The original Eclipse goal was to make us use both SF and Wrath in PvE rotations:

With Cata Eclipse, we'll use both spells, and we'll spend about the same time on both spells (unless Lunar Shower becomes too OP, and we stick to MF spam).

One might argue that they put the two Eclipses on separate cooldowns, because they wanted us to use both Eclipses. My memory (can't find a quote) is that it was one of several early attempts to buff Eclipse so that it couldn't be ignored.

I haven't tried it yet, but I think it is reasonable to have choices about the Eclipse strategy:

1) Full Solar for Patchwerk (or maybe go to Lunar just for Starfall) and for AoE phases (all three AoE spells are Nature).
2) Full Lunar when lots of movement is required (buff MF).
3) Rapidly alternate Eclipses for the most mana regen (frequent Euphoria procs). This is in serious danger of being overshadowed by Lunar Shower as a mana maintenance tool. However this could be better DPS than Lunar Shower, and have better mana than parking in one Eclipse.

You might be right and they will force us to alternate Eclipse procs. I'd think the biggest argument against that would be PvP, where
1) Gaining either Eclipse will be very difficult.
2) Solar Eclipse is very weak (unless Mushrooms turn out to be much stronger than I expect).

Even with 100% solar you have to switch between Wrath and SF all the time.

But there is no way you want change between the eclipses in a Patchwerk fight unless there is a penalty when you don't switch. But then you will have the eclipse which does more dmg for 45s and run through the other one.
If it doesn't matter, because Wrath and SF and MF and IS do the same dps you will always lunar in a fight where you have to move.

Ok, let's assume that they intend for us to alternate Eclipses. Now, the easy way to do this would be to force it--after the first Eclipse in combat, only allow Eclipse to alternate until combat recycles. This is how it worked the first time the Twitter Eclipse went live. I actually objected to this the next time I gave some feedback, just thinking it would be irritating if I were interrupted for a long time or for whatever reason was still on the near side of Eclipse after playing naturally and had to move all the way across the bar. Also, I thought that allowing reproccing of the same Eclipse might lead to interesting gameplay decisions. I assumed that a longish cooldown would achieve the desired effect just as well (of making sure that the optimal stationary rotation uses both Eclipses). It seems this has caused more trouble than it's worth though.

So do we just need a longer cooldown? Or does it really even matter--maybe there's nothing to be gained from being able to proc the same Eclipse multiple times except for a lot of unintended complications in the rotation.

One other solution I've mentioned before which might be more elegant than using cooldowns at all is to make Eclipse clear slightly beyond the midpoint of the bar, rather than at the midpoint. That way you're always inclined to move the shorter distance across to the opposite Eclipse. The only real downside I see is a further increase in Eclipse uptime, but it's not clear that that's a major problem. I even suggested at one point that this could be the function of talents like Planetary Alignment or Euphoria that have lost their purpose in the new Eclipse model, e.g., "Planetary Alignment: allows Eclipse to persist until you have 10/20/30 of the opposite type of Energy."

Technically, doesn't it already do this, as you still get the benefits from Eclipse for the first spell cast once you're on the wrong side of the bar? I'm not seeing how extending/forcing this behavior is going to change the apparent trend of trying to cycle within a single Eclipse. Really, the only reason why that's preferable to oscillation is because Nature spells are currently tuned to be higher DPS than their Arcane counterparts, and I don't see that persisting to launch intentionally. A really ham-fisted solution might be that when you have Solar Eclipse, arcane spells generate 0 solar energy (making Surge the only method of extending Eclipse), and similarly with nature spells while Lunar Eclipse is active. You could still stall Eclipse and only discharge it when it's convenient to do so (i.e. if you have to move), but it puts a finite benefit on each Eclipse proc.

The next hassle is going to be that you'll always want Lunar Eclipse up when moving if it affects Moonfire. That's going to cause a lot of grousing if they enforce oscillation, but it's not going to have a simple solution. I don't see them applying Lunar Shower-type effects to Insect Swarm.

I don't know if you're on the beta, but my experiences with the rotation are anything but boring. On paper, it sounds simple - keep up both DoTs, keep Surge on cooldown (there's probably going to be an art to this actually, determining when's the best time to surge, but it's going to be very close to cooldown in any case), and otherwise spam whichever nuke is either being buffed, or is going to get you the next buff quickest. Pop your major cooldowns when they're both up and buffed.

In practice, with durations of our DoTs closely matching both the cooldown of Surge and the natural period between Eclipses, it's actually fairly complex. Especially since the period of Eclipse is somewhat randomized. If you pop Surge just off cooldown, you may waste some of the energy it generates, especially if you pop it right after an Eclipse proc. You'll want to wait until Eclipse procs to refresh DoTs, while minimizing downtime and number of clipped ticks. Yes, you're going to have longish (~7-8 seconds, maybe) periods where the only thing you're doing is spamming a nuke. But in between, you're going to have very short periods of time where there are 4-5 valid options for which button to press, and the priority of those buttons will vary depending on whether you're charging or draining, which Eclipse is active, and how far you are from the endpoint.

I'll grant that it's been about a year since I raided, and my experiences are based only on target dummy work at level 80, with a half-functioning SquawkAndAwe and no Insect Swarm (for some reason pre-mades didn't get loaded with it, and though the spellbook is telling me to see a trainer, the trainer isn't offering it). It's possible that once the rotation gets burned into muscle memory, it could get pretty boring. But I'm hard-pressed to see why choosing to stay in one Eclipse* is going to make life any more interesting. You're cutting out the major factor that changes up our rotation on a regular basis. If you always have Solar Eclipse up, a large number of those interactions go away, and then you're back to a very simple priority list.

* For any reason; i.e. even if your idea is stay in Solar when nuking, Lunar when moving, it gives you a little bit more of the medium-term decision-making (for only one decision), and completely guts the short-term decision-making.

I was perhaps a bit unclear.

If we are forced to move from one eclipse to the other then the decision to choose what eclipse to proc is removed from us. We need to go from Eclipse A to Eclipse B back to Eclipse A and so on. I am not saying that the tools we have in beta is boring as today, only the eclipse rotation is as boring, as it is today on live.

The decision for us to choose what eclipse we want to procc should be based on what will happen during the next 30 seconds in the fight. Will we move, will we AoE etc. That is something i would like to be implemented, not "oh heck i had my Lunar Eclipse so now i need to proc Solar".

I agree that we now have more tools to use, i was only refer to eclipse "rotations".

You missed my last point. By gaining that one decision in the medium, 30-45s term, you lose a ton of decisions in the short, 5-10s term. While I can perhaps stand behind the idea that a player should be able to make that trade if they wish, the more optimal play style should be the one with the short term complexity. Yes, on paper, what you're suggesting is a little more interesting. In practice, it would be more boring.

More notes:
--Looks like only Spirit from gear counts for BoP. The screenshots linked above seem to show that base Spirit at level 85 is 173.
--Glyph of Starfire tooltip still says that the extension is 3s, not just one tick. Can't tell yet because haste isn't taking effect.
--AUGH they put back in an IIS effect (Glyph of Wrath). Well, I'm not going to put in the elaborate machinery again; I'll just assume 100% DoT uptime for now.
--Starfire seems to have a damage range now.
--Lunar Eclipse ends one full spell after crossing the midpoint. In other words, if a Starfire leaves me at 10 Solar energy, Lunar keeps ticking. It goes until I complete another spell that's either a Starfire or a Wrath. I can cast Moonfire during this period and it gets the bonus, or I can cast Starsurge and it generates Solar energy, both of which leaves Eclipse running until I cast one more (buffed) Starfire. It's possible to end Lunar Eclipse at over 50 Solar energy this way.
--Solar Eclipse does not work this way; it ends as soon the Wrath hits and generate energy that takes me past 0. I can still cast an IS during the final Wrath's travel time and have it buffed, but that's all.
--E&M is still giving 13%.

----

Okay, here's today's spreadsheet update. Put in glyphs, set bonuses, simple talent changes, basic effect of Eclipse on DoT's. Next project is probably more rotation options. First I have to figure out a good way to model Starsurge even in the basic model though (the basic model--which I suspect will become the standard rotation even if Blizzard has to force it somehow).

----

The biggest piece of missing data now is base spell values at level 85 (and should probably check coefficients again too). Anyone who's 85 and willing to help can help out there, but ideally someone who has a good sense of how to collect useful data. We basically need damage values for each spell at two very different spellpower levels. Completely untalented/unbuffed/unglyphed is ideal. Unlike at the beginning of this thread, tooltips for most spells seem to no longer contain damage values, so you have to actually test at a dummy.

The most important tip for collecting spell data is that when you're repeatedly casting some spell with a range, only the min and max values you observe matter; there's no need to record every single hit and take a mean or anything like that. Just cast a whole bunch in a row and note the min and max.

Anyway, if anyone could test all our spells that would be great. Again, note absolutely any buffs/talents/glyphs/procs of any kind that you have for each spell, and record any spellpower value. Two different spellpower values for each spell is probably fine. Testing multiple combinations of buffs/talents to confirm they all stack like we expect can probably wait until later.

Ok, got that big spreadsheet update in.
--Refined the standard Eclipse cycle, including much better modeling of Starsurge and its energy effects.
--Added options for more Eclipse cycles.
--For right now, all the fancy options for using spells at certain times (DoT's, Starfall, Starsurge) only work in a "standard" Eclipse cycle. Customizing specific spell times for each possible Eclipse cycle would have been a pain--I'll probably add some in if non-standard Eclipse cycles show themselves to be worth examining in detail later on. For example, casting IS 4 times during a "stretched" 45 second Solar, but at no other times--stuff like that. But by and large, DoT's and Starsurge all seem like good enough spells that "use on cooldown" is going to be optimal anyway.
--If one Eclipse (e.g. Lunar) is set to "stretch" and the other to "standard," it assumes you manage energy during Lunar so that you wind up in at the center of the bar right as 45s expires, then proceed towards Solar. If instead Lunar is set to "stretch" and Solar to "unused," then it assumes you manage energy during Lunar so that you're right at the Lunar end of the bar when 45s expires, to proc it again.
--Starfall can be used on cooldown or once per cycle (during Lunar). This is slightly sketchy since the cycle length will dip under 60s at higher levels of haste, but good enough for now.
--If you use one of the specific cast timing options for Starfall/Starsurge/DoT's, it now correctly adjusts for the proportion of casts that will be affected by the Eclipse bonus. If you use "on cooldown" for any of these spells, it uses a mean Eclipse uptime computed from the rotation.
--Note that this means DPET of each individual spell on the front page (these were moved to the front page) will reflect partial Eclipse uptimes. Maybe I should just put uneclipsed values there. Eventually I should break out all other spells into Eclipsed and uneclipsed versions, like it's always been for Starfire and Wrath, but this is good enough for now.
--Right now, SS is given priority over MF, and MF over IS, if all are set to "use on cooldown." I'll have to figure out some way to make that customizable, but it will be tricky.

----

As far as actual interesting results so far: I won't speak in too much detail yet since we don't have base damage values for all the spells. But,
--For all the work playing with different options for Eclipse, they don't have a drastic impact on total DPS. With the new Eclipse model, you can't fundamentally do too much to affect your Eclipse uptime, since you more or less generate and consume Eclipse energy at the same rates regardless of how you rearrange it. The only real variable is Starsurge.
--I tried to model the fact that you go slightly past 0 energy as you end each Eclipse. This does slightly disincentivize turning around after each Eclipse and eliminating one Eclipse entirely.
--Solar is the slightly stronger Eclipse (I think because Starfire generates energy a touch faster than Wrath generates energy). So the candidate for elimination would still be Lunar. But "stretch" Solar and/or "unused" Lunar are showing to be a near-wash for DPS purposes.
--Sticking with the standard Eclipse setup, the best spell usage seems to be IS/MF/SS on cooldown, and Starfall during Lunar Eclipses.

Note that the length of even the standard Eclipse cycle is 67s. There aren't that many times where you shoot at something for 67s. This was generally true of the WLK Eclipse cycle as well, which is around 45s, but even more so now. Managing DPS correctly will always entail some degree of improvisation, just as it does now. The goal of the sheet or any modeling tool is to get basic principles: when DoT's/SS should be refreshed, and the most efficient way to traverse the Eclipse meter. That's part of the reason I'm not worrying about things like sequencing exact DoT casts into the Eclipse cycle.

Anyway, people should be able to have fun with for a bit. The model will probably need some tweaking--I changed the guts of the rotation quite a bit today so be on the lookout for possible errors. But all the basic elements are in place; once I have the base spell damage values it should practically be ready for live.

Note that Wowhead has spell tooltips that they show as being 85. The seem to be at least a little larger than your 82 numbers, so there is at least a chance that they're right. If somebody could confirm one or two, they'd be an easy source for base numbers.

I think confirming spell coefficients (at least to within a couple of percent) is probably more important for meaningful theorycraft.

It does show one coefficient: Thorns: 446+SP*.421. Thats enough to easily see Thorns hitting for over 2k in a raid.

About thorns, I just tried it in a duel on my level 85, while in feral spec;
With 3703 spellpower it hits for 1348.
With 151 spellpower (gearless) it hits for 343.

And versus a few level 9 raptors;
With 3703 spellpower it hit for 2006
With 151 spellpower it hit for 511

I also ran some with gear, without gear tests for each spell (versus the training dummy), but the resulting log is a bit messy, and my parser does not take it like I expected.
I'll rerun the test and note the values manually later today, if anyone wants to have a look at the combatlog regardless, you'll be able to find it at http://mnavers.net/WoWCombatLog.txt , my player name is Kongz.

Also, from what I understood you'd rather have me doing the tests in feral spec, so as to get base values right?
This would mean only the non specced spells get tested. (wrath, starfire, moonfire)