Which Fairy IS better?

[cainejw]

Just a note to those who come here, Rainsford is plugging this incredibly bad theorycrafting work as somehow proof that Selene is always better than Eos. As someone who has worked with and taught statistical analysis for a decade, do not buy it. There is no theorycrafting on faeries...

But thanks to Rainsford, I'm going to start.

[Rainsford]

*snip(

I'm not sure what exactly you want. Tell me how you would improve upon it? You can't chart out healing like you can damage because spamming heals just gives you lots of overheal. I hear "Eos is better for heal intensive fights" a lot though, and it's just not the case. 20% for 8 casts over 2 minutes if you chain-cast heals is what Eos gives you for Coil content. That's the equivalent of 1.5 extra Physicks. Again, especially if you take overheal into account, Eos just isn't attractive.

For Selene, I'm using BLM as the example since that's what I have (and what I can personally test). My BLM has 466 Spell Speed. Using Purostriders chart, Fey Glow gives a 6% increase in damage.


Sure, the benefit will vary based on class, but a quicker skill activation rate brings other benefits that you can't put into numbers (a Monk keeping Greased Lightning up easier, a WHM getting a cast off before a follow-up attack that would kill a tank, etc). Why is there such a fetish with trading utility and damage for unnecessary heals?

[YuriRamona]

Just a note to those who come here, Rainsford is plugging this incredibly bad theorycrafting work as somehow proof that Selene is always better than Eos. As someone who has worked with and taught statistical analysis for a decade, do not buy it. There is no theorycrafting on faeries...

But thanks to Rainsford, I'm going to start.

Unless you bring in specific counterexamples or counterarguments, you are contributing nothing to this discussion. Citing yourself as somehow having advanced knowledge of stats does not help. In fact, your lack of substantiated claims along with your self-reference to profession is a contradiction worthy of ridicule.

Selene: Fey Glow shaves .15s off my GCD, and is up for 60s. Over that time, the decreased GCD gives me the equivalent of four free actions. Assuming both me and the WHM use all 4 actions on heals, that's a roughly 8000 increase in HP.

Eos: Fey Illumination is up for about 8 actions over that 2 minutes. Assuming you and the WHM use every action for a heal, it's 200x16 for 3200 added healing. Assuming a 650~ Embrace, that's another 1040 healing done. Two Whispering Dawns will heal for about 4200 overall. Adding them up, using Eos gives you an extra 8450 healing on the tank over two minutes.

The obvious and most glaring omission is the consideration of MP. Chain healing puts me out of MP within a short period of time.

This is the critical division between Selene and Eos. Yes, Selene boosts your rate of output by a greater degree, but Eos' effectiveness is an MP savings rather than Selene's MP sink.

Fey Covenant carriers very little utility but again, it results in MP savings rather than letting you spend MP faster.

For this reason alone, Eos is not trumped by Selene.

I'd also like to argue, conceptually, that making fights faster is not always superior. For example, in high-movement encounters such as Titan, being able to land large cures at critical moments (Fey Illumination) and being able to AoE cure (Whispering Dawn) can be highly useful if you don't have the time to cast Adlo/Succor.

[Rainsford]

*snip*

To be clear, I don't think chain-casting is necessarily the way to go. The issue is that, with Eos, the buff is only up for, at maximum, 8 casts. If you're not chain-casting while it's up, half of Eos benefit is nearly wasted. Considering again how, in content like BC, the name of the game is to keep your tank topped off, the 20% for 20s is going to have very, very little actual impact.


While sure, it increases your healing per MP, is there a difference between a 1000 hp heal and a 1200 hp heal when the tank is missing 1000 hp? What about when they are missing 2000? in either scenario, you're casting the same number of heals. While FI is nice, what you give up to get it is a pretty bad trade-off.

For FC, when would you ever use it? On Titan, Plumes and his Astral Flow are the only magic attacks. Pre-Turn 5, I'm not sure what, if any, attacks are magical.

And for WD, like I had said, it's only really useful if you're trying to solo-heal Titan (or are healing Titan/Garuda with an undergeared partner). The HoT is too small to be depended on in progression content.

*snip*

Well I don't mean to come off so strong! And if someone really wants to play with Eos because they like a blue fairy more than a green fairy, then awesome! But I keep seeing things like "I'll switch to Eos for more healing etc" when, in reality, any healing increase is marginal if it exists at all! At the same time, people are so quick to dismiss Selene's benefits without seeming to understand. It's just frustrating >_<.

[YuriRamona]

While sure, it increases your healing per MP, is there a difference between a 1000 hp heal and a 1200 hp heal when the tank is missing 1000 hp? What about when they are missing 2000? in either scenario, you're casting the same number of heals. While FI is nice, what you give up to get it is a pretty bad trade-off.

For FC, when would you ever use it? On Titan, Plumes and his Astral Flow are the only magic attacks. Pre-Turn 5, I'm not sure what, if any, attacks are magical.

And for WD, like I had said, it's only really useful if you're trying to solo-heal Titan (or are healing Titan/Garuda with an undergeared partner). The HoT is too small to be depended on in progression content.

Keeping the tank topped off is not really necessary in Coil. It is better to have them floating between 90-95% HP with an Adlo shield up than to overcure and toss that MP down the drain. In the example you gave, it might normally take two Physicks to bump up the tank to full, whereas under Fey Illumination, you might be satisfied with one Adloquium that brings the tank close, but not at, HP cap. Hope that makes sense. (Indeed, we might even be freed to wait 1-2 seconds for the next auto-attack/skill to hit the tank, and cast back-to-back Adloquiums that free up healer time to move around/reposition/cast abilities, etc)

Again, my concern is MP. It is certainly conceivable that Fey Illumination can bridge that gap and allow you to cast less heals, even if it is only for 8 GCDs. For encounters such as Caduceus and Turn 4/5, landing large cures is critical often in the fight, and so is conserving MP.

In terms of Fey Covenant, aren't move such as Ballast, Repelling, and Vacuum Wave magical? That could be useful for ADS.
You do mention that Whispering Dawn is useful for solo heal. So you agree that Eos is not totally sidelined???

[cainejw]

Unless you bring in specific counterexamples or counterarguments, you are contributing nothing to this discussion. Citing yourself as somehow having advanced knowledge of stats does not help. In fact, your lack of substantiated claims along with your self-reference to profession is a contradiction worthy of ridicule.

The counterargument is clear, but I guess you missed it. My counterargument is that we do not know, with statistical backing, which faerie is best in every setting. Rainsford, however, is stating without exception that Selene is the best. Instead of doing that, I'm going to work on theorycrafting the faeries. However, I ran into a problem with FFXIVApp and FFXIVLogRep. The data failed to properly log, likely due to an error on my part.

So far, my work has been to validate Valk's calculator against Scholar, work on ideal rotations based on validating MP recovery information, and investigating the effect of Mage's Ballad on Scholar heals. My next stop is Eos/Selene and then food. All of which is being cataloged in an off-forum guide until the drafting is complete.

If you'd like to examine the draft, I'd love to find a way to share it with you.

[cainejw]

Unless you bring in specific counterexamples or counterarguments, you are contributing nothing to this discussion. Citing yourself as somehow having advanced knowledge of stats does not help. In fact, your lack of substantiated claims along with your self-reference to profession is a contradiction worthy of ridicule.

Mkay, got some good testing done!

In testing Eos vs Selene, I used each faerie once in Coil 4. I did not control either faerie and I kept movement to a minimum. This was to attempt to control for Embrace usage.

Now, a disclaimer is that HPS and DPS is still highly variable in recording. I'm working on trying to validate my logs more readily, but it's difficult to work this unless everyone in my group runs the same parser.

For this work, I elected to use FFXIVLogRep as it records DOTs and HOTs.

In running Selene twice, I received the following information for healers:

HPS:
Selene A - 5m 12s - 329.93 WHM, 376.06 HPS SCH
Selene B - 5m 15s - 277.85 WHM, 302.45 HPS SCH

Eos A - 6m 45s - 277.33 WHM, 423.09 SCH
Eos B - 7m 10s - 583.53 WHM, 592.19 SCH

Average HPS over these is 395.30. Neither Selene heal attempts matched this average HPS. Both Eos runs appeared to match this. This would lead one to conclude that Eos is best for HPS despite Selene's buff.
Total Healed:

Selene A - 5m 12s - 102938 WHM, 117330 SCH
Selene B - 5m 15s - 87523 WHM, 95273 SCH

Eos A - 6m 45s - 112320 WHM, 171351 SCH
Eos B - 7m 10s - 250918 WHM, 254643 SCH

This would lead one to conclude that Eos is better for overall healing, but keep in mind that the time length presents a problem here. Now let's look at the averages healed in each fight, non-crit:

Selene A - 5m 12s - 1225 WHM, 696 SCH
Selene B - 5m 15s - 1234 WHM, 751 SCH

Eos A - 6m 45s - 1314 WHM, 644 SCH
Eos B - 7m 10s - 1228 WHM, 867 SCH

This would present one to conclude that Eos's heal buff may actually help depending on other circumstances including RNG.

Crits:

Selene A - 5m 12s - 1868 WHM, 1041 SCH
Selene B - 5m 15s - 1444 WHM, 1017 SCH

Eos A - 6m 45s - 1685 WHM, 1007 SCH
Eos B - 7m 10s - 2051 WHM, 1203 SCH

Some mixed results, but this one also shows pretty decent evidence that Eos is probably best overall for healing. However, the mixed results mean that, honestly, you can't say that Eos is vastly superior to Selene in healing. There are too many potential confounds here.

For the DPS, I chose to utilize a statistical test to see if Selene improves DPS significantly across the board. The tests ended up showing me that, quite frankly, there's no surefire outcome.

For all fights, only the Monk ended up with significant outcome...and he is the most geared in the raid. The bard ended up the worst DPS due to running Ballad. Other than that, it was mix and match.

Utilizing the logger-given DPS, statistical significance at the .05 level:

Selene A had significance with the Monk.
Selene B with a BLM and Monk.

Eos A had significance with a both BLM, Monk.
Eos B only had significance with the Monk.

Utilizing an attempted corrected DPS by dividing damage done by time of fight:

Selene A had significance with both BLM and Monk
Selene B had significance only with the Monk.

Eos A had no significant outcomes.
Eos B had significance with Monk and BLM.

All of these were tested against the average DPS of each user over 4 fights for 16 DPS measurements.

As a result of my findings, I decided that much more data is going to be needed to actually see which is statistically better than the other. As it stands, and as I've stated, there is no surefire way to know which faerie is better than the other. There are entirely too many confounding variables to test. You simply cannot equalize every single variable. Even with statistically testing Eos for healing, the outcomes are mixed.

The behavior of you and the DPS is much more important than the faerie you use. There will need to be more extensive testing to tell which faerie is "the best," but as it stands right now, either faerie can provide what is needed for both healing and DPSing.