This is a topic which I've touched upon before, but just as quick refresher course, here are the generation numbers for all of our spells. Wrath generates 13 Eclipse points per cast and an additional 4 on critical strike. (It is interesting to note that the combat log quantifies this as a negative number and separately; upon casting Wrath, the combat log will read "Murmurs gains -13"; if the spell crits, it will then read an additional line of "Murmurs gains -4".) Starfire generates 20 Eclipse points per cast and an additional 8 on critical strike. At its baseline, Starsurge generates 10 Eclipse, or 16 with the Lunar Guidance talent. Starlight Wrath once again reduces the cast time of both Starfire and Wrath by 0.5 seconds. This means that, discounting the additional gains from Euphoria, Wrath has a base Eclipse generation of 6.5 per second, while Starfire has a base generation of 6.66 per second.
An Eclipse proc will occur at either -100 or 100 on the Eclipse scale. This means that to rotate from one Eclipse proc to the next, a player must generate 200 Eclipse. For Wrath, this should take approximately 31 seconds' worth of casting -- barring any additional time spent on DoTs, cooldowns, movement, haste or additional gains from Euphoria. For Starfire, it would take approximately 25 seconds, again with the same stipulations. Starsurge is a bit of an oddity when it comes to Eclipse generation as it generates "smart" Eclipse rather than flat gains as with Wrath or Starfire. In the case of Starsurge, Eclipse will be generated in the capacity that is most useful -- if you are currently under the effects of a Solar Eclipse proc, then the bar will be moved back towards that proc (essentially registered as a negative gain in the combat log), while if you are moving towards a proc, it will move in that direction.
Predicting Eclipse gains
Now, Euphoria is an interesting little mechanic to a degree. The function in order to predict the gains from Euphoria is rather simple. All you must do is multiply the additional gain by your critical strike chance; to get the whole formula, you merely add it to the base gains of the spell. Therefore, the expected gains would be as follows:
Wrath - 13 + (4 * critical strike chance) = average Eclipse gain
Starfire - 20 + (8 * critical strike chance) = average Eclipse gain
A simple example: With a 50 percent critical strike chance, the average Eclipse gain of Wrath would be 15 and the average Eclipse gain of Starfire would be 24. While averages are nice for creating predictable, theoretical rotations, the unfortunate aspect of RNG is that averages occur over the course of an extended amount of time. RNG won't really equalize itself until thousands and thousands of spell casts have been made, far more than what would occur over the course of a single encounter -- far more than what would occur over the course of an entire raid, for that matter. This is slightly problematic due to the way in which RNG can still completely work against us.
While critical strikes are a fairly major increase in damage over the span of an encounter, where they fall within our rotation can actually be somewhat detrimental to our damage -- or perhaps not. Getting a string of crits during an Eclipse proc will consume that proc faster and thus cause you to lose it sooner, but then again, it still pushes you toward the next proc faster. Without more data, we cannot really predict how this might influence our damage. Even still, I am not too keen on finding out.
A "smart" proposal
Instead, there has been a proposal to rework Euphoria to generate "smart" Eclipse as Starsurge does. I wish I could call this suggestion my own, but, alas, I cannot. I believe the first to make the suggestion was Relevart, but I could be mistaken (and if I am, then I apologize). Either way, the plan is brilliant. Euphoria needs to be made into a "smart" system in an effort to make the talent actually attractive to us as a form of DPS increase.
As it stands right now, Euphoria is supposed to be a tool that adds a sense of RNG to our rotation, to make it more random, but it simply doesn't have that effect. Theoretically, we should never see a 50 percent critical strike chance in Cataclysm
, yet even with that variable, the impact it would have on our rotation is minimal. At a 50 percent critical strike chance, Wrath would, on average, generate 7.5 Eclipse per second and thus take approximately 27 seconds to proc Eclipse -- a reduction of only 4 seconds. For Starfire, the average Eclipse generation would be 8 per second, while a proc would take approximately 21 seconds -- again, only a reduction of 4 seconds. At best, with a 100 percent critical strike chance, Euphoria could change Wrath's Eclipse proc time to approximately 24 seconds and Starfire's Eclipse proc time to 18 seconds -- a difference of 7 seconds for both.
Having Euphoria work as a "smart" gain would have a much larger impact on our rotation than it currently does. Given that Euphoria gains are parsed independently of the standard Eclipse gains of our spells, it shouldn't be difficult for this change to be implemented. Clearly, the coding already exists for this form of structure, so it is merely a matter of changing Euphoria itself, which shouldn't have any impact on how Wrath nor Starfire function. Making this change would take significant strides toward making Eclipse itself more engaging, increasing the value of critical strike rating and serving to make Euphoria functional for more than just mana regeneration.