I doubt there is a 2 minute cooldown on the proc, I checked wowhead and they have no cooldown for the proc at all so it is likely MMO has listed it incorrectly. Item - Coliseum 25 Normal Caster Trinket - Spell - World of Warcraft

Blizzard developers always are wanting us to have to make challenging decisions. I think that how a player decided to handle their mana regen is one of those things, and there won't be a clear cut answer.

Personally, up until 3.2 (right now I have a strange build until I get the 2pc T9) I had nothing for mana regen, not even Omen of Clarity. I came to expect that on some fights I would end up using a [Powerful Rejuvenation Potion] or a [Potion of Nightmares], the former being used much more often. I would always choose a time when popping the potion saved our healers from a cast, and in some situations kept me alive. I would tend to think of it as a health stone that gave me back mana. The lack of mana regen talents helps balance out sacrificing a [Potion of Speed] for one of the pots that return mana, plus not dieing helps tremendously. For us Moonkin, we bring a lot of utility and buffs to the raid, so staying in the fight is a big deal in and of itself.

Patchwerk is a bit of a special case because the healers are so loaded down right at the beginning of the fight, but I find that a lot of zerg fights it ends up being the DPS casters that need mana toward the end and not the healers. I have no qualms about Innervating myself if the healers are doing fine, especially with the 3.2 change to Innervate.

At the core of all of it, I really think it comes down to figuring out what works best for you and your raid. A quick solution would be to drop [Glyph of Insect Swarm] and pick up [Glyph of Innervate] instead. Check with your raid leader, but I'm guessing they would prefer if the tank, and to an extent the entire raid, was hit 3% less often at the sacrifice of a small amount of your overall DPS.

Advocating "do what feels right, not what math says" isn't going to get far here. What is the point of theorycrafting if you aren't ever going to use it?

You bring up a good point that the DoTs are doing relatively less damage than nukes in 3.2, but the notion of "strategic times" for refreshing them is a lot less clear than before. I'm struggling with this in my head...

HEAR ME MUSE:
In 3.1, it was easy to know that it was worth it to cast dots when eclipse is on cooldown by simple comparison of a normal starfire's DPET to MF/IS. Similarly, when eclipse was up, you could compare the damage dealt by the dot ticks during the eclipse to the difference between an eclipsed starfire and a non-eclipsed starfire to know whether its worthwhile to cast dots.

Now in 3.2...
There is no eclipse cooldown, so DPET for non-eclipsed W/SF must include the dps gained by the chance to proc a new eclipse. (I think everyone agrees that a starfire that can proc eclipse is worth more dps than one that can't.) I haven't wrapped my brain around how much dps a 3.2 non-eclipsed nuke is worth, but I'm hoping hamlet has been thinking about it.

During eclipse, I see things getting pretty crazy too. Choosing to cast a dot during an eclipse means of course that you didn't cast an eclipsed nuke, and that instead you refreshed the dot after the eclipse ended... so you'd have to compare an eclipsed nuke to a non-eclipsed nuke with chance to proc a new eclipse + the dot ticks during the eclipse.

Maybe i'm making this too complicated, but it seems like this topic hasn't been addressed enough. Any ideas?

Yeah, I hope so

Moonfire (if extended) is better DPET than Eclipsed Starfire. Before the idol or 2T9, I can see it being quite marginal to cast during the Solar part of your cycle. But once you have the Idol, I don't see how there would be any question.

I agree with that sentiment, but sometimes the math is just too complex or there are too many variables that the answer isn't clear-cut. With regards to the above discussion, one of the hardest decisions to make are whether or not to refresh dots where there are semi-predictable movements. Examples include Ignis flame jets, IC rune of power/death, Mimiron p2/p4, etc.

For example: your MF/IS just fell off. Ignis is going to cast flame jets in the next few seconds, do you recast MF/IS now or just wait until you get launched into the air? You can simulate the example exhaustively, with scenarios like your eclipse status, cooldown on your other instants, etc. but sometimes it's better to just have a general rule of thumb.

If the DPET between a nuke and refreshing a dot is very similar, it's pretty easy to see situations such as movement where you obviously will get a higher overall DPS if you cast dots in the times where you can't nuke. I think figuring out that threshhold is the tricky part and is much more difficult to model.

Agreed, 100%.

My bent is that, right now, MF just isn't worth it most of the time, almost all of the time. However, with the new [Idol of Lunar Fury] and 2pc T9 bonus (MF ticks can crit), MF can be brought back into the rotation.

Let's keep in mind that in order to make MF worth casting in any situation, it requires two major glyph slots and 3xSF casts which is not as convenient as it was before. For a Moonkin working their way up in 10 mans or 25 Naxx before the induction of tons of +Spirit on Ulduar gear, I would go with replacing one of those glyphs with [Glyph of Innervate] until you get into a lot of T8, at least... and I'd say probably up until that 2pc T9 bonus.

IS should stay in a rotation as much as possible (something I need to work on, admittedly) because of the debuff it gives the boss (3% decreased chance to hit w/ melee and ranged attack). However, I've seen a lot of Moonkin on these forums with it glyphed, increasing their overall DPS by no more than 2%. We're designed as a utility class and especially so with a DPS role, so I have no qualms about a minor reduction in DPS to help the raid overall.

As for the rotation that includes MF with the new idol and/or 2pc T9 bonus, I'd venture to suggest that most of the time refreshing your dots right at the end of an Eclipse is your best bet. With a decent casting bar addon such as Quartz w/ the Eclipse addition, I can visually get a sense of timing as to when I'll get in my last Eclipse cast and immediately afterward I'll refresh my DoTs. At the least, MF should get 2/3 of it's 3 second extensions this way. Granted, this is a best case scenario.

The only caveat I can think of is when you're on a bad RNG string of Wraths trying to get a Lunar proc and your DoTs are close to falling off. As you aren't casting SF, MF will fall off first and with the bonuses it is getting I would recast it immediately. Because IS, via Improved Insect Swarm, doesn't give a buff to SF, I would let it fall off if Lunar proc'd with more than 5 seconds left on IS.

Then again, there are fights like Hodir in which I DoT on movement but otherwise am going to find a campfire and nuke as fast as I can. Sometimes you have to figure out for yourself what works best.

So this thought occurred to me, but since I'm not quite as experienced with the math behind determining stat weights, I wanted to bring the theory up here and see if somone else could model it.

I was thinking conceptually about 3.2 Moonkin rotations, and realized that Haste should be devalued in comparison to Crit. With the new rotation, IDEALLY half of your DPS uptime should be spent casting Eclipsed Wraths, the other half with Starfire. Due to the Haste soft cap, anything beyond 400 hurts half of your rotation. Crit on the other hand not only makes the Wrath end of the rotation stronger, it makes back to back proccing far more viable.

In an ideal world, the entire DPS cycle would be spent in an Eclipse, and I'm willing to assume that that is the highest DPS potential of the spec right now. So regarding that, if I'm not missing any major loopholes, how would that affect the value of Crit vs. Haste now?

I saw that it was less valuable, which after the soft cap I had always assumed was the case. But does crit not gain any additional value considering 3.2 changes? (.9 vs. 1.1 seems so much closer than I was expecting)