I guess it's possible that if the signal clears when the driver is trundling at around 25mph between the 2 TPWS▸ points then he could then start to accelerate, pass the second point just under the trigger speed and continue accelerating.
I'm not suggesting this did happen, just that it seems to be possible given the description and the diagram. TPWS appears to be based on trains that are travelling fast on the approach to a hazard and need slowing in the event of the driver having a lapse in concentration, whereas this one wasn't.
You are quite corect. In this case the TPWS calculations would be based upon a calculated maximum approach speed againt the required permanent speed resritiction ignoring the effect of any signals. I'm not going to speculate further on what or may not have happend but this has always been a known risk. TPWS cannot, and was not designed to, prevent things like this ever happening.