Serenity is a potency barrier, which means it needs to be applied to a spell before the system checks that spell's potency. To have Serenity apply after the hit is determined would make it worthless--it would be lowering a number that is no longer relevant.

Since potency plays a role in accuracy (or the stat contest, for contested spells) this means that Serenity applying before the hit is checked makes the spell less likely to hit at all. If you want protection that is checked last, that's what Tenacity and Toad are for.

As for Serenity being vulnerable to spam snap casts, I think you'd be on firmer ground asking that low potency spells cause less IF loss.

I'm on the fence with Serenity.

On the one hand, my testing indicated it to be an exceptionally powerful potency barrier that probably needs a downside like other powerful barriers such as Aether Cloak and Ghost Shroud. On the other hand it's possible to be played around with such an easy cheese strategy, but the cheese is more a function of the game's spell-casting mechanics and how mana pools scale to high level, where min-casting a spell over and over again as fast as the game engine/RT will allow is a trivial cost at 1000+ magics (not infinitely, but well within the margins of a typical PvP fight).

I think the inner fire hit aspect should be removed completely, and the negative modifier to roars should be increased substantially, even to a point where roaring for all practical purposes is an ineffective strategy. That puts it more in line with AC, where a powerful defense has a commensurate cost to the one enjoying the barrier's protection.

I'm not sure it's a 1:1 strategy, you still have a class with a huge potency barrier that's weapons primary, but frankly I think Barbarians could use something that's a little broken in PvP. Backstab is broken, WM TM is broken. If there's one thing that should be broken in PvP for Barbarians, it's magical defense.