Hmm, I guess I overestimated Starfall's importance in that regard. I suppose the better way to think of it was "we can really use starfall anywhere we want, so we might as well use it here where we get the most perks from NG uptime."

No. As has been elaborated multiple times, the only buff that takes effect when the spell begins casting is Haste. All other personal effects are calculated when the spell is cast. Effects on the target are calculated when the spell lands.

I had a couple thoughts on 3.2 rotations, idol swapping, etc.
1 - I feel like starfall would be best used right before a solar eclipse, because with 2t8 you have such a high crit chance during lunar eclipse that the crit chance will result in a better NG uptime than the lower crit, more casts every 3s solar eclispe. That sentence is kinda convoluted but basically the NG uptime aspect of starfall would probably help more during solar than lunar.
2 - In order to sorta simplify the idol swapping, my current plan is to make 2 new macros and bind them to shift and control D (my current insect swarm key) - 1 will swap to starfire idol, 1 will swap to wrath idol, and both will cast insect swarm of course. I'll probably refresh IS and swap to the new idol right as the other eclipse ends, giving up any damage on the casts that it takes to proc eclipse. The other option would be to refresh IS and swap idols only once you proc a new eclipse, but that gives up eclipse damage. The former will be easier to keep track of, but the latter will probably be better damage. I'd leave moonfire unmacrod and just refresh it whenever it drops.

Yeah, Starfall refreshing is mostly academic. You just want to get in as many as you can per fight.

Idol swapping requires a fixed-DoT macro. The way do it would seem to be:

Lunar (SF)
IS + swap to W Glyph
Pre-Solar (SF)
Solar (W)
IS, no swap
MF, swap to SF Glyph
Pre-Lunar (W)

Like I said above, pretty low DoT uptimes, but might work well. Another possibility is to change the second half to:

Solar (W)
MF, swap to IS Glyph
IS, swap to SF Glyph
Pre-Lunar (W)

But this is mostly for fun and not really going to be a significant DPS increase.

----

The point above about the added complexity is worth addressing. I don't agree that the complexity of boss fights means we should somehow simplify our cycle to save attention. That wouldn't fly on this forum regardless.

But, Ulduar fights in general don't allow us to maintain the ideal rotation in most situations. Our Eclipse/DoT cycle runs on quite a long time scale, so there's in practice more improvisation than anything else. But working out the Patchwerk DPS is always the starting point. For one, anytime you get some seconds to stand still and shoot something, you're trying to use whatever portion of the cycle you can. But perhaps more importantly, by working out the best cycle in ideal conditions, you learn your spells and interactions and their values better for figuring out what to do in other situations.

Hmm, yeah, will check that.

e: yeah, I think you're right.

This also catapults Glyph of Starfire to the top. Also, 2T8 is monstrously valuable (worth over 350 DPS by itself)--it will take quite a serious 4T9 bonus to give it up.

I disagree with your statement that in-depth combat analysis and applicable startegy for each and every encounter is something that "wouldn't fly on these forums". Because that would mean that these discussions are just about applying (simple) high school math to not so difficult gaming concepts. Neglecting to introduce human error into the theory just makes it a flawed theory.

I'd consider myself a bad academic if I'd stuff everything that's not pure math into a "big bowl of improvisation", without feeling the urge to reduce the "randomness" to minimal levels.
For example, getting a single moonbeam on the other side of the room is random. The spells you cast while getting to the moonbeam and while standing in it is not random, it's strategy. And it requires as much analysis as whether swapping an idol produces 2% more damage on a theoretical Patchwerk fight.

I do not consider myself so skilled in real (Ulduar 25) combat that I can toss it aside as "Mastered! lets move on".

Idol twisting is not looking promising. Probably for the best, as it would have been quite a pain to have to do that to maximize our DPS.

Unless idol twisting yields a much higher DPET for eclipsed W or SF than either MF or IS, I can't imagine a situation where you'd end up with abysmal dot uptimes. I like the concept of having 3 separate IS (one plain, one macroed for W-idol, one macroed for SF-idol), though playing around with it this afternoon has been a bit of a headache. Maybe I'm just fat-fingering spells before the unhasted GCD kicks in, but I'm sorely tempted to just stick with one idol, take the buff to Solar Eclipse, and say "screw it".

The loss in idol twisting has little to do with dot uptime and more to do with lost dps time. Napkin math:

Loss: Idol damage on ~3 non-Eclipsed Starfires, ~1000 damage. ~0.9 seconds of DPS, ~5000 damage.
Gain: Idol damage on ~12 Eclipsed Wraths, ~2500 damage.

Not swapping is 3500 damage more/cycle or approximately 70 DPS.

I'd have to be overestimating the time loss effect quite badly to shift the balance so that it is in favor of idol swaps, really. Hamlet probably has much more accurate numbers in the works.

That's pretty accurate. Here's my attempt at modeling it. (Here's generally updated spreadsheets which might be all for now since the weekend is over. Major missing thing is 4T8 in both the 3.1 and 3.2 versions).

This isn't what I said at all. Boss-specific strategy and optimal behavior in imperfect situations are both important (though I'd say that working out the Patchwerk DPS is the first goal). I was responding to your idea that we should "simplify" rotations because there's often a lot going on. The game places a pretty hard cap on how rapidly you have to execute anything--1 spell per second at most. There's an expectation at the top level that players can process things fast enough to decide on that one spell while paying attention to the environment. That being the case, analysis of human error is mostly irrelevant (and would be far beyond the scope of these discussions anyway). We want to figure out what the optimal spell selection is.

Some fight-specific strategies are very conducive to customized theorycrafting (e.g. Vezax). Some are not (Mimiron, Freya). For the wide variety of situations where you're running around a lot or switching targets, the most useful thing is for players to have some basic knowledge of how much damage each of their spells does and a few other things that help choose. Trying to model some kind of general-purpose rule probably will not go very far.