Whats better?

[el_tio]

I'm guessing the idea is that the faster the weapon, the less rounds it takes for you to see a guaranteed return from the haste. I.e. getting an extra attack round in (excepting killshot, which always = return from haste) or swinging before another person's attack round when otherwise you would've swung after.

Edit: Meh, nvm, I think he was maybe making a point about the roundings used in formula? No clue really...

[el_tio]

Go 5 DA, 5 berserk.

And why would KJ be better on harder mobs? Harder implies more defense or higher level. KJ doesn't up your pDIF. Raging Rush does. Use Raging Rush.

Well, with RR, on harder mobs, even though you get the attack bonus, you also get the lower crit rate, which might weaken the improved attack bonus from crits, even though you get relatively more attack from the crit. Leading to the base damage increase out performing a crit increase.

Though maybe my thinking's incorrect, because even though the mobs are higher I don't think AGI would change as significantly as I was thinking. Though even a relatively small increase in AGI would result in a drastic decrease in % of crits due to the way crits work.

So, I guess idk really. Disclaimer: I don't have KJ, just models other ppl have done and models I've done show them to be very very close in performance.

[Wolfknight]

Fuck I just did all the math and lost my post.

Gonna sum it up as fast as possible, 'cause I don't feel like redoing it right now.

Sam roll = +25% WS damage.
AH sam = +38% WS and TP damage from marches.
Chaos roll = +17% TP damage.
Better haste gear = +16% WS and TP damage from marches.
.

Fair enough but I think would be close still and I disagree that's minimal value chaos should given from a good cor, like I said with relic hat you get 25~41% w/proc and 16~25% w/o, so +16% seems like low amount to me, considering should be 25% usually.

[mdkuser]

I've been meaning to ask this for long time, I've never been able to convince myself that one is better than the other. My own considerations tell me set-up A wins but I don't now much about pdif. I use both depending on mood.

For SAM:

Set-up A:

Hagun, sword strap, usukane haramaki, rajas, brutal, usukane sune-ate.

Set-up B:

Hagun, pole grip, haubergeon +1, rajas, brutal, usukane sune-ate.

All other slots what they should be, same WS gear in both cases except grip.


Set-up A is a 6-hit build, has more accuracy, more dex, more str and more attack speed. set-up A has less accuracy dex str attack speed, but more attack, and more DA during WS.

My approximate maths:

Spoiler: show

Set-up A:

Edit:
Its just enough stp for a 6-hit build ( 54 stp in melee, 41 on WS). Your Hagun will have 436 delay, so 11.1 base tp instead of 11.5. Melee hits give 17.0 TP and WS hit gives 15.6 for a total of 5x17+15.6=100.6 TP. So sword strap works like haste: straight delay reduction w/o tp penalty.

You do the same dmg as you'd do w/o grip, but in 97% of the time it would take you. 1/0.97=1.031. It's 3.1% more melee and WS , so 3.1% more total dmg.

Set-up B:

Going from no grip to pole grip moves you from 15% to 17% DA. 1.17/1.15=1.7% more hits, thus 1.7% more tp and 1.7% more ws. The DA can proc during WS. A DA'd WS is increased by ~50% ( 2.875/1.875=1.5x) so 2% of your WS are increased by 50%, which leads to 1% more ws dmg.

Total: 1.7% more melee dmg, 2.7% more WS dmg.

But I didnt take the body swaps into account. Accuracy doesnt matter for set-up A, it will be 3.1% increase with or w/o caped accuracy. In uncapped accuracy set-up B will have 1-2 less acuracy, 2 less STR, 2 less DEX, but 10-11 more attack. So this should actually enhance melee dmg a bit. by how much idk cause it will be dependant on what you fight.. But im guessing it will be much less than 3.1% in the end.

I'd like serious maths, not guestimates like mine thanks.

[Miraun]

100 hits landed at 95% Acc takes 106 swings (106 x .95 = 100.7)
100 hits landed at 93.5% Acc takes 107 swings (107 x .935 = 100.045)

The only rounding I did was the round down function on the reduction of delay. Since we're arguing fractions of delay at high haste values, .655 of a swing is a significant difference.

[Max™]

K, consider it after 10,000 hits landed.
10,527*.95 = 10,000.65
10,696*.935 = 10,000.76

Is .11 small enough?

(95%/49%) 10,527 Swings takes:
500 Delay = 2,684,385 Delay
400 Delay = 2,147,508 Delay

(93.5%/50%) 10,696 Swings takes:
500 Delay = 2,674,000 Delay
400 Delay = 2,139,200 Delay

(500*10,527) - 2,684,385 = 2,579,115
(500*10,696) - 2,674,000 = 2,674,000 minus the above = 94,885, /500 = 189.77 rounds

(400*10,527) - 2,147,508 = 2,063,292
(400*10,696) - 2,139,200 = 2,139,200 minus the above = 75,908, /400 = 189.77 rounds

Still not sure what you're getting at...

5,348,000 - 4,278,400 = 1,069,600, difference of the base delays.
2,674,000 - 2,139,200 = 534,800, 1/2 the difference between the base delays.

If the lower delay got more out of it, shouldn't it have been more than half the difference, since the 400 would wind up at a smaller total delay as you seem to be implying?

[Max™]

50% Acc 0% Haste 100 Hits takes 200 Rounds
47.5% Acc 1% Haste 100 Hits takes 210 Rounds

150 at 0% = 30,000
150 at 1% = 31,185, 1,185/150 = 7.9

500 at 0% = 100,000
500 at 1% = 103,950, 3,950/150 = 7.9

80% Acc 10% haste = 100 hits in 125 rounds
79% Acc 11% haste = 100 hits in 127 rounds

Spoiler: show

150 at 10% = 16,875
150 at 11% = 16,954.5, 79.5/150 = .53

500 at 10% = 56,250
500 at 11% = 56,515, 265/500 = .53

150 at 20% = 15,000
150 at 21% = 15,049.5, 49.5/150 = .33

500 at 20% = 50,000
500 at 21% = 50,165, 165/500 = .33

150 at 49% = 9562.5, 150/37.5 = 4
150 at 50% = 9525

500 at 49% = 31875, 500/125 = 4
500 at 50% = 31750

Here is the point where Haste wins at 80% Acc:
12000 (40000) at 36%/80% for 150 (500) Delay
12001.5 (40005) at 37%/79% for 150 (500) Delay
^ longer delay for +1% Haste -1% Acc

11812 (39375) at 37%/80% for 150 (500) Delay
11811 (39370) at 38%/79% for 150 (500) Delay
^ shorter delay for +1% Haste -1% Acc

Note that it's the same exact amount, you're incorrect in stating haste benefits different delays differently in ANY way other than the absolute amount knocked off.

[Selim]

Scorpion Harness +1 VS. Dusk Jerkin for Octave Drk.
Basically 20 HP vs 12 accuracy - not sure how much of an issue accuracy would be with Madrigals and Feint. Lacking Homam and a little bit of Club skill compared to a 75 Kraken setup, so not sure if I should stick with SH to be sure. Drk still sitting at 37 for now so can't really do any tests myself, just looking to plan in advance.

Assuming the rest of the gear looks like this:
Octave Club | Wyvern Targe | - | Happy Egg
Aurum Armet | Ritter Gorget | Wyvern Earring | Cassie Earring
Body Piece | Dusk Gloves +1 | Blitz Ring | Bombqueen Ring
Gigant Mantle | Speed Belt | Sable Cuisses | Dusk Ledelsens +1

[Max™]

Just working it out for 1500 Hp, but 1500 > 1520 is 180 > 182.4 on SE Dmg (it is 12% when Enhanced, yes?). Assuming you get full Acc (10 from Sh +1 < Dusk body, it has 2 acc on it), say 90% Acc x 100 hits for 182 = 16380, 95% Acc x 100 hits for 180 = 17100.

Would need 94% Acc to hit 17108 for that figure, prolly go with SH +1, I may be totally wrong though, not a Drk.

[Arximiro]

Armada Mufflers vs. Dusk Gloves when Acc is at 90%

I assume Dusk is better, but what I want to know is by how much exactly? I despise Dusk so I'd like to stick to my Armadas if the margin isn't huge. Let's also assume 19% haste with Dusk, + the 15% from the spell, so 34 haste.

[tsola]

Question on H2H "tiers". What skill levels are they?

I'm basically wondering for my PUP that has 4 H2H merits (that won't change), if it's better to use Faith Torque or PCC.

For reference: 225 base skill+8from merits+7 from torque=240 total.

Thanks.

[FailureMidgard]

Fair enough but I think would be close still and I disagree that's minimal value chaos should given from a good cor, like I said with relic hat you get 25~41% w/proc and 16~25% w/o, so +16% seems like low amount to me, considering should be 25% usually.

I guess I coulda used a relic hat cor, but I have no earthly idea what the actual average would be there. Best case - I overstated the difference by a bit originally.

However, I am still dead on about the actual difference between a single bard and a single cor. If you add Byakkos + Usu Feet to the bard's sam, and take away V Belt/Dusk+1 from the cor's sam (or even better, give those to the bard's), the gap will inflate hugely. The cor's sam loses that last 16% dmg increase entirely, and then it flips to more than 16% in the other direction (that already had a huge lead).

A single bard is just way better than a corsair.

As far as StoreTP goes, it's "better" than haste on gear, because you're only needing to drop ~3 haste to get your next hit build. When you start looking at sam roll vs marches, you're losing way more than that.

I've been meaning to ask this for long time, I've never been able to convince myself that one is better than the other. My own considerations tell me set-up A wins but I don't now much about pdif. I use both depending on mood.

For SAM:

Set-up A:

Hagun, sword strap, usukane haramaki, rajas, brutal, usukane sune-ate.

Set-up B:

Hagun, pole grip, haubergeon +1, rajas, brutal, usukane sune-ate.

All other slots what they should be, same WS gear in both cases except grip.


Set-up A is a 6-hit build, has more accuracy, more dex, more str and more attack speed. set-up A has less accuracy dex str attack speed, but more attack, and more DA during WS.

My approximate maths:

Spoiler: show

Set-up A:

Edit:
Its just enough stp for a 6-hit build ( 54 stp in melee, 41 on WS). Your Hagun will have 436 delay, so 11.1 base tp instead of 11.5. Melee hits give 17.0 TP and WS hit gives 15.6 for a total of 5x17+15.6=100.6 TP. So sword strap works like haste: straight delay reduction w/o tp penalty.

You do the same dmg as you'd do w/o grip, but in 97% of the time it would take you. 1/0.97=1.031. It's 3.1% more melee and WS , so 3.1% more total dmg.

Set-up B:

Going from no grip to pole grip moves you from 15% to 17% DA. 1.17/1.15=1.7% more hits, thus 1.7% more tp and 1.7% more ws. The DA can proc during WS. A DA'd WS is increased by ~50% ( 2.875/1.875=1.5x) so 2% of your WS are increased by 50%, which leads to 1% more ws dmg.

Total: 1.7% more melee dmg, 2.7% more WS dmg.

But I didnt take the body swaps into account. Accuracy doesnt matter for set-up A, it will be 3.1% increase with or w/o caped accuracy. In uncapped accuracy set-up B will have 1-2 less acuracy, 2 less STR, 2 less DEX, but 10-11 more attack. So this should actually enhance melee dmg a bit. by how much idk cause it will be dependant on what you fight.. But im guessing it will be much less than 3.1% in the end.

I'd like serious maths, not guestimates like mine thanks.

I'm not going to do exact math, I don't think anyone on these forums has an actual good approximation on the current pdif formulas.

That being said, the WS difference is +.26% total damage (Assuming 65/35 split). The bottom set would need to do .7% more TP damage than the top to roughly tie, meaning another 2.1% increase in damage just from -2 str, -2dex, +12 attack.

Assuming best case scenario, both are accuracy capped, the dex doesn't affect crit rate, and -2str doesn't affect your fSTR (and only lowers attack by 1). 11 attack would be 2.1% at 524 attack. If you are either /war or have a minuet or chaos roll, you should be averaging higher than that with Pizza really.
Add in the fact that the dex probably affects your crit, fSTR varies from mob to mob potentially, so it will be +1 dmg some of the time, and the extra 1-2% accuracy will matter still on flashes/warm ups/wind walls, and I think we can pretty confidently say that there is no way for the Pole Grip setup to actually be considered better. When we assume best case scenario for it, it still isn't really better except in low buff situations.

All that being said...The difference is so marginal. I've considered switching it myself, but I'm not entirely sure that it's worth the -1 inv space to own a Sword Strap, because I'd still need Pole Grip to keep my true 5hit Pole setup. Then again, I could probably just sell Haub+1 and get a lot of gil back if I did decide to make the switch.

[Ashokan]

Armada Mufflers vs. Dusk Gloves when Acc is at 90%

I assume Dusk is better, but what I want to know is by how much exactly? I despise Dusk so I'd like to stick to my Armadas if the margin isn't huge. Let's also assume 19% haste with Dusk, + the 15% from the spell, so 34 haste.

Using the 1acc ≈ .5% accuracy

Without delay, it looks a little silly because normally you'd use something like ( time / ( delay * ( 1 - haste ) ) ) to represent number of swings over a given time. Instead, I'm just using 1 for time & delay, which will just give a % of increased swinging speed 100% being non-hasted speed, then just using the percent as number of swings.

Dusk:
90% acc
1 / ( 1 - 34% haste ) = swinging @ 152% of non-haste speed
Use the percentage as a swings = 151.5152
Connecting swings: 136.36

Armada:
92.5% acc
1 / ( 1 - 31% haste ) = swinging @ 145% of non-haste speed
Use the percentage as a swings = 144.9275
Connecting swings: 134.05

Conclusion: 136.36/134.05 = Dusk give 1.7% increase in connecting swings.

Although it's deceiving, b/c if you add march or hasso, the numbers will continue growing, full hasted, probably around 5%.

[Miraun]

Note that it's the same exact amount, you're incorrect in stating haste benefits different delays differently in ANY way other than the absolute amount knocked off.

You're right... Just rechecked my old spreadsheet function... I only updated delay in one of the two instances in the equatoin. Thank you for checking my math, I'm glad you did...

I reran the numbers to double check my ACC argument though, since I was wrong on delay...

From 93.5 to 95% ACC, Turban will be a better once @39% Haste.

The lower your ACC, the more haste is required for Turban to be better.

@ 50% versus 48.5%, The two head pieces are equal at 60% Haste.

Working on mathing out the equation to graph it...

(1-Rounddown(1*(Haste+4),0)/(Hit%+.015) = (1-Rounddown(1*(Haste+5),0)/Hit%

Solve for Haste ><

[FailureMidgard]

You're right... Just rechecked my old spreadsheet function... I only updated delay in one of the two instances in the equatoin. Thank you for checking my math, I'm glad you did...

I reran the numbers to double check my ACC argument though, since I was wrong on delay...

From 93.5 to 95% ACC, Turban will be a better once @39% Haste.

The lower your ACC, the more haste is required for Turban to be better.

@ 50% versus 48.5%, The two head pieces are equal at 60% Haste.

Working on mathing out the equation to graph it...

(1-Rounddown(1*(Haste+4),0)/(Hit%+.015) = (1-Rounddown(1*(Haste+5),0)/Hit%

Solve for Haste ><

The math has been done a million times, when you say 39, you mean 39 post turban, or the 34 number everyone has figured over and over again? Meaning in any merit/tp burn situation that you have a bard, turban is always better unless your gear is just awful. Oh ya, and any time you eat pizza (which should be pretty much anything that isn't birds).

Wish people would just accept Turban is better in any situation worth mentioning.

[Ashokan]

Rounddown(Delay*(Haste),0))

Just an unrelated question about this. Have we proven delay is truncated? It seems like if it's anything it'd be rounded up or put into applied as fraction. I don't think it's been proven either way, since it seems like it'd be impossible to test.

I realize everything in this game is truncated, rounded, in a fraction, etc... and perhaps that's the answer to my own question, but I'm interested in if there is an accepted conclusion.

[blkswrdsmn]

Nuking as 75 Rdm/Nin ~> Aura Statues (ones near Ulli pop) my Z mitts are out performing my Yigits... I thought it was odd since it is generally accepted that for T3 nukes Yigit > Zenith. Unresisted doing 608 (Z) vs. 604 (Y).

HQ Ice/Bugard +1/X/Phantom
AF1head/ele torque/phantom +1/Moldy
errant/hands in question/snow/snow
merciful/penitents/mahatma/goli

Anyone have the math answer? I suppose there is a point where more MAB will give an extra boost compared to int.

(In this case the applicable diff is 5MAB vs 5INT + 2MAB)

Edit: Mainly tossing this out here b/c someone recommended nuking in Yigit over Zenith last page of posts...

[Dathus]

Just an unrelated question about this. Have we proven delay is truncated? It seems like if it's anything it'd be rounded up or put into applied as fraction. I don't think it's been proven either way, since it seems like it'd be impossible to test.

I realize everything in this game is truncated, rounded, in a fraction, etc... and perhaps that's the answer to my own question, but I'm interested in if there is an accepted conclusion.

If they treated all other values the way they do STR, then we can tell that they would round down(as you need 2 STR = 1 ATK, and when you have 3 STR it doesn't give you 2 ATK). Just a small example, but could be others? Maybe?

[Miraun]

Just an unrelated question about this. Have we proven delay is truncated? It seems like if it's anything it'd be rounded up or put into applied as fraction. I don't think it's been proven either way, since it seems like it'd be impossible to test.

I realize everything in this game is truncated, rounded, in a fraction, etc... and perhaps that's the answer to my own question, but I'm interested in if there is an accepted conclusion.

That was rounding down the reduction... same thing as rounding up the final number.

[mdkuser]

I'm not going to do exact math, I don't think anyone on these forums has an actual good approximation on the current pdif formulas.

That being said, the WS difference is +.26% total damage (Assuming 65/35 split). The bottom set would need to do .7% more TP damage than the top to roughly tie, meaning another 2.1% increase in damage just from -2 str, -2dex, +12 attack.

Assuming best case scenario, both are accuracy capped, the dex doesn't affect crit rate, and -2str doesn't affect your fSTR (and only lowers attack by 1). 11 attack would be 2.1% at 524 attack. If you are either /war or have a minuet or chaos roll, you should be averaging higher than that with Pizza really.
Add in the fact that the dex probably affects your crit, fSTR varies from mob to mob potentially, so it will be +1 dmg some of the time, and the extra 1-2% accuracy will matter still on flashes/warm ups/wind walls, and I think we can pretty confidently say that there is no way for the Pole Grip setup to actually be considered better. When we assume best case scenario for it, it still isn't really better except in low buff situations.

Thats what my intuition told me to do. It's hard to convince other sams that the sword strap works with usu body. In high buffs situations I use usukane body, in einherjar mainly, I always have at least 2 bards and sometimes 2 bard+cor., and always go /war. In limbus salvage I use Hauby+1.