Originally Posted by Frenzi
|
His math says it is better to refresh MF with about 9s or more of Eclipse. IIS has very little impact on his results (a fraction of a second).
His math shows it is better to refresh IS only during the first three or four seconds of Eclipse. However that was with the old IS coefficient (.76 for 6 ticks). Plug in the new coefficient (1.2 for 6 ticks) and IS becomes very similar to MF (worth casting with as few as 9s left).
His math doesn't consider 2t8. 2t8 increases the value of spamming your nuke during Eclipse. I think that that if you plug 2t8 into his forumula, you need about 1 more second of DoT time to be worthwhile (rougly 10s left on Eclipse before refreshing a DoT).
4t8 is obviously going to favor High IS uptime. I haven't attempted to determine where it moves the break-even point for IS during Eclipse.
Graylo also doesn't consider the dynamics going into your next Eclipse. Suppose you proc Eclipse every 34s to 40s. If you refuse to cast MF during Eclipse, you have a choice between 1 MF per cycle (no clipping, average about 13s downtime per cycle) or 2 MF per cycle (clip two to five ticks per cycle).
If you apply MF early in Eclipse, you get to keep MF during the current Eclipse, the gap after the current Eclipse, and most (not all) of the next Eclipse, with no clipping.
All of the above assumes a target-dummy type of fight. If it is likely that you have to move soon, it becomes a bit better to save your instants for movement. If pushbacks or interrupts are likely, then SF (including SF during Eclipse) becomes less valuable.
If you know you are going to have to switch targets soon (say from Emalon to one of his Minions) you should look at pure DPET considerations, and cast DoTs even if it is late in Eclipse.