the DOW ... day by day ...

A while ago I was reading some stuff from TD/Waterhouse where they showed which days of the week were best and how that's changed over the years and ...

>Best? What does that ...?
Wait until I finish!
They looked at how often the DOW increased, on Mondays, over the period 195? to 1989 (can't remember exactly the years) and compared these percentages to some period (I forget which) in the 1990s. Anyway, I thought that was pretty interesting and thought I'd look at each decade to see the percentages of times the DOW increased on Mondays, Tuesdays, Wednesday ...

>Yeah, yeah. I get the idea. So?
So here's what I got:  

For example, during the 1950s, the DOW increased 63.3% of the time - on Mondays.

>Increased? Compared to what?
Monday's closing value compared to the previous day's close.
During the next decade, the 1960s, only 53.7% of the Mondays saw an increase in the DOW.

>Wait! By the 1990s, are you saying that the DOW increased 80.6% of the time?
On Mondays. In fact, Mondays started, in the 50s and 60s, as the worst day of the week (increasing barely more than half the time during the 1960s: 53.7%) but, by the 1990s, became the best day of the week (increasing 80.6% of the time).

>In the 50s Friday's were best, eh?
Yup! Increasing 88.6% of the time ... but not by the 1990s.

>I assume you can use this to decide when to buy and sell, eh?
Wait'll I check.

>Yeah. Very funny. But what about a chart with everything, where ...
Okay. We'll look at the same stuff but each day's percentage is compared to the average for the week.
For example, during the 1950s when the DOW increased on Mondays 63.3% of the time, the average for the week was 76.4%.
By that I mean that the average (for Mondays, Tuesday, etc.) was (1/5)(63.3+70.0+81.4+78.9+88.6) = 76.4%
So Mondays were 63.3% - 76.4% = - 13.2% below the weekly average.
That's shown in the chart, below. It's the first red bar, for Mondays in the 1950s.

Repeating this for each decade (and each weekday) we get the following pretty picture: