A Storage Spaces pool carves its physical disks into slabs and allocates a virtual disk slab by slab across a number of columns, mirror or parity, with the map kept in a pool database replicated on every disk. Windows watches the physical disks and the database, and reports the virtual disk's health from what it can still do. Unhealthy with a disk missing is a two-way mirror or a parity space that can still read; No redundancy is the state past that, where Microsoft's definition is that the virtual disk has lost data because too many drives failed. In practice the drives are usually imageable, and the virtual disk is assembled from images with the slab map; Microsoft's documentation describes what Windows can do, not what a bench can.
Detached is different. Windows took the virtual disk offline, by policy, because an administrator said so or because integrity could not be assured, and on Storage Spaces Direct because the dirty-region tracking log that records in-flight writes filled up, in which case event ID 311 asks for a data integrity scan. The data and the map are intact; Detached is the system declining to attach until told. Incomplete, operational state 0xD003 in the MSFT_VirtualDisk class, means the virtual disk does not have enough redundancy remaining to successfully repair or regenerate its data: a repair should not be run, and the disks should be imaged first.
The 2020 issue is the documented case of a map fault with a healthy set. Microsoft's KB4568129 described parity Storage Spaces showing as RAW in Disk Management after the Windows 10 and Server 2004 update, noted that simple and mirror spaces were not affected, and applied a compatibility hold from 19 June 2020; KB4570719 gave a workaround; the 31 July 2020 preview update KB4568831 addressed an issue with in-memory parity bitmaps that could cause data integrity issues on parity spaces, and the hold was lifted on 13 August. Throughout, Microsoft advised against running chkdsk on affected volumes, because the volume was not corrupt; the parity space's map was being read wrongly, and chkdsk would have repaired the wrong thing onto the right data. Later reports of read-only pools after the 22H2 update and RAW parity volumes after 25H2 cumulative updates are user reports, not Microsoft-acknowledged issues, and this page does not treat them as more than that.