Count of Outside and Inside Days by Day, from Russ Sears
For SPY from 2000 on here is the count of "inside" and Outside days, if inside day is defined as High < prior day High and Low > prior day low. Likewise outside day is High > prior day high and Low is < prior day low.
It would appear that the inside days do not like to occur in middle of week. While outside days like to occur in the middle. But Mondays do not have many occurances of outside days.
Why?
Binomial Distribution
Inside Outside Count In Out
Mon. 68 34 542 86.07% 0.03%
Tue. 65 80 590 48.61% 99.05%
Wed. 54 82 593 6.00% 99.46%
Thur. 63 54 581 43.38% 17.42%
Fri. 72 55 577 85.60% 22.89%
Jim Sogi comments:
Very interesting study by Russ. Would imply that its going to be hard to beat Friday's high in the next ten minutes.
My initial thought was that Fridays were more volatile making it harder for Monday to be an outside day, but on counting the vol using SD, its not so.
M 15.74
T 15.64
W 14.86
T 15.32
F 14.27