That's taking a blameless post-mortem approach which is a valid tool, but not the entire answer.
The reality is to cause the issue, the operators had to drive the reactor well into a dangerous and hard to control regime which it would not get into under any normal operation circumstance.
So while yes, it shouldn't have been physically possible to do it, even with that design it took substantial, deliberate malfeasance to get that result (you can also only get that result with that design - a meltdown is not normally an explosion).
From what I’ve read some of the safety systems did not give real-time feedback and the critical error occurred because the operators didn’t know they were in trouble due to lack of real-time feedback. I’m not an engineer and know nothing of these things but the test was not unauthorized and it wasn’t an experiment. As far as I read about the accident.
The reality is to cause the issue, the operators had to drive the reactor well into a dangerous and hard to control regime which it would not get into under any normal operation circumstance.
So while yes, it shouldn't have been physically possible to do it, even with that design it took substantial, deliberate malfeasance to get that result (you can also only get that result with that design - a meltdown is not normally an explosion).