opnion, that is a strange figure. I checked my data stash, and except for the day you mentioned -- 7/29/02 -- when NYSE advancers minus decliners hit 2086, there has not been another day of +2000 or greater at least as far back as 5/29/84 (which is when my data begins).
On the other hand, taken by itself, the fact of an extreme in NYSE advancers minus decliners does not seem to make a huge difference over a one to four week period.
Here is a graphic of my (very quick) analysis of this stuff. The graph shows the percentage change in INDU at one, two, three, and four weeks after each of the forty largest extremes in advancers minus decliners since 05/29/84.
Market Data 
Markets 
AI
