Poofah wrote:Crush is 296, Sandstorm 370 (ie 296*1.25), and the shields are both 74 (ie 296*0.25). If you're slower, then you have to reapply Stoneskin every 5 turns. So on average
Sandstorm, Crush, Crush, Crush = 296*3.25/4 = 240.5 damage per turn
Stoneskin, Crush, Crush, Crush, Crush = 296*4/5 = 236.8 damage per turn
This is assuming that the Sandstorm shield is always up, whereas in practice you'll get full damage out of your first Sandstorm, so that pushes the Sandstorm damage cycle up a bit more.
If Sandstorm has a value of 1.25 Crushes, then a Sandstorm + 3 Crushes should equal 4.25 Crushes not 3.25, right? That would change the numbers to:
Sandstorm, 3x Crush = 296*4.25/4 = 314.5 damage per turn
Stoneskin, 4x Crush = 296*4/5 = 236.8 damage per turn
At first, I thought the 3.25 number was factoring for misses caused by Sandstorm, but if the Stoneskin case is using 4 then you aren't adjusting it for any kind of base chance. Therefore the only miss chance that should be factored into the Sandstorm case would be the 10% miss chance from Sandstorm itself and 4.25*.9 is not 3.25. If we factor in the sandstorm and base (using 5% as base miss chance as per your section on Rupture) miss chances then we would get.
Sandstorm, 3x Crush = .85*296*4.25/4 = 267.3 damage per turn
Stoneskin, 4x Crush = .95*296*4/5 = 225.0 damage per turn
While both shields block 74 points of damage, I think there are a couple of other factors to consider when comparing their effectiveness. The first is that Sandstorm should add to the miss chance of the opponents attacks causing some to miss that otherwise wouldn't. While the benefit can only be modeled on a pet by pet basis, especially when a secondary attack like Moth Dust has invovled, it is true to always say that the damage reduction offered by Sandstorm is always going to be better than that from Stoneskin.
The other factor is that a Sandstorm hit will trigger the Humanoid passive healing you 69 (IIRC) health that you wouldn't get on a Stoneskin turn. On the other side, the greater miss chance from Sandstorm could cost you some of those heals. I suppose the per turn average is not that really that different, but with the granularity of combat, and the Idols lack of speed, the absence of a heal on a Stoneskin refresh could decide which pet kills the other.
Sandstorm, 3x Crush = .85*69*4/4 = 58.65 healing per turn
Stoneskin, 4x Crush = .95*69*4/5 = 52.44 damage per turn
Its unfortuneate that Sandstorm nerfs our Ruptures chance to stun so much, but I would never make a tamer strat that relied on a 25% proc chance anyway. While the average dps loss to Rupure vs a Mechanical from Sandstorm might close the dps gap some, you might just lead off with Rupture and lose one turn of Sandstorm to mitigate it.
The only reason I can see to ever use Stoneskin is if you didn't want the Weather effect. For example, Aki's Stormlash uses Lightning Storm which adds an extra bit of damage to all attacks; using Stoneskin to block that extra damage instead of changing the weather to Sandstorm lets your attacks keep that extra damage.
Setting the weather to Sandstorm is also be a problem whenever you want to follow the Idol with a pet that uses multiple attacks per round mechanics whose damage gets nerfed. The other side of that, though, is usually another point in Sandstorm's favor; that it can continue offering protection to your team after death or swap.
In general I think Sandstorm is definately better than Stoneskin. For Obalis specifically, I'd go Sandstorm + Rupture. The only things you'd really want to Deflect are Moth Dust and Counterstrike (which isn't always predictable). If the expectation is that the Idol is coming in to finish off a weakened second pet, then I'd rather have the option of Rupture for extra burst damage to try to kill them quickly.