raughing out roud,,,
We could use htz/msec that could be divided by overall time of the initial event, and we could incorporate different levels of measured decibels for exits. Also we would have to use a time forecaster, I've noticed most initial high velocity breakouts on 1 hour time frames...