I don't know if this is the reason, but supposedly their btrfs fork contains a custom integration with mdraid/lvm so that when btrfs detects a bad block, it signals lvm to do a repair. This is their solution to avoid using btrfs raid5/6 which is still marked unstable.