Zeal
A net damage multiplier is a huge deal. Same reason why savage is so good.
The stat increase is almost irrelevant. The problem is the illusion of % stats. Since the modifier is to base stats, the net effect is much less. C.dmg starts at 150% (you have to add 100 to the displayed stat to get the actual damage multiplier!) and any sort of competitive build will have it at 300% or higher. So that stat bonus is just +8% net damage. That gets added on top of a conditional +30% damage, yielding +38%.
For savage, calculating exact damage mitigation from def is really hard, but I'm pretty sure it's sub-linear. So Savage is at most a 33% damage increase.
Conclusion: it depends on the defense of the target. But in any case, the 25% cdmg is much worse than it seems, almost all the power is concentrated in the net damage multiplier.
Impulse
Gives you half a speed set (per piece) in exchange for 1/4 of a reflex set. That requires some analysis:
Let's say you have a 3 turn CD skill that you want to hit as often as possible. You need one trigger in 2 turns. Reflex has a 1 - 0.6^2 = 64% chance of triggering. Impulse has a 1 - 0.7^2 = 51%. So:
- No set: 33% skill uses per turn
- Reflex: 42.37% skill uses per turn: 1/(.64*2+.36*3)
- Impulse: 40.16% skill uses per turn: 1/(.51*2+.49*3)
Reflex/Impulse = 105.5%. for impulse to be worth it, that 12% base speed needs to give you 5.5% net speed. So it's only good if you're running champs at ~2.18x base spd or below.
Conclusion: it's really bad. Worse than old instinct, giving tiny spd bonuses at the loss of a significant fraction of its other effect.
ETA: that analysis is for long battles where average throughput matters (like hydra). It might be that impulse is ok in arena or something, where hitting certain spd thresholds is a must, and the reset just gives you a few percentage points of winning the battle.
ETA 2: Hmm, not sure about my math. Will run some simulations to double check.
ETA 3: fixed math.