except:

    You never 'lose' anything that will never change a result!

    If you have overage potential from DA on any given cycle your rounds per WS was already closed anyway, so it can't be said that a DA on a last hit is ever costing you something.

    The only way those overages happened is a DA on the last hit of the round. This means that the rounds to TP was the same. You hit twice when you still needed to hit at least once anyway.

    These last hit DAs can't be construed as lost WS potential!

    There's a big flaw to thinking you 'lost' a cycle because of DA overages. Let's say that you had 60 delay (One second).

    The clean 9000 damage took 30 seconds. The 8050 damage took 25 seconds. The latter still has five more rounds to make, so really it gets the same number of cycles in.

    9000 vs 9550 damage.

    Overage from DA does not lose you anything! Those hits are tertiary and are never 'lost potential'. They are pure bonus TP phase damage while also slightly increasing your WS damage. If you have overage from DA your number of rounds to WS was the same, so you got an extra hit in for free.

    Stop trying to shoot down OAT weapons with this 'TP overage/overflow' myth. Even when misapplied the 'losses' are minimal, but also functionally nonexistent. This does exist for OA2-4 weapons, but those have vastly more varying round counts anyway.

    Mild crosspost for the sake of all the WARs and DRGs that don't read my usual forum and have seen discussions on 'DA TP overflow' or 'OAT TP overage' in their own circles.

  8. #8
    First invited, last in the zone.
    Join Date
    Sep 2008
    Posts
    1,449
    BG Level
    6
    FFXI Server
    Lakshmi

    The "Overages" occur if you naively calculate rounds/ws averages from swings/round average.

    As an example, say you have a 3 hit (not realistic, but it makes the numbers dead simple), and a 2-4 weapon with a 40% single attack rate and an average of 2 hits per round.

    A naive person will go "average 2 hits per round, I need 2 hits between each ws, obviously I will WS on average after a single attack round!" and that's where the notion of "wasted attack rounds" comes in, because 40% of the time you'll need 2 rounds, so you average 1.4 rounds/ws, not 1.


    If you do the probabilities out fully it's never an issue, but it will *always* show up if you try to do calculations based off just average hits/round.

  9. #9
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    Quote Originally Posted by Raelia View Post
    This does exist for OA2-4 weapons, but those have vastly more varying round counts anyway.
    Not to be rude, wait no yes I'm being rude: RTFP next time.

    Plus you're talking about proper calculation of effective hit builds based on probabilistic tiers, which is my forte, not TP overage/overflow. When someone says "OA2-4 has a 2.0 average hits per round" can you guess who they're quoting?

    This is specific to OAT weapons and high DA rate Warriors. Double attacking on the last round of a cycle does not impair damage at all, but is entirely bonus damage because it wouldn't have benefited your effective hitbuild anyway.

    It's people trying to say "You have a 40% chance of wasting a swing past 100TP and losing a hit towards another WS" that are driving me nuts now.

  10. #10
    Flowery Twats
    Join Date
    Jul 2008
    Posts
    3,583
    BG Level
    7

    This is a post out of ignorance, a genuine question, not debating you.

    When people make gear comparisons and use everything in terms of "% DoT", isn't tp overflow a factor? If im comparing 50% Double Attack to something else, then I'll need to concider:
    1) It's affect on TP damage
    2) It's affect on WS frequency

    Obviously the first point has nothing to do with TP overflow. But regarding it's effects on WS frequency, changing it to a standard "% DoT" unit so that it can be compared to other stats like attack/accuracy/haste/crit/etc, wouldn't it be innaccurate to say "every time you DA, you shave off a hit to your climb to 100%TP"? Because If that 50% DA procs on any hit except the last, then yes, it will save you a hit, but if it procs on the last, it wont. Thus saying 50% DA is equal to 50% increase in WS frequency" couldn't be accurate surely.

    A DA on a last hit is not "lost potential", but surely has to be accounted for when calulating the boost of OAT/DA in terms of "% DoT" to compare it to other stats. If you have a 5hit setup, then wouldn't you agree that a double attack proc would decrease your swings to 100%TP only on the first 3 out of 4 swings, and thus only equals 50% WS frequency boost 75% of the time?

  11. #11
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    I greatly value genuine questions.

    What precisely I'm saying here is that OAT and high DA rate TP overflow is actually a positive factor on overall damage, but is being parroted as a negative instead. The sneaking misconception is that any multihit weapon is wasting swings by going past 100TP when this is not equivalent to going past 100 TP with excessive rounds of a vanilla weapon. Going over because you multiattacked is not the same as going over because you derped and didn't hit your WS macro.

    Anyway that's why you need a properly calculated 'Effective Hitbuild' as Foldypaws mentioned. He's right about that. You can't actually have 1.5 hits in a round, and 2.66 isn't the true average rounds per cycle.

    We can get into that here though. Let's say you have a 5-hit build with an OAT weapon and a little extra DA for a smooth 50% double attack rate (not actually 10% DA, but that's the other thread...)

    There are three 'cases' that need to be assigned probability.

    1. Two double attacks (Two rounds)
    2. A double attack and two single attacks in any order. (Three rounds)
    3. Four single attacks. (Four rounds)

    You might ask 'What about a single followed by two doubles' and that gets into exactly what this thread is about. The second double doesn't matter, it's still a three round cycle and only the probability of the first two rounds matters (or three, as in the third case).

    So the chances of #1 are simply 0.50^2 or 25%.

    The chances of #3 are 0.50^3, 12.5%, because you must have three single attacks in a row and anything after will always still be a four round cycle, again the very subject of this thread.

    You can't have more than four rounds or less than two rounds, and the two and four round cases are exclusive to what was already calculated, so the remainder is actually the chance of a 3 round cycle and all possible permutations.

    100% - 25% - 12.5% = 62.5% chance of a 3 round cycle.

    Now let's weigh everything together:

    0.25*2 + 0.625*3 + 0.125*4 = 2.875 average rounds per cycle.

    Interestingly I'll have to rework OA2-4 since we found a more appropriate distribution set in the other thread:

    40/30/18/12, 5-hit
    Spoiler: show
    One round
    Only one way
    .12

    12%


    Two rounds
    3+any
    2+2

    If first is triple, or a single or double is followed by triple or quad, or two doubles
    .18 + (.4+.3)*(.18+.12) + .3^2

    18 + 21 + 9

    48%

    Three rounds
    Will be calculated from other results subtracted:

    33.6%

    Four rounds

    Only one option again
    1+1+1+any

    .4^3

    6.4%

    .12*1 + .48*2 + .336*3 + .064*4


    2.344 Rounds per cycle average.

    But in the end this isn't to be attributed to TP Overflow. This is the diminishing returns of Double Attack and other multiattack stats at work. Something that has been understood for a long time.

    People using 'TP Overflow' as an excuse to scoff off OATs being devastatingly powerful with Merit WS are misappropriating the term completely.



    And for my own reference: Average rounds per cycle for OA2-4 Scythe with 10% DA, parsed raw 32.6/39.2/16.5/11.5 distribution:
    Spoiler: show
    1
    11.5%

    .165 + (.326+.392)*(.165+.115) + .392^2
    16.5 + 20.104 + 15.366

    2
    51.97%

    3
    33.07

    .326^3

    4
    3.46%

    0.115 + 1.0394 + .9921 + 0.1384

    2.2849 Rounds/cycle.

  12. #12
    Flowery Twats
    Join Date
    Jul 2008
    Posts
    3,583
    BG Level
    7

    Cheers for the clarification there, I guess maybe a lot of the "TP overflow is a waste" stuff stems from using a too simplistic view of trying to calulate ##% Double Attack as ##% WS frequency increase, instead of calulating the difference of your average rounds to 100% TP.

  13. #13
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    Quote Originally Posted by noodles355 View Post
    Cheers for the clarification there, I guess maybe a lot of the "TP overflow is a waste" stuff stems from using a too simplistic view of trying to calulate ##% Double Attack as ##% WS frequency increase, instead of calulating the difference of your average rounds to 100% TP.
    Yeah well I had to blist and ragequit to not have to put up with this crap from the person I was arguing with about it in-game, and then found more of it on FFXIAH forums shortly after, and then put the OP together over on alla and I decided to extend it here. End up putting together a decent baseline for calculating true rounds per cycle of 5-hit OA2-4 weapons from parsed or estimated distributions (I just checked reduction in 3-round cycles previously because it was half the work and produced useful comparisons).


    Quote Originally Posted by Foldypaws View Post
    The "Overages" occur if you naively calculate rounds/ws averages from swings/round average.

    As an example, say you have a 3 hit (not realistic, but it makes the numbers dead simple), and a 2-4 weapon with a 40% single attack rate and an average of 2 hits per round.
    I made this mistake in the past and chewed on my parse data for a while before running the tiers and putting together a 2.4 R/C value that I never really used anywhere. This was with my old 40/30/20/10 estimate. I then knew the 2.0 average producing a 2.0 R/C on 5-hit to be false and stopped posting it, but by then others may have picked it up.

    I instead decided OA2-4s were just too volatile to calculate in this manner anyway, since I was dealing with all of 5-hit OA2-4 Scythe, 6-hit 2-4 Scythe and GS, and a 7-hit GS applications, hence my '3-round cycle reduction' method I mentioned instead, since that's the case that changes most radically when you add Apoc atma and other things into the mix. Largest reduction in 3-round cycles was then the fastest way to find the largest reduction in Effective Hitbuild from Atmas, maybe I'll detail it here but the application is very narrow (OA2-4 GS, GAxe, Scythe, and maybe Polearm).

    Interesting though that the exact method needs to be altered pending details in the other thread regarding DA overriding OA2-4's triples and quads as it definitely does Kraken's Triples and higher, so things like this may require another look. Short version is you calculate the chances of having consecutive singles or doubles as your Single and Double rates are reduced by increased other multihit sources like DA or TA from Atma, generally only applicable to 5-hit but works on 6-hit too, but with DA from gear being shown to override multihit weapons completely this may actually be working against reducing R/C for 6-hit and higher builds.

  14. #14
    First invited, last in the zone.
    Join Date
    Sep 2008
    Posts
    1,449
    BG Level
    6
    FFXI Server
    Lakshmi

    Quote Originally Posted by Raelia View Post
    Not to be rude, wait no yes I'm being rude: RTFP next time.

    Plus you're talking about proper calculation of effective hit builds based on probabilistic tiers, which is my forte, not TP overage/overflow. When someone says "OA2-4 has a 2.0 average hits per round" can you guess who they're quoting?

    This is specific to OAT weapons and high DA rate Warriors. Double attacking on the last round of a cycle does not impair damage at all, but is entirely bonus damage because it wouldn't have benefited your effective hitbuild anyway.

    It's people trying to say "You have a 40% chance of wasting a swing past 100TP and losing a hit towards another WS" that are driving me nuts now.
    I did "read the fucking post", I just don't agree with it. This isn't, in fact, limited to 2-4 weapons like you say, It comes up with just plain OA2 as well, I just didn't use OA2 because you can't have a 2.0 average on an OA2 weapon and I wanted to be lazy/simple with the math.

    In a sort of back of the envelope way, if your swings/ws is high, you can treat multi hit as a flat bonus to all of your melee damage. (I'm assuming no DA, just pure OA2 so I don't have to worry about DAs on WS affecting damage)

    As your swings/ws goes down (either from Save TP, or Store TP, or Regain or whatever, or more multi-hits), more of your hits are "last hits of the round" , on which a multi hit will only be a bonus to TP phase damage, and *not* WS frequency. So if you've got a 60/40 WS/melee damage split, *some* of that x% damage increase from multi-hit is only applying to the 40%, not the 60%.

    You show this above when you math out the 5 hit. Average rounds/cycle is 2.875 instead of the 2.66 you get from (4/1.5). In other words, about 8% of your multi-attacking is going *only* to TP phase damage, and not increasing your WS frequency, So if you want to calculate the damage boost, you get (.92*.6+.4)=95% of the "naive" damage estimate, or a 50% OaT only boosts your damage by 42.5%, not 50%.

    If you wanted to compare OaT to some other weapon, (one with a giant attack bonus, say, but the same damage) It wouldn't need to boost pDif by 1.5 to win, it would only need to boost by 1.42 (or some other way of boosting damage that added up to a 42% boost that was more realistic than that much attack).

    This is not the same thing as the diminishing marginal returns of DA (because as I indicated above, it can be based on Save TP or Regain and not just multi-hit) but it is the basis of the notion that extra multi hit can have" "wasted" rounds that contribute less to WS frequency.


    Of course, this is still just back of the envelope calculations, and if there's a DPS spreadsheet for the job, it's going to do way better with less effort. (Especially because if you've got a crit or damage varies with TP ws (or even Acc in some rare cases), the extra TP is going to have a non-zero effect... etc.)



    Anyway, tl;dr version. It is a real diminishing marginal return, it's not the same thing as the general DA diminishing marginal return, but if people are trying to claim it's a penalty rather than just a diminishing marginal return, they are dumb and I'm not trying to say otherwise

  15. #15
    An exploitable mess of a card game
    Join Date
    Sep 2008
    Posts
    13,258
    BG Level
    9
    FFXIV Character
    Gouka Mekkyaku
    FFXIV Server
    Gilgamesh
    FFXI Server
    Diabolos

    Removed posts, but:
    - Did not hand out infractions (I don't do that)
    - Did not remove posts starting the initial ad hominem battle since that would remove math-relevant discussion
    - Remind you that ad hominems will meet ad hominems (Talking down people on the internet only strengthens retaliation)
    - Despite BG's "Tough it out" attitude, the math forums have stricter guidelines

  16. #16
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    The average number of hits per round (TP gain per round) determines your average overflow. If you have 50% DA with a 5-hit (1 from WS), then you need four more hits to get 100TP and WS again. The simple way to calculate this would be 4 / 1.5 = 2.66666 rounds to 100TP. However:

    Ways to get 100TP -
    2 rounds: DA + DA (25%)
    3 rounds: SA + DA + SA/DA (62.5%)
    4 rounds: SA + SA + SA + SA/DA (12.5%)

    2*0.25 + 3*0.625 + 4*0.125 = 2.875 rounds to 100TP

    So people would call the .21 rounds difference "Overflow." This isn't to say that the TP is totally lost. It benefits you through the TP modifier of the weaponskill (as you said), along with the "regular" differences between your x-hit build and 100 TP. If you look at the guts of Motenten's DPS spreadsheets, he has substantial probability tables devoted to this.

    Some people also use the term to mean "TP lost when I'm not paying attention." They would be correct in stating "Builds that give more TP/hit lose a larger fraction of a WS every time I play inefficiently."


    As far as approximating overages with large hits/WS and low swings/round, you could say:
    * If X% is my DA rate, I will get the final hit X% of the time with a DA.
    * Therefore, (1-X)% of the time I will get the final hit with a new attack round, and X% of the time that round will DA.
    That case would predict Overflow (1-X)*X to be 0.25 swings with a 50% DA rate. Unfortunately, almost every 1H job dual wields. That means every round is from 2-4 hits, not from 1-2. It makes it much harder to calculate. Also, hit rate is never 100%.

  17. #17
    Lv.99 Mjollnir
    Join Date
    Sep 2008
    Posts
    254
    BG Level
    4
    FFXI Server
    Gilgamesh

    Quote Originally Posted by Byrthnoth View Post
    So people would call the .21 rounds difference "Overflow." This isn't to say that the TP is totally lost. It benefits you through the TP modifier of the weaponskill (as you said), along with the "regular" differences between your x-hit build and 100 TP. If you look at the guts of Motenten's DPS spreadsheets, he has substantial probability tables devoted to this.
    I like how you explained the 0.21 difference. Is this the same as what I found on KC?

    Quote Originally Posted by RyaWHM View Post
    Now if you were to run these x-hits for KC's "dummy" average (3.82 swings; so 4.8 w/DW), you get:
    16-hit: 3.33 rounds (vs 3.74 real, from above)
    15-hit: 3.13 rounds (vs 3.54 real, from above)
    14-hit: 2.92 rounds (vs 3.33 real, from above)
    13-hit: 2.70 rounds (vs 3.13 real, from above)
    12-hit: 2.50 rounds (vs 2.91 real, from above)
    11-hit: 2.29 rounds (vs 2.71 real, from above)
    10-hit: 2.08 rounds (vs 2.51 real, from above)
    9-hit: 1.88 rounds (vs 2.32 real, from above)
    I did look at spreadsheets but they did not accommodate KC, and I did not know how to bastardize them enough to make it work. If I can better understand it in the less-complicated "OAT-land" then perhaps it will be easier for me to understand KC.

    Please assist! Either here or in my KC thread, my question applies to both.
    Thank you so much :)

  18. #18
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    You can look at Post #5 in the spoilered sections for how I worked out OA2-4, but this still requires calculating the probability of all but one case which you can just difference for. Probably the only or at least best way to do the same for KC would be to simulate it, and from there you might as well put your whole Store TP analysis together in the same workup.

    Wish I could change the thread title to "OAT TP Overage is not a penalty and I'm sick of hearing it described as such".

  19. #19
    Lv.99 Mjollnir
    Join Date
    Sep 2008
    Posts
    254
    BG Level
    4
    FFXI Server
    Gilgamesh

    Quote Originally Posted by Raelia View Post
    calculating the probability of all but one case
    Yes, I actually generated all 117649 combinations (6 rounds) in a spreadsheet :)

    That's how I could easily grab portions of the data, complete with probabilities for each. I also saved off a new sheet for each round I added, so that I could run accurate probabilities and come up with those neat little columns for each round (%chance of each # of hits).

  20. #20
    Custom Title
    Join Date
    Nov 2008
    Posts
    1,066
    BG Level
    6
    FFXI Server
    Diabolos

    Crosspost relating to some personal speculation that ODD might not make up for using Quietus to activate aftermath with Redemption.

    Just for sake of argument, let's 5-hit both.

    A cycle is a 4.0 TP weight, some number WS weight.

    With 30% ODD, this goes up to 5.2 TP weight.

    With 50% ODD, this goes up to 6.0 TP weight.

    With Quietus you have a 3.0 fTP, with Entropy a 3.8 at 100tp, 4.55 at 300tp.

    Quietus: +85 STR, +15 MND

    Entropy: +60 INT

    So it's
    STR - 100+85
    MND - 70+15

    62+28, +90 WSC

    vs
    INT - 85+60

    +123 WSC

    163 base damage from 99 Redemption, 189 after fSTR

    Quietus WSC factor: 1.476
    Entropy WSC factor: 1.651

    Quietus WS weight: 4.428
    Entropy WS weight: 6.274, 6.893, 7.512 (100tp, 200tp, 300tp)

    So by using Quietus at 100tp you lose 1.846 on that WS and gain 1.2 per cycle afterward for 30 seconds of Aftermath.

    By using Quietus at 300tp instead of Entropy 3 times you lose 9.25 on that WS and gain 2.0 per cycle afterward for 90 seconds.

    By using Quietus at 300tp as an opener you lose 3.084 on that WS and gain 2.0 per cycle afterward for 90 seconds.

    502 delay, 65% 'nominal' haste, 5-hit (4 rounds per cycle), 2 second WS delay: 13.7 seconds per cycle.

    100tp: 1.846 / 2.19 cycles * 1.2 per cycle = You lose 70% of your ODD output just making up for using Quietus.

    300tp running: 9.25 / 6.57 cycles * 2.0 per cycle = Still the same 70% loss of ODD output making up for using Quietus if you're dumb enough to run up 300tp for a 50% refresh.

    300tp opener: 3.084 / 6.57 * 2.0 = Even best case you still lose 23.4% of your ODD just making up for that Quietus, and this doesn't even consider the Sekkanoki self-Darkness you could do instead:

    200tp Entropy SC closer = 2.465 loss PLUS a 6.893 Darkness loss, 9.358 total.

    300tp opener instead of Sekka Quietus -> Entropy and 100tp AF: (2.465 WS loss + 6.893 Darkness) / (6.57 cycles * 2.0 per - 2.19 cycles * 1.2 per) = 89% loss, so 99 Redemption brings 300tp opening Quietus back to beneficial.

    90 Redemption: 136 base, 159 after fSTR

    Quietus WSC factor: 1.566
    Entropy WSC factor: 1.774

    Quietus WS weight: 4.698
    Entropy WS weight: 6.741, 7.406, 8.072 (100tp, 200tp, 300tp)

    Losses: 2.043, 2.708, 3.374

    100tp: 2.043 / 2.19 cycles * 1.2 per cycle = 77.7% loss of ODD gains.

    300tp opener instead of Sekka Quietus -> Entropy and 100tp AF: (3.374 WS loss + 7.406 Darkness) / (6.57 cycles * 2.0 per - 2.19 cycles * 1.2 per) = 102.55%, meaning the self-SC and 100tp aftermath outweighs the 300tp aftermath by 2.55% and we've shown that 100tp refreshes are still valuable.

    So the best method, when capable with 90 Red, is indeed to sekka Quietus -> Entropy and screw 300tp aftermath, because even the two subsequent 100tp aftermath refreshes are still more valuable.

    Naturally this falls flat if you can't cleanly self-SC, but then 300tp AF isn't so far behind, but 90 Redemption is still only gaining a pathetic 2.68% total damage from ODD on 100tp refreshes!
    tl;dr 90 Redemption is actually better off starting with Sekkanoki Quietus -> Entropy over AM3 if possible, and it wouldn't take much more than a decent Miser's+Discipline setup for pseudo-3-hit or 4-hit to push the ODD gain below the loss of WS damage from using Quietus.
    Naturally, the FFXIAH tards are spewing crap all over this and using my old Diabolos snapshot as some means to 'discredit' me

    Starts here.