Hindsight is 20:20 of course, backseat driver etc. but if the on-board systems detected a bad rotation, then started a burn to correct it, presumably it could have been possible to detect that the burn was not "helping" during the burn? And halted the burn? The thrusters aren't usually that powerful, so the erroneous death-spin probably took a fair while to spin up. Even if the sensors were wrong, if you're a computer trying to get variable 'X' into a range, and you apply control 'Y', but 'X' moves further and further away, let go of the controls and ask ask an adult help! (Easier said than done, I know. I'm amazed space engineering works as often as it does! Super hard stuff.)
Of course it was using a feedback loop, but GIGO applies:
The STT and IRU disagreed on the attitude of the satellite.
In this case the IRU takes priority, but its data
apparently was wrong, reporting a rotation rate of 20
degrees per hour, which was not occurring.
Starting from there it would have no way of knowing about the true value of 'X', so the feedback loop was fed with wrong data and just kept taking decisions† that made things worse, especially given that:
The satellite configuration information uploaded earlier
was wrong and the reaction wheels made the spin worse.
[...]
the ACS attempted to correct a rotation that didn’t exist.
The erroneous configuration information led the ACS to
aggravate, not correct, the rotation.
† Even without misconfiguration, stopping an object from spinning in a vacuum isn't as direct and linear as accelerating/braking in a car, requiring precise coordination of multiple fixed thrusters and/or reaction wheels.
