Thread: Some interesting numbers on 4.0 BLM

... or, if you prefer, "how SE broke my heart but I'm having trouble getting out of this abusive relationship" (what, I don't like RDM qq).
I'm not bringing this here to offer solutions, but just to show how BLM is a disaster design-wise and, I'd dare to say it, fundamentally broken (hi, my SMN/DRG/MCH brothers and sisters ).

I want to present two sets of numbers which entwine to bring us the core issues of 4.0 BLM.
Yes, there has been much talk about this, but I kinda wanted to compile all that I've posted here and on Reddit in a single thread. First, AF/UI timers for 3.0 VS 4.0, or "why I can't keep god damned AF III up while I vomit Fire IVs whenever the boss looks in my general direction":


- AF leniency and windows of mobility (aka: "windows of mechanical mercy"):
In 3.2 onward, we had 12 seconds of AF III. We had 3 primary AF rotations (disregarding procs for now, but we'll get into it). F4 has a base cast time of 3 seconds, F has 2.5. Thus, we had:

- F4x3>F>F4 (9+2.5, 0.5s window for the AF III refresh)
- F4x2>F>F4x2 (6+2.5, 3.5s window for the AF refresh- this one is important)
- F4>F>F4x3>B3 (9+1.5, 1.5s window to get the B3 off and go into UI III)

We could do whatever suited the situation best. The middle one was the most common rotation. As can be seen, we had around 3 seconds of leeway during the initial part of thw AF rotation and were left with AF III on a 12s timer, in which we had to cast 2xF4 and B3 for a total of 7.5s of casting things. This gave us a 3.5s window in the first part, and a 4.5s in the other, and thus we could fit 2 procs there (or just run around for mechanics without losing F4s).
With higher speed and some slidecasting, this would become even more comfortable.
The merit of 3xF4>F, as tight as it is, is that if one could do that, we were left with 12s of AF in which we only needed one F4 and one B3- you had 12-4.5=7.5s of AF leeway in which we could shove Thundercloud procs while running around.

The current 4.0 scenario has only one rotation:

- 3xF4>F>3xF4>B3 (9+2.5s, meaning you have 1.5s for the AF refresh and 2.5s for the UI III refresh)

Even if one predicts heavy movement, there's way to alter our rotation for it. Since the cast time of 3xF4+F is 11.5s, and AF now has 13s of uptime, one has a 1.5s window to work with. Compare with the previous 3.5s for F4x2>F>F4x2 or the massive 7s for the 3xF4>F>F4.
No matter what we do, that 1.5s will haunt us. So if anything happens we're forced to cut the rotation a F4 short. Which is a disaster, siiince...


- PPS gains in 4.0:
Let's now check the expected pps of the 2.X, 3.X and 4.X rotations- the reason why I introduced the previous section was so we keep in mind the progressive difficulty of each of these rotations.
Let us consider a simplified calculation where the Umbral cycle always consists of B3>Foul>T3>B4 is a magical world where mana ticks always align so we have full mana (bad ticks only penalize the 4.0 rotation, so we're throwing it a big bonus doing this) except for the 3.X rotation.


2.X:
With 3 Umbral Hearts, we use 6 Fire. We do not use 7 as that would force us to transpose (which might be a dps gain... I'll check it later).
Since I want concrete results, let us consider the cases where we get 2 and 3 Firestarter procs (expected value between these two).

Pps for 2 procs (AF and UI cycles): 134,67
And for 3: 137,35

Seems pretty good... for what it's worth, the no B4 rotation with just five Fires has about 140pps with two Firestarter procs, but I'd probably still gamble on six Fires to get a better chance at nabbing extra Firestarter procs.
How does our old HW rotation fare?


3.X:
We don't need Umbral Hearts; we do 4 Fire IVs, one Fire and back to UI to Foul into T3;

No firestarter: 145,80
With firestarter: 148,30

This is a gain for something that's harder to do.
It starts to luck ugly though- if we get lucky rng and get a free F3 per fire, pps comes slightly ahead. Overall, we see a 7% pps gain, which isn't that amazing, but I'll take it.
... what about the 4.0 rotation?


Stup- I mean, 4.0 rotation:
You know how it goes; F4x3>F>F4x3.

No firestarter: 142,33 (????)
Firestarter: 144,44 (!!!!!!)

And if you get an unlucky mana tick, or need to move for a mechanic and lose a F4? 140,91 pps. So, this is actually pointless.
(Link to the table that I cannot embed here to save my life: http://imgur.com/7ENn04N)

Now, the astute observer might look at the overall cast times for the 3.X rotation and say "but Galvuu... you won't have Foul up every time!".
But you will. First, I'm not accounting for TC procs. You can Thundercloud without loss of any spell in your AF/UI rotation for the 3.X rotation. Each extends the duration by 2.5seconds, thus bringing you close to a Foul (2 procs = Foul).
Second, you can further mitigate this by using a B4 opener and also using B4 whenever you have Triple and Ley Lines up. Why? Because the overall rotation time for the 8/6 F4s is greater than 30 seconds- meaning that you get "extra time ticking on Foul".
This is why the top parse does the extended opener. To get more leeway on Foul. Any movement/transition favours the 3.X rotation.
To sum it up: the 3.X rotation is 1~2 pps ahead the 4.0 one with no TC procs, and will gain a pps advantage per each TC proc on the AF phase. Further, by doing a 8x F4 opener and using B4 if you have Triple/Ley Lines up, you more or less can always Foul in UI, thus negating its sole disadvantage. Also, no mana tick issues.

And now, some credentials about the author; this is a random run I did farming my Susie EX staff last night with random people on the PF:
https://www.fflogs.com/reports/bH42t...iew=analytical
https://www.fflogs.com/reports/thpyq...1&type=summary
Both ilv315, and there was no padding. Just random pf clear/farm parties (like to help qq).
I'm not posting this to brag- no, I'm posting this to show I'm a good BLM.
Not a perfect BLM, but a good BLM.
And why does this matter towards my point and the overall picture?

It's not realistic that my decision making and mechanical execution will improve my performance much more.


I'll try doing the 3.X rotation from here on, but how much will that bring me? 100 dps?

How is it good design that most new tools of our kit penalise us?
How counter-intuitive is it that performing perfectly on the hardest rotation hurts our damage, when damage is all we have?

This is me venting by now- the facts are over.
But in 3.X, I struggled with these, and I remember how I felt when I pulled 2.4k on A12S in a party with no Bard. I felt amazing.
Looking back at these runs from last night, all I can think is "why am I doing this?".
And that's not a good feeling...
Back to my Mhachi cave now- not so fast!
For now I bring...


- Potential solutions to these problems:

We have four main issues in our hands- reliance on mana ticks, insufficient AF/UI leniency, insufficient dps gain VS the much easier 2.X rotation (and an actual dps loss with the new rotation) and overall weak single target dps given our diminished raid utility.
That said, I do not want to make BLM easier to execute- which, imho, is not only part of our identity but crucial for our balance if we're to compete with MNK in single target. Nor do I want to change the fact that dealing with AF refreshing is the "hard part" of the 3.X and 4.0 rotations (as Fire IV does not refresh it). I wish to keep these difficulties mostly intact, and thus reward a player that masters BLM in a practical setting. That said, I also wish for a player who messes up and ends up in the 2.X rotation to retain most of BLM's dps (~85%, as I'll show soon). Thus, I suggest:


- Raise our base MP by any number between 600~1000
This addresses the mana tick issue and changes nothing (we don't get an extra Fire/Fire IV on the AF cycle, so it's not really a buff). We shouldn't have to struggle with this rubbish (that's completely out of our control and very obfuscated) when no other job has to. Makes the job much smoother, too;


- Return the 20 potency we lost on Fire IV and Blizzard IV
This makes the 3.X and 4.X rotations slightly better than the 2.X one. Not a lot, mind you (<10% dps gain...), but at least it justifies mastering and using this harder, more punishing rotation. Our personal single target dps also increases a bit (around 5%... so not a lot, but...);


- Increase AF/UI duration by 1 second
14 seconds gives us a bit more leniency to deal with mechanics and such. It may not seem a lot, but it brings us closer to the 3.X windows;


- Changes to Umbral Heart: gains in damage and AF refresh duration
We're still faced with the issue that the old, easier to perform 3.X rotation slightly outdps's the 4.0 one. Not only that, but it can fit TC procs on the AF phase (better movement and more dps) and is easier to execute. So how can we address this?
Make Umbral Hearts have another effect- if you cast Fire IV with a Umbral Heart stack, it increases its damage (I was thinking by 20~30 potency, which is around a 10% damage bonus). If you, however, use it on a Fire, you get a bonus 2 seconds duration on your AF refresh instead. What's the point behind these?
From a pure dps perspective, the F4x3>F>F4x3 rotation will now become the most powerful one, beating the 2.X one by about 18% dps if it gets 2 Firestarter procs, and beating the 3.X one by a bit under 10%. This is the hardest one to pull off, and therefore should yield the greatest reward.
However, if a mechanic forces you to Fire early- thus doing F4x2>F>F4x4- the extra 2 seconds ensure it's possible to get all 4xF4 off. You will lose on damage (since the Umbral Heart would only buff F4), but you can at least salvage it by getting your 6 Fire IVs off.
This also gives us depth- you adjust your rotation based on mechanics (and possibly procs, since now you could do F4x2>TC proc>F>F4x4, for example). We'd be back to HW in terms of rotation options


These are the important suggestions, however, I'd also like to add...

- Make Foul 680 potency
"But why? It's already pretty strong!"; if you also do this we're about the same level as MNK in single target dps. Still under SAM, but at least now I can justify my existence in a efficiency-based raid group. This is the least important one, though, since it doesn't affect anything much.


Where would these changes leave us?
The 4.0 rotation would outdps the 3.X rotation, which would outdps the 2.X one. The gap between 4.0 and 2.X is less than 20% (on average, around 16%), which seems fair to me, given the greater difficulty of execution.
A BLM ignoring all this junk would still end in 85%+ percentile, so I wouldn't say that the gap between a good and bad BLM is huge (none of that 40% HW stuff).
Further, our rotation would become smoother and we'd handle mechanics more fluidly, as we'd have more options to approach any given situation.
These are all relatively minor, easy changes. They'd also bring us to like 3rd single target dps (slightly behind MNK, unless I messed my math up).


Edit: the corollary is this: ignore this stupid, clunky 4.0 rotation at all costs. Hell, do the 2.X rotation and you'll lose ~10% dps over the course of a lengthy fight while having quasi-RDM mobility and no burden of decision making. Well-designed job (my god I'm so salty about this. Started in that dumb reddit thread).

Double edit: for some weird reason the Reddit folk included B4 in the 3.X rotation even though we don't need that (which brings the pps a bit lower). That said, there may be Foul timing issues with that (though I'm guessing you can just push it back and then double Foul in a UI cycle; since you only Foul and T3, you have plenty of time to double Foul).

Triple edit?!: so I don't want this to be all doom-and-gloom, so I added a section with possible solutions. I know this a very lengthy post, but there was a lot to be said about BLM. We're broken in a slightly less obvious way- our new tools hurt us. We gained things in 4.X that have no place in the optimal BLM world. How sad is that? That one of our new main mechanics is useless 90% of the time xD?

Originally Posted by Galvuu

SNIP

Hey, I've also been lurking at that reddit post and found some numbers. (http://i.imgur.com/xTF8ELG.png)

This chart is a comparison between the old 3.0 rotation (no Blizzard IV) vs the new 4.0 rotation. Apparently it's assuming perfect mana ticks and ignoring some constants that are present in both rotations (like Foul, Enochian).

The results are:

132,8181818 PPS (3.0 rotation - No Blizzard IV)
132,8387097 PPS (4.0 rotation)

This is less than 0,1 PPS difference between the two.

Still, you have to take into consideration that you won't have Foul up for every UI cycle if you choose to use the 3.0 rotation meaning you might not have full mana after using only Thunder III during UI.
But at the same time, you're still at the mercy of mana tics by doing the 4.0 rotation since you may land at 0 mana after doing all 6 FIVs.

But the thing I really don't like is that the new rotation does almost the same damage as the old one, but is way more difficult and tighter to sustain.

Originally Posted by Lilyth

snip.

Corrected everything and put the link to the table with everything on it so the numbers are crystal clear.
Also that Reddit thing has no Foul anywhere- the number is a bit silly (why wouldn't you have Foul?).
One thing that became apparent is that the Foul issue that you brought up isn't a problem at all. You just do the 8x Fire IV opener (like the top parsing BLMs do) which puts the next Foul on a 10s timer by the time you go AF.
If you get one proc per entire rotation (TC or FS, which you probably will), you're 2.5secs behind each time.
This means that by the time you've "caught up" to Foul, you'll have your 8xF4 opener again, and this can extend the timer another 10s.
This is, ofc, discounting that any movement will further alleviate this issue, and that if you're on proc-happy land, you also get more time to work with. The Foul problem should realistically almost never happen.

Originally Posted by soslinky

@Galvuu
Would 4.0 rotation win out over 3.0 if the potency of F4 was back to the pre-nerf value of 280?

It would lose narrowly. It'd be 148.78 (4.0) VS 151.67 (3.0).

Originally Posted by soslinky

snip.

I have the tables up on the post now, you can check them yourself (like, please do to make sure I didn't goof).
But those numbers are weird- is that using the "nerfed" Fire IV?
Note that both rotations use the same F4x8 opener (no reason not to, and in fact, 3.X should always do this to ensure they get Foul aligned for UI every time).
Further note that you can have Foul up almost always on 3.X if you do the extended 8xF4 opener and use B4 if you have Ley Lines and Triple up (Triple bypasses the movement issue)- this is to "buy time" on the Foul clock.

Also, do note that discounting TC procs hurts only 3.X. 4.X cannot fit them (trades a F4 for them in AF for a marginal pps gain), but 3.X can fit them without penalty, so that further brings it ahead (and gives it extra movement).