I am not sure I can add much to what you have read. I have one criticism as this report is a narrative and does not provide references on how each statement was developed. As a result I think we need to review the vast amount of claims, statements, descriptions and timelines to find support or conflict. OTOH, this is exactly the kind of event summary that can make that a manageable exercise.
The narratives of units 1 to 4 are all plausible as far as I can see. I was sceptical about the Unit 3 source for the hydrogen explosion in unit 4, but the recent photos of damage on the 4th floor are persuasive, if not conclusive. I thought that the reverse flow through the unit 4 SBGT filters was problematic. But this report explains that unit 3 was able finally to vent at a fairly high containment pressure. It clarifies that the SBGT system dampers fail open. It confirms there are no backdraft dampers to prevent flow to the adjacent unit.
From the descriptions I note a new issue that may affect other multiple unit sites. The military term is fratricide and refers to casualties to friendly forces close to a an intended target. Unit 1 explosion damaged response equipment at unit 2. Unit 3 explosion damaged additional response equipment for other units. The issue here is the physical separation of units or location of emergency hookups to shield the equipment and workers from failures in adjacent units.
On Unit 2 There are two things to point out. First RCIC probably failed due to inability to cool or rediuce pressure in containment. RCIC exhausts to the suppression pool and the combination of low steam pressure and high exhaust pressure was the probable cause of RCIC failure. I see nothingin the report that indicates unit 2 containment venting was successful. Containment failure of the torus is still a likely explanation for the "loud noise."
I had seen the reports that unit 3 had HPCI running when the preferred system should have been RCIC which failed. This is another example of an issue that needs careful investigation. The number of work-arounds and alternative methods tried by operators to vent containments, operate SRVs, provide alternate power to instruments, and to reenergize other systems was very impressive. But it is probable that some of these methods were subject to delays to figure out how to do them. In my experience, some of these methods are proceduralized and exercised at US plants.
Hope this helps the discussion.