How to Dragoon Per Second: A DPS Paradigm

[Outfoxed]

It is actually possible to get 3 GKs in and not miss any BotD application and to sustain HT + PH uptime 100% with 2.5GCD, if my model is correct... but it's A LOT harder to play as you get 0.1s leeway

my numbers don't agree with you.

0	h	bl	b4b            botd timer		
2.5	i	pot					
5	d	botd	ir		---
6.5                                     15 ---BOTD CAST---
7.5	c	ps	leg		14		
10	4th	gk			16.5		
12.5	p	jump			14		
15	t	dfd			11.5		
17.5	t	ls			9		
20	t	ssd			6.5		
22.5	4th				19		
25	h				16.5		
27.5	i				14		
30	d				11.5		
32.5	c				9		
35	4th				21.5		
37.5	p				19		
40	t				16.5		
42.5	t				14		
45	t				11.5		
47.5	4th	gk			14		
50	h				11.5		
52.5	i				9		
55	d				6.5		
57.5	c				4		
60	4th	gk	botd falls off after GK				
62.5	p						
65	t					botd up @ 66.5

according to my numbers you have closer to a 4 second leeway for GK reapplication.

http://pastebin.com/2gmaa66B <-huge writeup

[Dervy]

snip

0.1s leeway for HT reapplication, not GK. Should of made it clearer. http://puu.sh/iE2XM/fafa65901c.png Column "C" is the duration remaining on HT when you successfully deal damage to your opponent. As you can see, C2 has 0.1s left on HT when you go and actually damage your opponent. Animation buff delay is awesome sometimes.

Your 2.5GCD GK usage is exactly the same as mine, there's absolutely 0 difference. I never said it was impossible to do so, it just requires a lot more management and "know-how", rather than a simple minded approach which works with 2.4GCD (which is listed on the OP). Well, I did kinda say it was impossible a few pages back... But I made a faulty assumption about 2.5GCD + GK, and I retract that statement. Should of double checked my model before hand.

Now. A fully looped cycle, including GK management lasts 122.5s. This is the point where your entire rotation starts from the beginning again and you're using BotD just before Chaos Thrust.

You can also use the same BotD layout for your 2.5GCD example, with 2.4GCD. So, to get a full loop with both of these, here's the duration and potencies.

2.5 GCD
Potency: 21175
Duration: 122.5s
PPS: 21175/122.5 = 172.8571

2.405 GCD (607ss/i180 geared)
Potency: 21102 (includes %chance to clip a DoT tick)
Duration:117.82
PPS: 21102/117.82 = 179.10371

Even with the DoT clips, the PPS is much higher at max i180 SS than it would be at 2.5GCD. If you get one unbuffed HT, it amounts to 28.05 Potency loss (170*0.15*1.1). That will definitely hurt your overall PPS if you continuously miss out on those and in a full 122.5s cycle, there's 5 HTs being used.

The next argument is something which Aiurily has tested at 2.4GCD. You can actually Pop BotD before Disembowel and use your Drac-Pot before Chaos Thrust. This results in you having a 104+ STR buffed 4th Tier after your Full Thrust, versus a Disembowel. This is significantly more damage. Unfortunately, my model doesn't allow me to calculate the additional Damage I'd gain from using a Drac-Pot during certain parts of the rotation, as my model is based on PPS...

My hard wired version comes out to it being a difference of around 30 DPS, give or take rounded numbers. It's hard to say what the exact number is as I believe SE have also changed some of the coefficients for Strength and Weapon Damage behind the scenes... This is the difference at 2.405 GCD, not a comparison between 2.5GCD, by the way.

As for how I calculated DoT clips before people get confused, I calculated a probability of you clipping a tick, whilst you have your DoT on your enemy. The closer you reapply your DoT to a value of 3, the greater the chance you have of clipping your DoT tick. The actual math I use is (DoT_Duration - Animation)/3. This will give me a percentage, which I multiply by one DoT tick, then subtract it from the overall DoT potency. It's an average, not an exact value, but it's good to calculate the average PPS.

So yes, the math does check out that at 2.405 GCD, you're doing significantly more PPS. As for the numerical DPS value (which we're all focused on...) we still need to figure out how Skillspeed affects DoT ticks, as well as Critical Hit Rating and how Determination scales Auto-Attacks. And you are correct, we don't know what Alexander Hardmode gear will look like.

But this is a guide for 3.0, not 3.01. Once Alexander Hits and we evaluate the gear differences, plus when we have more knowledge on Skillspeed, Crit and Determination, then we can make a lot more accurate and calculated assumptions for what the rotation will be like and of course, the guide will be updated for then.

But as I said previously, even if it is possible to get the rotation going with 2.5GCD, the numbers add up to the developers designing the rotation around 2.4 GCD. If Alexander gear is i180 and has more Determination, I could probably see us going down to 2.45 GCD, as gear is just absolutely loaded with Skillspeed at the moment.

[Zathir]

So am I to understand the new ideal rotation for Dragoon is one that's pretty much impossible with the tiniest bit of latency?

I tested out Train Your Dragoon pretty extensively in 2.55 to test the benefits of the double off-GCD and found that without a doubt it's a DPS loss on average if you get one latency tick anywhere during the opener or the time it repeats, with fights expected to last 10-14 minutes.

Double off-GCDs simply are not feasible unless you have literally a perfect connection to the server. To use double off-GCDs so extensively in the opening rotation is creating multiple opportunities for massive DPS loss, especially with this game's latency problems and DRG's issue specifically with abilities like the jumps sometimes lagging and animation locking you for 1s in to something that never casts.

This just seems like theory crafting for the most part from people with perfect connections, on a rotation that is very likely to not be possible with latency. I'd like to see an actual workable rotation come out that priorities consistency and compatibility with latency over theory crafting that doesn't hold up at all in a latency based environment.

[Sleigh]

Question for you DRGs out there, would you say level 60 DRG has fallen in the LB priority department? In ARR it was the defacto melee LBer IMO, but with how sensitive this BotD system looks (currently leveling DRG) I'm gonna guess NIN has gained that honor? Or would there be pieces you could skip in favor of LBing in an optimal way (like skipping a FT combo in ARR)?

[Tilgung]

If it's a LB3 fight I do it anyway because seeing Dragonsong Dive is worth the DPS loss on the parse. :>

As for optimum, dunno.

[SunnyHirose]

Think it'll be easiest for NIN, regardless of who loses the most damage in the end.

Still, if you blow your BotD for any reason, you're almost an ARR goon anyway. And it's gonna happen.

[ChaozK]

Question for you DRGs out there, would you say level 60 DRG has fallen in the LB priority department? In ARR it was the defacto melee LBer IMO, but with how sensitive this BotD system looks (currently leveling DRG) I'm gonna guess NIN has gained that honor? Or would there be pieces you could skip in favor of LBing in an optimal way (like skipping a FT combo in ARR)?

Imo yes. It should be NINs job now.
I would still do LB3 for a MNK though, since we can get back BotD faster, if we have its CD ready.
But yes, NINs should do it and my FCs NINs are alreay doing it in place of DRGs.

[Aiurily]

dragonsong dive is the coolest melee LB, so yeah

but yeah it's definitely really rough for DRGs to LB now compared to NIN. MNK as well because of the demo change

[ColdestHeaven]

Is there currently any point in holding any of our oGCD/buff to line up with other skills/cooldowns in 3.0? Aside from the obvious Power Surge+Jump. I don't think anything has changed, but the CD have stopped lining up like they did back in 2.55 at base skillspeed so I'm having some troubles. Blood for blood best used after HT, that's all I guessed.

[Dervy]

Yeah. Geirskogul is another skill that comes into mind. It doesn't really matter when you use Geirskogul, as long as you're fitting in 3 per minute.