that's a fairly benine thing, 20%, but it's the shape and look of the ascent that precedes the drop that'sth that, let's look at epic tops and see how they also are quite correlated so let's go back to 1937. of course, we had the '29 crash and spent the most of the '30s coming off the lows but then we encountered trouble in '37. frankly in february, and we plunged basically 50% through to the end of the year. what's important, again, is the correlation with the current two years that we're in right now. it's about 90 plus%. take a look at a 68. in 1968, a very similar two-year trajectory just like the one we're in now and then the trouble started. in this case it was december of '68 and over the next 18 months, a little bit longer we dropped about 40%. now, let's lock at a 87, and it's -- it's the same thing. in fact, of those four or five periods that i've cited, the one that's the most correlated, running at a 96% correlation is this one, and if you just pokes the current two-year s&p against the two years preceding the crash in 5087, this is the overlay, and so does it have to play out