You are missing the point, completely.Furthermore, even with multiple controllers, if one fails, the entire stripped mirror will become unusable.
Or, am I missing a point?
Consider a single mirror, with each disk connected to a different controller. Now cause a single disk, single data connector, single cable, or single controller to fail. The mirror survives because it has a second redundant "side" which is still working.
Now construct a second, and successive, mirrors, identically.
Now cause one disk in each mirror, one splitter cable, or one controller to fail. Each mirror, individually, will survive because each has a second redundant "side" which is still working. Because each mirror survives such a scenario, a stripe of such mirrors survives.
My ZFS server has four SATA controllers and many mirrors (one mirror for the "root" on SSD, and many striped mirrors to form the "tank"). Each "disk" in each mirror is connected to a different controller. I can lose up to two controllers (or their associated splitters, cables, and connectors), and the system will still be running 100% correctly. RAID-Z is much more vulnerable to data loss under these circumstances -- because you would need a separate controller for each disk (or, to be fair, every two disks for RAID-Z2) to be able to guarantee similar redundancy.