Researching THF and subs.

[DelennFFXI]

Thank you for testing Ruic!

Also, this explains why my partial static monk can keep up in TP with me (brown belt, +8% haste, etc).

Bottom line, haste is good

Also, I'll ask my monk friend to test TP per fist with and without haste gear for more confirmation.





Just to re-summarize then: (copied and pasted with haste clarification)

• TP gained per hit is proportional to weapon delayTP is calculated before haste/slow effects are resolved
• The minimum TP from a swing is 5 (the infamous 5 TP floor)
• Multi-hit WSes only provide 1 additional TP for successive hits (from patch)
• A constant amount of TP is gained from taking any physical dmg (2 I think, 0 dmg = no TP)

Ways to increase TP rate:

• 5 TP Floor: Attack fast enough to warrant <5 TP, but get 5 TP anyway
• Multi-Attack Traits: Double Attack, Triple Attack
• Random Speed Up Trait: Rapid Shot
• Multi-hit Job Ability: Barrage
• Misc TP Abilities: Store TP, Meditate
• Haste equipment/spell: This increases the rate of attacks without decreasing the TP per hit.

Dual Wield and Martial Arts affect the weapon delay, and are not considered haste effects. Therefore, they will reduce the TP per hit up to the 5 TP floor. Hence, a thf/nin with low delay daggers will gain tp faster than a war/nin dual wielding axes.

Finally, don't forget you need to hit to gain TP, so accuracy plays a large role.

[Spider-Dan]

Regarding hastes and TP gain:

I don't believe Haste status changes your TP gain (the logical correlation would be that Slow status increases your TP per hit). Temporary effects don't change your TP per hit, but permanent effects (like Dual Wield and Martial Arts) do. This is easily proven with a 25+ NIN or 31+ MNK: with the NIN, equip one sword and note your TP return per hit, then equip another additional sword and note the change. With MNK, change your job to THF/(non-MNK-sub) and notice your TP return with 480 base delay, then change main to MNK and notice your TP return with 360 base delay.

So then, what we're really looking at with respect to TP gain per hit is the 5% floor. TP return per hit is somewhere around delay / 40-45. We'll assume it's 40 (if it's larger than that, that's actually more advantageous for /NIN).

Chicken Knife (D21/176) + Hornetneedle (D16/150) is a fairly popular combination for /NINs. Net delay per round: (176 + 150) * .85 = 277.1. Divide that by 40, and you wind up with 6.92 TP per round, or about 3.5% TP per *hit* that you should be getting. Because of the 5% TP floor, that's a 30% bonus in TP that you are getting, relative to delay. That's huge.

Additionally, Dual Wield adds an extra hit to all WSes, at normal TP generation rate (which is usually 5% for a THF). That's an additional TP advantage for /NIN.

THF/NIN is unquestionably the best choice for TP generation while exping. Whether or not that extra TP generation makes it better than /WAR is up for debate, but the facts stand as is.

[Ruic]

Originally posted by Spider-Dan
Regarding hastes and TP gain:
Chicken Knife (D21/176) + Hornetneedle (D16/150) is a fairly popular combination for /NINs. Net delay per round: (176 + 150) * .85 = 277.1. Divide that by 40, and you wind up with 6.92 TP per round, or about 3.5% TP per *hit* that you should be getting. Because of the 5% TP floor, that's a 30% bonus in TP that you are getting, relative to delay. That's huge.

Haste calculation isn't a 15% direct reduction as stated before. The actual delay you will get with those weapons is (326)/1.15 = 283. Or "7.075% TP per round" as you put it.

But this drops below the theoretical 300 static cap, if it exists, then a THF using a hornet needle would gain equal TP to a THF/NIN DWing this combo of weapons.

Also you're NOT looking at a 30% bonus in TP. I am sorry to repeatedly pick this argument apart, but it is simply not true. A THF/NotNIN wielding a chicken knife is already swinging a 176 delay weapon, which is BELOW the TP floor. 176/40 = 4.4. Vs our theoretical 3.5% per hand. That seems like a tremendous difference but...

While your numbers work out in theory, they fall apart in practice, simply because we're interested in a TP over time approach. And as examined over and over again, a /NIN is only going to get 2-3 more attacks, not DW hits, but single dagger hits, IE 1 to 1.5 ROUNDS of attack per minute extra. These extra attacks are made even less, if the 300 delay cap exists on dual wield using your weapons. It wouldn't matter if you were only supposed to get 1 TP per hand, but you got 5, if you only get 3 more swings a minute. Does that make sense? You'd predict a 500% boost in TP, but only see 15%.

If both THFs are at or below floor, then these 2-3 attacks can translate into -at most- 15% extra TP per fight. Which is exactly what you would expect, due to the fact that DW is granting a 15% haste bonus, and all parties involved are using TP floored weapons.

I'm beginning to think I must be crazy with how many people think that NIN is going to result in a 30%+ TP/time gain rate. Can anyone explain to me how/why we're thinking this, and give some numerical examples please? Maybe I've missed something else vitally important... If not then DW is going to grant a simple 15% bonus...

[Spider-Dan]

Originally posted by Ruic
Haste calculation isn't a 15% direct reduction as stated before. The actual delay you will get with those weapons is (326)/1.15 = 283. Or "7.075% TP per round" as you put it.

You're working from a different set of numbers. Everything I've seen says that DW is (delay1 + delay2) *:

0.9 @LV10
0.85 @LV25
0.75 @LV45
0.65 @LV65

But this drops below the theoretical 300 static cap, if it exists, then a THF using a hornet needle would gain equal TP to a THF/NIN DWing this combo of weapons.

The problem with this is that no THF would (or should) use a Hornetneedle in their main. Using a D16 dagger as your main weapon (even at LV48) would gimp your SATA damage. Therefore, Hornetneedle is a /NIN only weapon.

And where is the solid evidence for this 150/300 delay cap? If you're going to argue against hard numbers, I think you need more than anecdotal evidence of people failing to notice any difference; you'll need numbers to back it up. You seem to have cited animation as a limitation, but the existence of a 120 delay weapon pretty much refutes that theory; when you fire a Loxley bow, your animation is quite visibly interrupted to fire it in time.

Also you're NOT looking at a 30% bonus in TP. I am sorry to repeatedly pick this argument apart, but it is simply not true. A THF/NotNIN wielding a chicken knife is already swinging a 176 delay weapon, which is BELOW the TP floor. 176/40 = 4.4. Vs our theoretical 3.5% per hand. That seems like a tremendous difference but...

You don't actually address this difference at all, ever. While it's true that /notNIN can try to equip a low delay dagger to keep up, they just can't match two low delay daggers + DW2, especially if one of those daggers is a Hornetneedle.

While your numbers work out in theory, they fall apart in practice, simply because we're interested in a TP over time approach. And as examined over and over again, a /NIN is only going to get 2-3 more attacks, not DW hits, but single dagger hits, IE 1 to 1.5 ROUNDS of attack per minute extra. These extra attacks are made even less, if the 300 delay cap exists on dual wield using your weapons. It wouldn't matter if you were only supposed to get 1 TP per hand, but you got 5, if you only get 3 more swings a minute. Does that make sense? You'd predict a 500% boost in TP, but only see 15%.

You're leaning heavily on this 300 delay cap, which you still need to prove the existence of. Your calculations also assume lots of other things; for example, you say that NIN wouldn't have non-stop Hundred Fists with 100% haste; how do you know this? (Especially considering that you need Soul Voice to pull it off...)

[Kastel]

Originally posted by Mithrankittycat
thf/drk
no, really. stun works even without having magic skill from the main job.

thf/drk is the most fun i have had with a SJ. Not only was being in stun order for gods making them more interesting, but I used souleater and my damage went from 0-10 to 50-70 x.x this was with a dagger ...

[Ruic]

Originally posted by Spider-Dan
You're working from a different set of numbers. Everything I've seen says that DW is (delay1 + delay2) *:

0.9 @LV10
0.85 @LV25
0.75 @LV45
0.65 @LV65

Just about every source I've heard or read from has agreed with the (Delay)/(1+Haste%) as the haste formula, including Japanese players with guidebooks that have the game formulas down to the hundredths of decimals.

And where is the solid evidence for this 150/300 delay cap? If you're going to argue against hard numbers, I think you need more than anecdotal evidence of people failing to notice any difference; you'll need numbers to back it up. You seem to have cited animation as a limitation, but the existence of a 120 delay weapon pretty much refutes that theory; when you fire a Loxley bow, your animation is quite visibly interrupted to fire it in time.

Arrows have delay. The fastest arrow in the game, the Grand Knight's Arrow has 50 delay. 120+50 = 170 combined delay for Loxley Bow + fastest arrow, which is still slower than a beestinger or hornet needle. You're welcome to try to increase the speed of the Beestinger, I've tried it, and noticed no difference qualitatively or quantitatively. I can't test the theoretical 300 delay cap until I get a hornet needle, or until someone who can use, and owns both of those weapons attempts some trials with dual wield, haste and other hastes, while stopwatching.

You don't actually address this difference at all, ever. While it's true that /notNIN can try to equip a low delay dagger to keep up, they just can't match two low delay daggers + DW2, especially if one of those daggers is a Hornetneedle.

And you don't address a single point about any of the benefits of other jobs. But let's continue to spin the TP wheel of fortune, I'll even prove your point for you, then from your proven point, prove my own yet again.

Let's examine a situation in which we pretend there is no 300 delay cap, and let's imagine that the person is using two 150 delay daggers.
300*(1/1.15) = 260.8 frames per round.
To be fair let's say that the other person is using one 150 delay dagger, after all, we've removed any caps. So these numbers may as well be arbitrary, right? And remember that haste has diminishing returns, so using smaller numbers is going to be more drastic.
150 = 150 frames per round.
3600 frames in a minute.
3600/260.8 = 13.8 (times 2 for two weapons) = 27.6
3600/150 = 24
We have an attack difference of 3.6 attacks a minute. Which if we're being strictly mathematical, and looking at a single minute, is 3 attacks, since you can't excute 6/10ths of an attack, or 15% TP. If not we're looking at ~18% TP over our arbitrary time measurement.
Now let's assume 200 delay daggers.
3600 frames a minute
400*(1/1.15) = 347.8
3600/347.8 = 10.3 * 2 = 20.6
3600/200 = 18
A difference of 2.6 attacks a minute. It's 13% per minute more if we allow 6/10ths of an attack. At this point we're not even fully utilizing dual wield!

So between 300 and 400 pre-DW haste, we're looking at a 5% per minute difference in TP gain.

But like you said, no THF is going to use a 150 delay dagger in their main hand. They're more likely to use a heavier dagger, between 176 and 200 delay for their main weapon. Let's just say 180, for the medium range on daggers.

Single dagger: 3600/180 = 20 attacks/minute
Two, assuming one is 150 delay, the other is 180 = 330 delay = 287 delay, = 12.5 times 2 = 25.
5 attacks difference.

If a THF/NIN gives up everything else they do except TP, and SATAWS, they can maximize their TP again roughly around 330 combined delay. If you move to 350, you begin to lose attacks per minute on ratio, and if you drop to 300, you can see the bonus drop to a possible 18% in the above calculation. The best you can do is roughly add 5 more attacks per minute. Of which you must maintain an 80% accuracy per minute in order to convert 100% TP per minute, which is really nearly impossible on XP mobs without owning sniper rings, scorpion harness, lifebelt, and other massive accuracy adjusting equipment. Especially as you pass through Genkeis, and monster evasion rises exponentially. But if you want to do the differentiation and do the maximization function on delay to TP, be my guest. I would say that the best a /NIN can do is have a -potential- for ~25% more TP per battle.

I'll cede this point. I understand the point of using swapped daggers. But if you're NOT using a 150 delay dagger, then it is arguable that NIN sub loses out on nearly -everything- it has to offer. So now that that's understood, let's move on ...

So /NIN wins the TP war... assuming it has the same accuracy as /RNG and that /WAR doesn't double attack any at all. Should we really continue talking in circles here? If WAR double attacks 2-3 times in those 20 attacks, /NIN's TP gain potential is dropped to 0-15% (~0 if using 200 delay daggers, 15% if using a 150 dagger and DA only happens twice) more than WAR, while /WAR is doing 20-40% more damage. This hasn't changed at all. And /RNG is going to yield drastically improved accuracy.

Again.
25 attacks a minute with 70% accuracy = 17.5 hits = 87.5% TP

22 (say WAR sub DAs twice so 20+2) with 70% accuracy = 15.4 hits = 77% TP. (Assuming that no WS hits trigger DAs for more TP and damage.)

20 (RNG sub) with 80% accuracy = 16 hits = 80% TP... (Not even counting chances for more hits of WS landing, leading to more TP/Damage.)

(We're also assuming all three jobs use no ranged weaponry, though it is obvious which job would be more proficient here.)

And yet again, we arrive with a number close to or, in this case less than 15%, as the TP difference between /NIN, /WAR, and /RNG.

After spending pages discussing this topic though, there still hasn't been any thoughtful consideration on the fact that WAR is still going to add attacks, and do more damage, while /RNG is going to convert MANY more attacks into TP than the other two discussed subs. Instead all anyone seems to want to discuss is Ninja haste, and TP, which are only one dimension of a THF's primary purpose, and does not even pertain to secondary or tertiary job roles.

-- Begin opinion
Having seen each sub in action on THFs, I have to honestly say I have been most impressed with /RNG. 85-90% accuracy is disgusting, and comes from having some of the gear I mentioned above, along with a willingness to pay for arrows and other equipment which improve TP gain. When my THF friend hits 60, and has a 30 RNG sub, it'll be like she just swallowed two sniper ring+1s and then she'll gain barrage, which if examined over time, is around 50% every 5 minutes, or 10% more TP/minute... we're closing the gap even further here. I expect that /RNG accuracy will remain steady throughout the higher levels, while /NIN and /WAR will decrease.

I prefer having a good THF/WAR in the party second, as the extra damage can lead to some sickeningly powerful renkeis, and gives better hate control due to the high numbers.

I've nothing against /NIN personally. I simply wished to discuss things here. If you hadn't noticed, I'm sub Ninja on my RDM. Many people consider me to be extremely gimp. However understanding debuff formulas, and decreasing my resist rate has given me a cast per cast effectiveness far higher your average RDM/BLM. I nuke nearly 6-8% harder on average, and I can contribute hundreds of damage a fight with my melee for free.

But this is just my opinion, based on conjecturing and my knowledge of the game. Take it for what you will. I'll stick to unorthodox party make-ups when it pulls in high XP/hour, and is easy.
-- End opinion

You're leaning heavily on this 300 delay cap, which you still need to prove the existence of. Your calculations also assume lots of other things; for example, you say that NIN wouldn't have non-stop Hundred Fists with 100% haste; how do you know this? (Especially considering that you need Soul Voice to pull it off...)

I discarded the entire cap formula just for you, and gave you another full comprehensive analysis. Make of it what you will. You're more than welcome to go and perform your own calculations on haste and dual wield. If I ever get a hornet needle, I'll run my own tests of course. But to be honest, I don't care very much about this topic anymore. As I said in the first paragraphs of the post, I've no interest in THF. And many people seem to have taken my posts as an attack on their job/sub choice. That was not my intent. Researching and gathering information was. I feel at this point that I've gathered all useful data and considerations though, beyond the point necessary, and that this thread is going to endlessly talk itself in circles.

... So as far as I am concerned, I'm through with this thread. If you have any particular questions you'd like to direct to me, please use the PM function.

I apologize if this seems rude, or if it seems I am backing out of an argument. I simply feel that the discussion has left all scope of a comprehensive analysis of the THF job and various subjobs.

Take care.
- Ruic

[Spider-Dan]

Originally posted by Ruic
Just about every source I've heard or read from has agreed with the (Delay)/(1+Haste%) as the haste formula, including Japanese players with guidebooks that have the game formulas down to the hundredths of decimals.

I don't think the formula you are using is referencing the right thing. Those formulas deal with haste (e.g. Haste spells, +haste equipment, etc.) whereas I'm talking about Dual Wield. Although we commonly refer to DW's effect as "DW haste," it obviously is not the same as haste, if only for the fact that (as you already noted) haste does not decrease your TP return, whereas DW does.

Arrows have delay. The fastest arrow in the game, the Grand Knight's Arrow has 50 delay. 120+50 = 170 combined delay for Loxley Bow + fastest arrow, which is still slower than a beestinger or hornet needle.

Ammo delay is the delay before you can shoot again. The Loxley Bow goes from input to execution in 120 delay (1.2 seconds). Your argument was that the 150 delay cap was there to prevent animation from being cut off; the fact that Loxley Bow clearly interrupts the bowstring draw animation to fire invalidates this argument. If Squenix cared about animation being interrupted, they would not allow Loxley to fire as fast as it does.

Let's examine a situation in which we pretend there is no 300 delay cap, and let's imagine that the person is using two 150 delay daggers.
300*(1/1.15) = 260.8 frames per round.

Again, it should be 255 (300 *.85), but there are much more glaring mathematical issues in your post, which I'll get to.

150 = 150 frames per round.
3600 frames in a minute.

Absolutely, positively incorrect. There is 6000 delay in one minute. 100 delay = 1 second, real time. This is easily proven.

My NIN/RNG uses a Pirate's Gun (600 delay) with a LV22 Bullet (240 delay). I use the following macro:

/equip ring1 [RACC ring]
/equip ring2 [RACC ring]
/ra <t>
/wait 8
/equip ring1 [melee ring]
/equip ring2 [melee ring]

Now, if 60 delay = 1 second, that means that it takes 10 seconds (600 delay) for my NIN to fire the weapon, then another 4 seconds (240 delay) before I can fire again. The above macro would be totally pointless, because I would switch back to my melee rings 2 seconds before I fire. In actuality, the second ring swap comes well after I fire, right before I can fire again. This clearly shows that 100 delay = one second.

(And, BTW, FFXI runs at 30FPS, not 60.)

Now, let's make the appropriate changes to your formula (I'll even use your 260.8 figure):

6000/260.8 = 23.006 (times 2 for two weapons) = 46.01
6000/150 = 40
We have an attack difference of 6 attacks a minute.

That is 30 extra TP for the /NIN, per minute.

The rest of your calculations on this topic all refer to the flawed 3600 delay/minute formula, which is incorrect. You can update it all, if you like.

[Rones]

Absolutely, positively incorrect. There is 6000 delay in one minute. 100 delay = 1 second, real time.

Actually Ruic has it right on the mark and you going off the common misconception. Here is a lengthy discussion of it for you. I'll include an excerpt from it to clear thigns up below. The 60 delay = 1second is also the commonly accepted equation by the majority of JP guide books that include most of the formulas for this game.

I was in the same group as those who believe 100 delay = 1 second. Me being the curious person I was, I decided to test it myself.

I grabbed my L15 THF, equipped an onion dagger (193 delay), and my stopwatch.

If 100 delay = 1 second, then the swing time should be 1.93 seconds.

If 60 delay = 1 second, then the swing time should be 3.216 seconds.

Obviously, a human hand and a stopwatch dealing with milliseconds = mucho human error. So, testing between two swings is kind of moronic, even with such a wide margin of error.

So, I'd simply find a mob that would survive several swings, and time from the first hit to the 10th hit, then divide the total time by 10 to get an average. Note: I'm timing from when the first and 10th hits show in the chat box. Going by animations is very unreliable.

10 swings = 32.25 seconds...

32.25/10 - 3.225 seconds per swing.

So, with human error figured in, which is this closer to, 1.93 (100 = 1sec), or 3.216 (60 = 1sec).

At this point, you cna take me at my word, or you can test it yourself. But anyone who argues that 100 delay = 1 second from hereon out, IMO, has done no testing other than perception.

And a quote from game guru Applepie:

The number is actually frames between swings.

Let's take a look at my favorite sword, Royal Guard Fleuret. It has 215 Delay. This means 215 frames are needed to be rendered between swings.

Since all FFXI motions are based on 60fps internally regardless of what fps we get in game, the actual delay (seconds) is calculated in this way.

215 [frames] / 60 [fps] = 3.58333333..... [seconds]

And just for the heck of it, another thread proving it.

[Spider-Dan]

Originally posted by Rones
Actually Ruic has it right on the mark and you going off the common misconception.

Did you miss where I posted exactly how I proved that it's NOT 60 delay? One can do all the stopwatch testing they like, but if 60 delay is 1 second, then my macro would not work, period.

If you take a gun + bullets out for testing, you can clearly see that it is NOT 60 delay per second. I just used the following macro with the aforementioned Pirate's Gun + Bullets (600 + 240 = 840 delay):

/ra <t>
/wait 9
/ra <t>

The result: two shots in a row. If 60 delay is one second, then that's a 10 second wait to shoot, followed by a 4 second wait before you can fire again (ammo delay). My /wait 9 would attempt to /ra again before I even fired the first shot!

I'm sorry, but no amount of stopwatch testing (complete with human error and internet lag) can convince me of the 5+ second difference between what you are claiming and what actually happens in game. Do any sort of testing with guns and it will be patently obvious that 100 delay = 1 second.

[Matkun]

Originally posted by Spider-Dan
Did you miss where I posted exactly how I proved that it's NOT 60 delay? One can do all the stopwatch testing they like, but if 60 delay is 1 second, then my macro would not work, period.

If you take a gun + bullets out for testing, you can clearly see that it is NOT 60 delay per second. I just used the following macro with the aforementioned Pirate's Gun + Bullets (600 + 240 = 840 delay):

/ra <t>
/wait 9
/ra <t>

The result: two shots in a row. If 60 delay is one second, then that's a 10 second wait to shoot, followed by a 4 second wait before you can fire again (ammo delay). My /wait 9 would attempt to /ra again before I even fired the first shot!

I'm sorry, but no amount of stopwatch testing (complete with human error and internet lag) can convince me of the 5+ second difference between what you are claiming and what actually happens in game. Do any sort of testing with guns and it will be patently obvious that 100 delay = 1 second.

I went out and timed again, with

this macro:

/echo start
/wait 5
/echo end

I started the macro when I saw the text message of my hit. I had no lag, and with my Greatsword with 453 delay, I needed a /wait 8 inbetween hits to match up with the damage text.

By 100 delay = 1 second, /wait 5 if started after the hit, should show up after the next hit. It never did, there was always a wait of 2-3 seconds after 'end' showed up before the text message of my next hit.

With /wait 8, it matched up properly, which matches up with the 60 delay = 1 second theory.

Did the same with a stopwatch, My greatsword and Scythe both take more then 7 seconds between swings.


I also tried out your method using a /ra <t> /wait macro with my crossbow and bolts,

Delay = 288 +192 = 480. /ra <t> /wait 5 /ra <t> Did not work, but /wait 6 did.. Which does not match up with what I found out from melee weapons.

I also noticed and realized that ranged weapon attacks function very differently from melee weapon attacks. Ranged have a 'startup' delay, where you are aiming and have a delay before you deal damage or miss. Melee Weapons have a delay AFTER you deal damage or miss.



Which leads to a few possible conclusions..

1. Melee weapons have a 60 delay = 1 second ratio, Ranged have 100 delay = 1 second ratio.

2. Ranged Weapons delay is somehow spread over the aiming period/after shot waiting period..

Basically something is different about ranged weapons, because from all tests, with /waits and timers and stopwatchers, ALL Melee weapons have complied with 60 delay = 1 second.

[Spider-Dan]

Put simply, stopwatching melee autoattack is unreliable at best. Autoattack is affected by lag. (And BTW Matkun, your test of hitting the macro when you see the text in the log is even more problematic... how many times have you seen a mob die before the log reports that you hit it?)

I regularly have the occasion to play FFXI on an extremely lagged connection (laptop + cellphone modem, 4 digit ping to my ISP). When I do this, even with my Hornetneedle my autoattack rate is all over the place. Now obviously, I'm sure we all are doing these tests on high-speed connections. But the point that I'm making is that autoattack is lag dependant at the very least.

It also might be affected by any number of other factors. I know for a fact that autoattack timers are put on hold when you activate a JA, and we've all seen how you "freeze" if you are hit by a spell and you aren't moving, etc.

We can all agree from the start that stopwatching (when dealing with such miniscule portions of time) is inherently flawed. You can try to counteract it by doing a lot of autoattacking, but you have to assume that there's nothing going on that is delaying your autoattack cycle, which is unproven. Ranged attacks are a simple and easy way to clearly see that 100 delay = 1 second. They have a defined and controlled beginning and end.

[Rones]

Here's a better thread you can study.

Now obviously, I'm sure we all are doing these tests on high-speed connections. But the point that I'm making is that autoattack is lag dependant at the very least.

I've found my attack rate to be unaffected by lag. I've lagged out in the middle of a solo fight and came back to find I had killed the monster and lvled up from it, meaning my character was attacking seperate from what my screen showed. It is the central server that does all the attack calculations and timing and sends back the info of what should be displayed on your screen. Sometimes the message of what to display can get lagged, but its still running on time for the server. This means one attack might get delay slightly and the next attack will come sooner (visually speaking). The best example I've seen of this was a fight where my R went to 0 for 20 seconds, suddenly it sprang back to life and saw the previous 20 seconds play out in about 3 seconds on my screen and was quite impresive (made mnk 2hour look slow). :sweat:

The conclusion of what I'm getting at is timing 50 swings against a too weak mob with a low dmg weapon and no one casting on you without noticiable lag will bring much more accurate results than you are giving it credit for. Some of the swings visually might get lagged on occasion, but over the course of the fight it will come out very close to the calculatable time.

Here is an excerpt from Applepie's Thread which calculates out his number of swings and it matches his video (which is a blast to watch) along with matching his parser.

Let's think of how many times we can hit them for 10 minutes.

Suppose we have Temple Knights Army Sword (Delay: 240).

No Haste
[60(fps) * 60(secs) * 10(mins)] / 240 = 150
- 150 times

With Haste
[60(fps) * 60(secs) * 10(mins) * 12.5%] / 240 = 168
- 168 times

With Haste and Wyvern Earring
[60(fps) * 60(secs) * 10(mins) * 17.5%] / 240 = 176.25
- 176 times

[Rones]

I got bored yesterday and found a nice too weak elemental to experiment on. I pulled out my 400 delay staff and began auto attacking. I cast no spells during the fight. After the 2nd attack I started my timer for 2 min and counted the number of attacks.
If 100delay = 1 sec, I should be able to get an attack every 4 seconds. So 120 sec/4 = 30 swings with my staff.
If 60 delay = 1 sec, it would be 400/60 = 6.67, meaning one attack every 6.67 seconds. So 120/6.67= 18 attacks. That's a difference of 12 swings between the 2 formulas.

Well, when my timer beeped after 2 min I had performed 17 attacks (18th was about to go). There was no detectable lag during that and no cures cast on me to slow down my attacks. There is no way I could be short 13 attacks unless the 100delay = 1 sec formula is wrong. Don't believe me? Go try it with any weapon besides your gun. This whole discussion was about melee weapons rather than ranged, so you should experiement with melee weapons.

Edit: Spelling

[Huck]

Originally posted by Omena
Wrong. Triple and double attack can not trigger on each other. If both trigger on the same swing, triple takes priority and no doulbe attack occurs. It is not possible to get triple or double trigger on hits given by triple or double either.

Acutually-No. Triple and Double attack CAN trigger each other. I've had it happen, I've seen it happen with other thf/wars. Double seems to go off about 1/6, and Triple is maybe 1/8. It's not one solid 6 hits, It's like this.
Let's say this is an attack X, You have Double attack X X, and Triple Attack X X X.
Ocassionaly you will see this happen X XX X, or XXX X, it's not a XXXXXX, so I guess you're part right. When triple attacks happen, It's less likely than 1/6 of the hits will double, It's like 1/40+. But it CAN happen. And it CAN happen more than once. Like this XX X XX, or XXX XXX. And each of those hits that happen off of that can trigger a double or triple respecftully. XXxx X X. It happens.

Originaly posted by Omena
With 200 delay daggers the TP gain bonus is around 30% and obviously it is even better with even faster daggers. I don't know why you are calling that "marginal". Getting 100% TP in 60 seconds is pretty good in my opinion. With some luck I can use WS in several fuidamas in a row. [/B]

With ~200 daggers it would be 25% TP bonus, You get 10% at 20, +15% at 50. This is true, this looks like 50+ thf/nin gets a lot more tp than 50+ thf/war.. except, for Double Attack. Double attack is quite possibly better than berserk. Double attack -> TP Machine, I gain TP faster as thf/war than thf/sam or thf/nin. (I've got them all to experiment, how about you?)
Most of the time as thf/war I can get 100%+ TP in about 45~50 seconds. As a thf/nin, you say you are "lucky to use serveral WS fuidamas in a row"? How is that 30% faster TP? If thf/nin gets such great TP compared to thf/war, How can 100% TP in 60 seconds be "pretty good".

[Huck]

Can someone Sticky this, and Un-sticky that worthless "thf sub guide"?