Yes, 90% points lost is also part of the equation, but they're rarely that far apart.
A 90% downside day followed within a week of a 90% upside day, like we got at the March 2003 bottom, has had only one failure that I know of, in late August 1990. It's a rare indicator, but appears to be a very good one.
Wish Paul Desmond had come up with something that identifies tops as well...