Screwed again: Proposed THF job adjustments.

[Motenten]

But I don't think this even accounts for triple attack rounds - the faster you hit the more chances you have for double/triple/quadruple attack to proc.

Unfortunately, that's not a valid way of looking at it. If you make 10 attack rounds, and get 2 triples in that time, it doesn't matter if each round is 1 second apart or 10 seconds apart, you still do identical damage. You don't get "more" chances to triple attack as you reduce delay, you just compress the damage you did into a smaller period of time.

Yes, you can technically view it as more chances to triple attack within a given period of time, but that's only because of the greater number of attack rounds to begin with.

I'm probably not explaining it well. Triple attack/etc chances do not scale with time, they scale with number of attack rounds. Attack rounds scale with time. Saying that your improvement is because of more triple attacks is viewing the damage schema at entirely the wrong level.

[Economizer]

Yeah, I'm not trying to say the extra damage is due to double/triple/quadruple/occ. attk. X or anything like that, merely that haste amplifies the effects of them more (in this game, what haste doesn't amplify is probably a shorter list then what it does).

Saying that your improvement is because of more triple attacks is viewing the damage schema at entirely the wrong level.

Lets say you have a fight that goes on for a somewhat limited time. You either get those extra rounds due to haste/delay reduction or you don't. If you have some form of multihit trait or weapon then the damage difference is further amplified from someone who doesn't have haste.

Of course, I'm not the legendary Professor Mathematical so I'm either misinformed in my own thinking process or performing some miscommunication in my thinking process. At this point in time I'm erring on the second one however.

[Babekeke]

Incorrect. It doesn't increase TP gain *as much* as it used to, but it does absolutely still increase TP gain.

Delay per weapon    TP/Hit      TP/Second
180                 5.0         1.667
175                 4.9         1.680
170                 4.9         1.729
165                 4.8         1.745
160                 4.8         1.800
155                 4.7         1.819
150                 4.7         1.880
145                 4.7         1.945
140                 4.6         1.971
135                 4.6         2.044
130                 4.5         2.077
125                 4.5         2.160
120                 4.5         2.250

Is this only true for delay under 180? I recall seeing a graph showing delay to TP a couple of years ago and it showed that weapons under 180 produce more TP, but don't recall if it was linear above that? Also, is there a floor at 120, or can TP gained go below 4.5?

[Arcon]

Of course, I'm not the legendary Professor Mathematical so I'm either misinformed in my own thinking process or performing some miscommunication in my thinking process. At this point in time I'm erring on the second one however.

You're correct with that statement, and while you get a higher total bonus with more Haste (or less Delay in general) you get the same percentual bonus.

Is this only true for delay under 180? I recall seeing a graph showing delay to TP a couple of years ago and it showed that weapons under 180 produce more TP, but don't recall if it was linear above that? Also, is there a floor at 120, or can TP gained go below 4.5?

TP always behaves linearly, but with different segments as it's defined partially in certain intervals. Overlapping intervals may produce weird behaviour when trying to calculate the percentual bonus (one such inflection point is at 180 Delay, so if you compare the interval [150;200] you'll get non-linear results). But for delay under 180, this is definitely true.

[Byrth]

Of course, I'm not the legendary Professor Mathematical so I'm either misinformed in my own thinking process or performing some miscommunication in my thinking process. At this point in time I'm erring on the second one however.

Well, if you gain 15% Triple Attack (from 0% TA/DA/QA/OAT/OAwhatever), it will increase your TP gain by 30%. It is true that you will see more Triple Attacks if you have more Haste, but they will still represent the same percentage increase in your TP gain. Thus, the impact of Triple Attack is independent of weapon delay.

This "decreasing delay increases TP gain" is one of the reasons why my Dancer (95 delay per hand) gains TP so much faster than my Thief (~122 delay per hand). All the extra Store TP and Haste Samba also help quite a bit.

[Aana]

I always think of it like dice. Everytime you attack, you roll a 10 sided die. If you roll a 1 you triple. 2-10 and you single. Pick up dice, roll it, see what happens. 10% of the time you triple attack.

Adding haste just makes you attack more, so you grab more dice. Now you pick up 5 dice. or 10, or 999. No matter what, you will still only triple 10% of the time. Addin more dice doesnt 'enhance' triple attack at all. You can roll 1 dice 100 times (it will take a while) or pick up 100 dice and roll them all at once (SUPER HASTE GOO!). Either way, the only way to make triple attack better is to say "Triple attacks now proc on rolling both a 1 and a 2" (20% chance to triple).

In this way, triple attack is completely independant of how many dice you roll. Its always a 1/10 chance to go off. Adding more dice doesnt change that. It just makes the process of rolling dice take less time (like haste makes attacking take less time). It doesnt actually affect your triple attacks. Roll 1 dice slowly 100 times, or m100 dice at once, it wont change how many times you roll a 1.

This "decreasing delay increases TP gain" is one of the reasons why my Dancer (95 delay per hand) gains TP so much faster than my Thief (~122 delay per hand). All the extra Store TP and Haste Samba also help quite a bit.

Little tangent that I have always wondered about. When you play dnc, does your style use steps/flourashes pretty much as soon as their up (assuming a pretty constant melee scenario), or do you just kinda ride haste samba+saber dance into the sunset. While its true dnc has a good chunk of atk speed over thf, i wonder how much of that is eaten up by spaming SOOO many JAs (compared to thf). I remember back in the old days of colibri parties when people started mathing in WS delay that soboro stopped beating hagun and became a wash because so much time was lost in the animations spamming so many weaker WS compared to hagun.

I just wonder if the constant barrage of dnc steps/flourishes messes with all that absurd speed with stop and go pauses. While, comparitively, thf has virtually no JAs.

2JAs a minute at best with SATA. Bully every 3. WS when TP. In a 3 minute cycle thf does (at most)
* ~7-8 JAs (3SA, 3TA, 1bully, 0-1 ACharge) and however many WS we can. Feint once in a blue moon if needed.

Compared to 12steps, 6reverses, ~1.5 of a flourish 3. ~1.5 haste samba. 1 saber dance. X for presto(iunno much about dnc post 75 if you use this or not, but seems to net gain of zero since presto+step=2 steps=same JA delay). THEN you add in the WS you get from all that DW/Samba speed and reverse flourishes. Good god thats a crapload of JAs.
~22+WSs in a THREE Minute period.

thats ~14 more 1 second JA delays in a 3 minute period (and however many mroe WS dnc is capable of doing on top of that). just an even 15 second/3 min is 8.3% of your melee time vanished.

Anyway, just wondering how JA intensive dnc really is. Not familiar enough with it to say what teh ost 'efficient' play style is, but it makes me wonder. The soboro vs hagun was the 1st example that adding JAs turned a clear winner (soboro) into a wash at high haste values. At a certain break point of haste and JA spam, less becomes more. Thf is definately on the 'simple' side. Curious if all the crazy JA spam on dnc ends up hurting it. Obviously capped haste things change and dnc needs to start toning down DW gear when it becomes a liability and thf ends up at the same 80% delay cap, so then its just TA vs saber dance and im sure at 80% zerg mode the JAs are really painful, but dnc and thf arent exactly high end zerg jobs, so we can stick to basic buffs :P

[Byrth]

The real answer is that it's very situational, but most of the time when I am playing Dancer (at high Haste lowman or when duoing with a mage mule against trash monsters), it is not worth using Steps/Flourishes at all, in my opinion. Obviously there are times when you want to use Steps/Flourishes even though it hurts your overall damage (like in Dynamis), but it is not very often. That is why I keep advocating for reduced JA delay. The job (played properly) is basically just Thief with more Dual Wield, Haste Samba, and Saber Dance (= Triple Attack, but you have to activate it) in most of the situations where it is useful.


When you do the comparison, you need to account for the fact that things like Presto+Step+Flourish+WS all end up chained and don't actually take an additional 2 seconds each. However, self-skillchaining will overkill everything Dancer fights or be impossible (Voidwatch). Box -> Wild -> Exenterator may be our best use of time, but even that isn't always useful.


If I was in an alliance or even full-party situation where Dancer could survive, I would obviously spam Steps for the sake of the group regardless of my own damage. Weasling your way into a Voidwatch alliance is the only way to get yourself into this situation though, and I'm usually so bitter to be in the mage party (inevitable) at that point that I don't really care anymore. Hell, most of the time mages don't even bother keeping Dia II on monsters, and that's just barely weaker than a level 5 Sluggish Daze and takes almost no time to cast.

I think I went into it more here.

[Motenten]

Is this only true for delay under 180? I recall seeing a graph showing delay to TP a couple of years ago and it showed that weapons under 180 produce more TP, but don't recall if it was linear above that? Also, is there a floor at 120, or can TP gained go below 4.5?

As Arcon mentioned, the scale changes at various delays, but the most interesting portion is 180 and below. There's a graph on wiki for it, I think.

There is no known lower bound. I just ran it down to 120 delay since that seemed like a good place to stop.

The lowest practical delay that I can come up with would be:

Nin (DW5): 35%
Suppa: 5%
Iga+2 head: 5%
AF body: 5%
Relic+2 legs: 7%
Nusku Sash: 5%
Iga Earring: 1%

Total: 63% DW

Edit: Actually, could add Blitzer's Roll to that for another ~15%, 78% total (just shy of the delay cap).

Weapons: 150 delay daggers (eg: Hornetneedle, Oneiros Knife, etc)

Effective delay: 55
TP/Hit: 3.9
TP/second: 4.216

Effective delay: 33
TP/Hit: 3.7
TP/second: 6.727

[Byrth]

Does Blitzer stack additively with Dual Wield? I was under the impression that they were separate categories.

[Ophannus]

Would be sweet if our merit 2hr ability gave us Triple Attack+100% for 30 seconds-1min.