No registration required! (Why?)

# Weekly pivot influence on first trading day

quote:
Originally posted by BruceM

...I have a friend who will only trade in the direction of the weekly pivot on the first trading day of the week....he uses it like a magnet...so for today he is only taking longs......his is by observation

I'm curious about testing the general theory of this. I'm typing as I think...

If we took the week's first trading day's opening price and measured the number of points that it traded in the direction of the weekly pivot point and also the number of points it traded away... We might be able to measure a statistical deviance in the direction that it trades. Problems with this theory are (1) the market has an upward bias in the long term (i.e. not symmetrical (2) Market Profile theory can replace this because we are probably trading towards the weekly Point of Control which is likely to be near the weekly Pivot Point.

Testing is difficult because we need to compute the first trading day of the week in case it is a holiday and also count the number of trading days in the prior week in case of holidays. For example, last week we had 3 trading days in the E-mini S&P500 because of Gerald Ford's day of mourning. This was an unexpected holiday that is not on the calendars.

What I am thinking of is running a test to see if a rolling 5-trading-day pivot point has any influence on the direction of trade for any particular day-of-the-week. For example, are Monday's more likely to be days where the price is "drawn to that pivot like a magnet" rather than any other day of the week. That should be easier to test...
<<edit start>> Data being used to perform these calculations: E-mini S&P500 daily data using RTH figures from 1/15/2002 to 1/8/2007 <<edit end>>

Okay, I've just run an initial analysis. Based on a 5 day rolling pivot point. These are the number of times that the pivot point has been traded through during the subsequent trading day.
```
Mon     Tue     Wed     Thu     Fri
94	122	122	116	120  (times traded through)
234	256	257	251	251  (total days in data)

40%	48%	47%	46%	48% (percent traded through)```

In other words, 40% of the time the pivot point will be at least touched on a Monday and the most likely day that the pivot point is touched is on Friday (47.8%) (Tuesday is 47.7%)
Here is another question that I asked myself when looking at this data. Can the 5 day rolling pivot point be used to predict if the market will close above or below the open? In other words, if the 5 day rolling pivot point is above the open then will that bias the close to be above the open as well?

Here are the results:
```
1249 (days observed)

647 (close on same side as pivot point)

51.8% (percent)```
I don't believe that 51.8% is significant in this case so my conclusion would be that this is not an edge.
And a final experiment for now is to measure the number of ES points that you would have made by opening a position at the open (at 09:30 EST) and closing it at the closing price at 16:15 EST and trading in the direction of the pivot point.

This resulted in a profit of 486 ES points over 1249 trading days averaging 0.39 points per day before commissions. If you factor in 0.1 points commission for each trade taken (i.e. \$5 per contract traded) you have a net average return of 0.29 points per trading day.

Plenty of assumptions have been made here. For example, stops have not been used or discussed, only the target and entry/exit methodology. Also it assumes that you can enter and exit at the opening and closing prices of the day.
DT, Could you explain "Rolling" pivot when you have time? I've been watching the weekly pivot point since this thread began and I think the market has traded back to it each week at some point. We hit it today. Would this be an easier test to make? Can you run a test to see how often the weekly pivot is tested at some point during the week? Perhaps this is what the rolling pivot is but my simple mind doesn't understand. thanks

Bruce
A rolling pivot is a pivot calculated off the previous N trading days. So the pivots that you would use for today would be those calculated off the values generated from the last N trading days. If N is 5 then it is from last Friday and the 4 trading days that we have seen this week.

The problem with calculating weekly pivots is that you need to run it through a program that can do the logical date calculations (or a script). I was generating these results with Excel without using any scripting (because it's fast to do) and so I can't really test anything beyond this without expending a lot of time on it to write the scripts and code to do this.
Ok DT, Thanks for the explaination. This would be cool information to know someday...for now I'll just keep tracking it and will post the first week it fails.

Bruce
Weekly Pivot hit today on Tuesday after being down 10 points below it on Monday...

Bruce
Are you using the weekly pivot off the daily notes page or a rolling 5 day pivot for this? i.e. 1445.92 for this week...
Off your daily notes page only.
quote: