One of the recurrent themes to majority of these events is the crew fails to hear the final bug backstop height at 100 feet. I am wondering if we have this backstop too low in that the state of arousal by then is so high it is simply not processed. Just a thought.
I've been saying that for years, and certainly in the period after ETAP. The whole reason for descending through 100ft is almost certainly a loss of situational awareness and a probable saturation of the crews. To expect them to hear the alert, process what it means, snap out of their lost S/A and then respond - and the aircraft respond to their response - seems ludicrous. At 1000 ft/min that's 6 seconds! Hearing also becomes ineffective when the brain is saturated, so it may not even be heard anyway. The mandating of AVAD with a fixed setting came about in response to the BA S-61 accident in 1983, but that was an inadvertent cruise descent (as was the 2004 ERA S76, at night from 2,000ft) and so the AVAD alert would have given a reasonable notice. If we look at all the accidents I've listed previously the time frame would have been absolutely minimal and of no practical use. It's one of the reasons research is being done into modifying the Mode-1 EGPWS algorithms to make them useful.