Some surprising data in regards to SMN dps vs BLM dps

[hallena]

who cares what your average dot ticks were - your total bio II dmg was 3500 and bio I 3200

I don't know what you're trying to prove right now. Look back at the log, then back at your comment. Repeat the process until you realize how pointless your last statement was

tip : He has not been dpsing 1:30+
tip #2 : 700x5 =
800x4 =

[Crescent_Dusk]

Not to mention total damage can be misleading as you could just be padding your numbers multidotting stuff that didn't need multidotting as it was assigned to somebody else that could reliably take care of it regardless in some encounters.

[Viridiana]

who cares what your average dot ticks were - your total bio II dmg was 3500 and bio I 3200

Look at his Bio II damage.

Now look at his Miasma damage.

Now assume that he's not a total idiot.

[KroLeXz]

yeah I get it - just assumed because he said he tested odd rotations and that the dps was lower because he allowed it to just sit there.

[pandabearcat]

What happened to this thread while I was gone over the weekend D=

Anyway I read about 2 of the 5 pages or whatever.

Here's the points I want to reiterate.

1. Yes, the calculated DPS is from a standstill fight with perfect execution, as you may have noticed from like the second line in my original post. This is an inherent flaw of a MODEL as opposed to a SIMULATION. These caveats are what are necessary to actually model the dps.
   
   You are WELCOME, and in fact I INSIST that you (or someone) run a monte carlo simulation of actual fight mechanics with BLM and SMN. However, I don't think anyone has written (or even is writing) a version of SimCraft for FF14, and we don't even know all the mechanics, which is why we are modeling instead of simulating.
   
   If you want to be precise, the exact assumptions are as follows.
   
   1. The GCD is 2.5 seconds. As GCD gets faster, you will do -more- dps than calculated.
   2. The player has 0 lag and 0 hesitation in casting.
   3. The player requires no movement
   4. The player has infinite mana. This is a null point if you have a bard, but it is an actual worry if you do not have one...you will run out of mana at some point. I have made edits to the model at the bottom of the post stating the amount of DPS lost if you do not ruin, etc.
   5. The player's pet doesn't suck and isn't stupid (this might be the worst of the assumptions lol)
   6. Pet damage is guesstimated at 75 potency a cast. If someone has a better guess/tested answer it would be more accurate.
   7. Shadowflare damage is calculated at 25 potency every 3 seconds. I haven't tested, but people are saying it either auto crits or is doing much more than that per tick.
   8. Optimistic view of cooldown usage, as cooldowns are assumed to be used "in a general sense", in that, they take no GCD (assuming you use them off gcd), and that they last their full duration of stacking (namely Spur, Rouse). In reality, this would not happen. However also in reality, this does not significantly affect dps, as even using no personal CDs is only a 2.5% dps loss, and using no pet CDs is only <5% dps loss.
   9. I assume no swiftcasting. In reality you will be swiftcasting at least the first Shadowflare, possibly even up to half of them, but I'll leave this one as a positive assumption.
2. As you can see from the above list, its has assumptions that lower dps and assumptions that increase dps. My personal belief is that it averages out to be a pretty good profile of SMN dps. Obviously you are not required to believe this, but since I don't see anyone else doing their idea of the DPS other than parsing (which we already know is flawed), or just anecdotal evidence (which you'll forgive me if I say is COMPLETELY USELESS).
   
3. The point of the model was to get a general ballpark estimate of comparison between SMN and BLM dps. This means that raw potency is used, rather than weapon stats and armor stats, crit, etc. This works because we know potency is a linear multiplier to the factors listed above, thus, 99 potency is lower than 100 potency in an absolute sense, and 99 potency is exactly 1% less than 100 potency in the absolute sense.
4. Again like Kevee and others have stated, the tick is completely independent of your DoT application times, and ticks purely when the server decides it is ticking time, based on what buffs/debuffs are active.
5. If you are dead set on saying DoTs cannot be as I stated, you are welcome to subtract from my analysis. Total Potency calculated over 300 seconds is 16709+5544 = 22253 pps. Every missed Bio II, Miasma, Thunder tick is 35 potency, and every Bio is 40 potency. You are welcome to subtract the potencies from the total and calculate your own pps numbers. Just a general guideline, every "full" rack of DoT ticks that you miss costs you 145 potency. This equates to a 0.65% dps loss per lost "full" tick. This is actually significant if you want, but it requires you to lose almost 20 ticks, or 60 seconds of DoTs in order to lose the 13% of pps that I calculated would make you equivalent with lowballed BLM dps. If you spend a 3 minute fight with 60 seconds without DoTs, there is something seriously wrong with you, and you probably should pick a different class.
6. I do not play BLM so my BLM analysis is probably really flawed. I welcome people to help me with that one, cause I really don't know anything about BLM endgame, just a general understanding of their mechanics.

[ChickensEvil]

As that chart listed an incorrect, flat 2.0 modifier for Astral Fire, the calculations for BLM potency should actually be lower than previously stated.

Sorry, was the best I could find posted on the internet, and the chart was made toward the end of Beta, so either they gathered incorrect data or the data has changed from Beta to Live, both of which is entirely possible. I didn't have access to the game at the time I wrote up the calculations, so I did my best to find what seemed like accurate information.

hmm people think the rotation is hard?

"hard" no, I don't think anyone claimed "hard", but as you can see even you are on the basis that it is better to do T3 instead of T2 and there are even some claiming you should do T1, and I never even got into the whole adding in Flare calculation... which is basically, whether you should Flare at bottom of your mana pool and then transpose, and continue, or if you should just ignore flare for single target... and even then, what about taking swiftcast into account (so ditching the 4 second cast time of flare), or taking a double flare into account (cast flare, swiftcast, convert, flare)? Hard? hardly... but which is the most optimal? That is a large point of debate, and hardly the focus of this thread, which is why I just tried to take the most basic of rotations.

but sometimes that ends poorly as you might lose count of your fires and end up not having enough MP for Blizz 3

I don't know about the affects of MP song, but the easy way to remember it is <1036 mana cast Bliz 3. If >= 1036 cast fire. 1036 is the breaking point that would leave you with EXACTLY 0 mana left at the end of a Thunder 2 cast (another reason to stick with thunder 2 over thunder 3... uses less mana) and you getting totally unlucky with no mana regen tick. But this is the simple way to remember it without actually having to count your casts.

For someone who uses either Thunder or Thunder 3, just add up the mana cost of Bliz 3 while in AF3, and the mana cost of whichever thunder spell you are using to determine your breaking point (this also works for leveling purposes too).

So the "burn rotation" is then 5 fire 1's, 2 fire 3's, lasting 7 GCDs, or 17.5 seconds.

Then we Blizz 3 (3.5 seconds) and thunder 2 (3.0 seconds), then fire 3 (3.5 seconds) for a full rotation of 27.5 seconds.

Consequently, if we remove raging strikes altogether from the summoner rotation, it does 120.5 pps.

Incoming TONS more information than you ever cared for about BLM, TL;DR fixing your calculation and my old one, I come up with ~121PPS which compares to your ~120PPS on SMN

You count on timing is off, So when in AF3 or UI3 the opposite spell with have it's cast time cut in half. So assuming no spell speed if you are in UI3 Fire 3 goes from 3.5 to 1.75, Fire goes from 2.5 to 1.25 (however you still respect the built in GCD which I believe is 1.5). When in AF3 casting Blizzard 3 it goes from 3.5 also down to 1.75. Finally, (which was pointed out to me by someone else, and I have tested it, and it seems to be the case), AF and UI doesn't actually change until the spell HITS the target. This means, if you hit your next spell fast enough, it will still "think" you have for example UI3 instead of AF3 for the purposes of your cast bar, however by the time it strikes the target it has identified the appropriate potency calculation. So basically, you end up with one quick cast of Fire at the full AF3 potency. Whether this is intentional design mechanic, or an "exploit of sorts, or even just a "flaw" in the game design, remains to be seen... but that is the way things work. Anyway, your timing should look like this:

(assuming we are starting in AF3 with our first Bliz 3 cast)
Bliz 3 (1.75), Thunder 2 (3.0), Fire 3 (1.75), 1st Fire (1.5), 2-5 Fire (2.5), Assumed 2 Firestarter procs (1.5 per) = 21 Making your thunder refresh EXACTLY perfect.

Calculating in the chance of thundercloud can get messy, because it is such a low chance it most definately does not happen every cycle, but sometimes RNG makes it happen 2x. This variance makes it very hard to calculate properly. The reason for this is you cannot just assume an average proc rate as simply as you are trying for one reason, if you get a double thundercloud proc on a single cycle, you would be getting the thunder 3 proc damage on that second hit. This is because any time you get thundercloud you ALWAYS hit thunder 3. However the amount of damage thundercloud does is based on whichever thunder (1, 2 or 3) actually triggered the proc. So a thunder 2 proc would be 295 potency, but a thunder 3 proc would be 340. This makes it far more complicated, to make a best guess average, since you would have to account for your chance of getting a thunder 3 proc in a rotation (or rather a double thundercloud proc), this also adds in overhead on your rotation time for each proc because you have to eat the 1.5 global to hit the instant cast.

Sorry, if I just confused you to death, I hope you were able to track all that. Anyway, if we ignore thundercloud altogether for the moment, using your numbers it should be:
Bliz 3 (1.75), Thunder 2 (3.0), Fire 3 (1.75), 1st Fire (1.5), 2-5 Fire (2.5), Assumed 2 Firestarter procs (1.5 per) = 21
55 (noone has tested this to contest the AF3 damage on Bliz 3... so I am going to stick with it reducing your damage down to .25, base is 220) + 50 (thunder 2) + 7*35 (245) (thunder ticks) + 55 (again .25 reduced fire 3 damage based on previous numbers noone has verified) + 150*5*1.8 (1350) (5 fire) + 220*2*1.8 (792) (firestarter) = 2547 total, for 21 seconds = ~121PPS

Which again, still puts us just about on par with summoner, completely ignoring thundercloud and raging strikes AND any addition of using Flare (or double Flare).

[Grids]

why would you use bane other than to spread and have that chance of resetting the cd? what is the percentage anyway on that? not high from what i've seen.

[KroLeXz]

pandabear smn sims were recently update - can you check for any errors in the sim on behalf of us smners.

http://chocobro.com/sim.php

[point09micron]

However the amount of damage thundercloud does is based on whichever thunder (1, 2 or 3) actually triggered the proc. So a thunder 2 proc would be 295 potency, but a thunder 3 proc would be 340.

People need to stop perpetuating this blatant lie. The tooltips for the spells and the trait say two different things, and five minutes of testing will tell you the spells are the wrong ones. Regardless of which Thunder DOT is on the target, the proc does the damage of the one you cast on the proc. It's easily observable as soon as you get Thundercloud at 28 and just have Thunder 2. If you cast Thunder 1 and use T2 on a proc, it does 295 potency. If you cast T2 and use T1 on a proc it does 240 potency.

[pandabearcat]

pandabear smn sims were recently update - can you check for any errors in the sim on behalf of us smners.

http://chocobro.com/sim.php

Oo I didn't know that existed, imma go play around with it.