How extensively did you re-optimize each profile post-ilvl jump last time around?
I did a quick and dirty run with raid_t15h.simc and global_item_upgrade_level=15 to get all of the profiles up to ~600 ilvl, and here's how things stood:
http://dread-gaming.com/simc/raid_mop_18.html
(live settings - can't get raid_t15h.simc to run with ptr=1)
... but there are some obvious issues with that. Every profile is quite a bit over the hit/exp caps, so dual-wield profiles will benefit from this extra ilvl more than 2h and caster profiles unless I go in and fake-reforge every profile back down to their caps. Additionally, the stat priority that feral druid and windwalker monk profiles established for rune of reorigination procs was ruined by the scaling (both now have haste > mastery, whereas before they had mastery > haste) - likely because the stats they have from gems aren't scaling along with the stats on the rest of their gear. Again, fake reforging would be necessary to correct this imbalance.
I'm sure those aren't the only inefficiencies introduced by this leap, but I'm not a qualified expert for every spec in the game, so those are the biggest changes that stand out. Still, I'm not sure that fury could optimize itself enough to close the ~40% gap between it and fire (before re-optimizing the fire profile, mind you), and I'm not convinced that Fury was nerfed enough to explain away that gap if what you guys are saying about fury scaling is true.