| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User
Joined: 12/31/2004
Posts: 163
|
Hi Bruce or Craig,
TeleCharts only has a Simple and an Exponential Moving Averages.
How do I add a Weighted Moving Average?
Bill
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
It depends on the period.
2-Period WMA:
(2 * C + C1) / 3
3-Period WMA:
(3 * C + 2 * C1 + C2) / 6
4-Period WMA:
(4 * C + 3 * C1 + 2 * C2 + C3) / 10
Start by multiplying the current value by the period and subtract 1 from each of the coefficients as you go back. Then divide the result by the sum of the coefficients. If you create it using a Custom Indicator, make sure to select Plot using price scale.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 12/31/2004
Posts: 163
|
O.K. Bruce,
If I want a front weighted moving average, I just reverse the weighting from the current price being the highest weight to the earliest price being the highest weight, right?
Bill
|
|
Worden Trainer
Joined: 10/1/2004
Posts: 18,819
|
No, he gave you the front weighted. Front weighted means the most recent price has more weight.
- Craig
Here to Help!
|
|
Registered User
Joined: 12/31/2004
Posts: 163
|
Oh!
Thanks Craig.
I decided to program a front weighted average and a reverse front weighted average (or is it Non-Front weighted or is it just weighted?), just to see the difference in the plot.
While doing so, I observed something interesting.
This calculation:
(15 * C14 + 14 * C13 + 13 * C12 + 12 * C11 + 11 * C 10 + 10 * C9 + 9 * C8 + 8 * C7 + 7 * C6 + 6 * C5 + 5 * C4 + 4 * C3 + 3 * C2 + 2 * C1 + C)
And this calculation:
(15 * C14 + 14 * C13 + 13 * C12 + 12 * C11 + 11 * C 10 + 10 * C9 + 9 * C8 + 8 * C7 + 7 * C6 + 6 * C5 + 5 * C4 + 4 * C3 + 3 * C2 + 2 * C1 + C) / 120
produce the same plot.
The difference: one is divided by 120 and the other isn't!
Is there an explanation?
I think I also noticed something else happening. I added a day's calculation then looked at the plot, then added another day, observed, etc, etc. The plot didn't seem to change much between 10 and 15 days. Is this just a spurious observation or a valid observation? The plot seemed to change more noticably when I was working on the days below 10.
Bill
|
|
Worden Trainer
Joined: 10/1/2004
Posts: 18,819
|
They will present the same pattern and appear identical because they are both on different scales. Check the "plot using price scale" on both and you will see a HUGE difference.
- Craig
Here to Help!
|
|
Registered User
Joined: 12/31/2004
Posts: 163
|
Stranger than fiction!
|
|
Worden Trainer
Joined: 10/1/2004
Posts: 18,819
|
LOL
- Craig
Here to Help!
|
|
Registered User
Joined: 12/31/2004
Posts: 163
|
O.K. Craig/Bruce,
I settled on a 50 period front weighted average. See if I typed it correctly:
((50 * C) +
(49 * C1) +
(48 * C2) +
(47 * C3) +
(46 * C4) +
(45 * C5) +
(44 * C6) +
(43 * C7) +
(42 * C8) +
(41 * C9) +
(40 * C10) +
(39 * C11) +
(38 * C12) +
(37 * C13) +
(36 * C14) +
(35 * C15) +
(34 * C16) +
(33 * C17) +
(32 * C18) +
(31 * C19) +
(30 * C20) +
(29 * C21) +
(28 * C22) +
(27 * C23) +
(26 * C24) +
(25 * C25) +
(24 * C26) +
(23 * C27) +
(22 * C28) +
(21 * C29) +
(20 * C30) +
(19 * C31) +
(18 * C32) +
(17 * C33) +
(16 * C34) +
(15 * C35) +
(14 * C36) +
(13 * C37) +
(12 * C38) +
(11 * C39) +
(10 * C40) +
(9 * C41) +
(8 * C42) +
(7 * C43) +
(6 * C44) +
(5 * C45) +
(4 * C46) +
(3 * C47) +
(2 * C48) +
(C49)) / 1275
I plotted it using the price scale and it looks real nice. It also compares nicely with the front weighted average in SnapSheets!
Bill
|
|
|
Guest-1 |