Exactly. They're still exploring limits and failure modes. Hence the effort to re-fly a used booster on an aggressive trajectory when they have multiple newer revisions ready to go. Looking forward to the first launch with Raptor 3.
Also really can't understate the importance of Starlink in the development program. Apollo jettisoned parachuting return capsules from the interstages so that engineers could get camera footage from inside the rocket during launch. SpaceX just gets an RTSP stream (or something similar).
It's the original definition of "pushing the envelope." You start with a known flight envelope, then push a little past that. Sometimes you learn that the real envelope is bigger than you thought. Other times you find that the envelope is exactly where you thought it was.
I notice I mis-typed and should have said you can't overstate the importance of Starlink on the development program. And I agree with you about pushing the envelope. It was really interesting to hear that they were planning on testing an engine-out scenario with this booster by lighting one of the middle 10 engines and two centers. If they can pull that off, they'll have a measurable increase in fault tolerance over Falcon 9.
We don't know all the things they did to the Booster, but among them were deliberately not igniting some engines as well as taking a more aggressive angle-of-attack on descent (the rocket is a fairly effective lifting body, as it turns out!).
There may be more things, but between those two I think the latter was a bigger problem. It would have gotten hotter and more physically stressed. And then weakened to the point to where re-igniting the engines caused it to fail.
They also used a new hot-staging maneuver, where the gases were directed out one side so that it flipped more rapidly in the other direction. It was a really fast flip! A rocket the size of a small skyscraper turning 90 degrees in just a few seconds. That could have jarred something loose, too.
Hopefully we find out in the post-mortem. SpaceX doesn't typically give the public as much detail as we'd like, but they're pretty good at sharing the high-level reasons why something failed.
Also really can't understate the importance of Starlink in the development program. Apollo jettisoned parachuting return capsules from the interstages so that engineers could get camera footage from inside the rocket during launch. SpaceX just gets an RTSP stream (or something similar).