| Tanstaafl |
|
Beta Testers, Member, Platinum User, TeleChart
|
| Registered User |
| Jim Dean |
| Duluth, GA |
| Trading (& Atlanta TC Club leader) |
| Science Fiction & SCUBA |
| Male |
|
| Thursday, October 7, 2004 |
| Thursday, September 23, 2010 11:19:33 AM |
799
[0.27% of all post / 0.11 posts per day] |
|
Architect:
Been a while since I posted those formulae ... and I suppose I should have included %b values as well, since they are useful.
Here is the formula for a 2-std-dev, 20-day band's width (note its relationship to the original posting):
2 * SQR((( C*C +C1*C1
+C2*C2 +C3*C3 +C4*C4 +C5*C5 +C6*C6
+C7*C7 +C8*C8 +C9*C9 +C10*C10
+C11*C11 +C12*C12 +C13*C13 +C14*C14
+C15*C15 +C16*C16 +C17*C17 +C18*C18
+C19*C19 ) - ( 20 * AvgC20*AvgC20 ) ) / 19 )
To make that into a percentage of the average price, you divide by the AvgC20 and multiply by 100:
100 / AvgC20 * 2 * SQR((( C*C +C1*C1
+C2*C2 +C3*C3 +C4*C4 +C5*C5 +C6*C6
+C7*C7 +C8*C8 +C9*C9 +C10*C10
+C11*C11 +C12*C12 +C13*C13 +C14*C14
+C15*C15 +C16*C16 +C17*C17 +C18*C18
+C19*C19 ) - ( 20 * AvgC20*AvgC20 ) ) / 19 )
That should do it for you.
Jim Dean
|
I've been busy with other stuff (and other software) for 2-3 years ... but I could not resist the fantastic RT Platinum Suite guaranteed-for-life deal. Will probably be a while before I'm active again. I've stayed generally current during the Loki/SnapSheets/Blocks (now SF), and of course TC is still it's tried-n-true self. But I've got some catching up to do re the greatly-enhanced language (and somewhat revised GUI) that SF has. Hello to everyone else, too! 
|
Nice to see that those ideas are still useful.
Hi, Bruce - long time - I can see you're still holding the fort !
|
My answer may not have been clear.
Use the formula that Craig provided with the Custom Indicator method I described, for ANY timeframe.
Or, use the PCF I supplied AS a PCF (not as a Custom Indicator), for use in EasyScans or a Watchlist Tab column.
If you study my formula carefully, you can see how to modify it for bar-periods other than 5-day "weekly" ones.
Jim Dean
|
Hi, Eternum:
First, thank you Craig for helping out.
The answer to the weekly question is NO. PCF's in TC do not interact with the timeframe chosen. If you create a custom indicator with that formula, and then set the chart to weekly, and plot it in the top pane, the databox will show you the current (and past) values of the ATR20 for whatever timeframe is being plotted.
The PCF for a weekly (ie 5-trading-day) calculation must be regenerated, to determine the Open, Close, High and Low of each week, from which the ATR is then calculated.
Here it is:
.05*( (MaxH5.0-MinL5.0+Abs(MaxH5.0-C5)+Abs(MinL5.0-C5)) +(MaxH5.5-MinL5.5+Abs(MaxH5.5-C10)+Abs(MinL5.5-C10)) +(MaxH5.10-MinL5.10+Abs(MaxH5.10-C15)+Abs(MinL5.10-C15)) +(MaxH5.15-MinL5.15+Abs(MaxH5.15-C20)+Abs(MinL5.15-C20)) +(MaxH5.20-MinL5.20+Abs(MaxH5.20-C25)+Abs(MinL5.20-C25)) +(MaxH5.25-MinL5.25+Abs(MaxH5.25-C30)+Abs(MinL5.25-C30)) +(MaxH5.30-MinL5.30+Abs(MaxH5.30-C35)+Abs(MinL5.30-C35)) +(MaxH5.35-MinL5.35+Abs(MaxH5.35-C40)+Abs(MinL5.35-C40)) +(MaxH5.40-MinL5.40+Abs(MaxH5.40-C45)+Abs(MinL5.40-C45)) +(MaxH5.45-MinL5.45+Abs(MaxH5.45-C50)+Abs(MinL5.45-C50)) +(MaxH5.50-MinL5.50+Abs(MaxH5.50-C55)+Abs(MinL5.50-C55)) +(MaxH5.55-MinL5.55+Abs(MaxH5.55-C60)+Abs(MinL5.55-C60)) +(MaxH5.60-MinL5.60+Abs(MaxH5.60-C65)+Abs(MinL5.60-C65)) +(MaxH5.65-MinL5.65+Abs(MaxH5.65-C70)+Abs(MinL5.65-C70)) +(MaxH5.70-MinL5.70+Abs(MaxH5.70-C75)+Abs(MinL5.70-C75)) +(MaxH5.75-MinL5.75+Abs(MaxH5.75-C80)+Abs(MinL5.75-C80)) +(MaxH5.80-MinL5.80+Abs(MaxH5.80-C85)+Abs(MinL5.80-C85)) +(MaxH5.85-MinL5.85+Abs(MaxH5.85-C90)+Abs(MinL5.85-C90)) +(MaxH5.90-MinL5.90+Abs(MaxH5.90-C95)+Abs(MinL5.90-C95)) +(MaxH5.95-MinL5.95+Abs(MaxH5.95-C100)+Abs(MinL5.95-C100)) ) / 2
Jim Dean
|
Hi, Craig:
Could you please make a PDF available of your notes?
Thanks!
Jim Dean
|
As usual, you did an *excellent* job Craig. It was refreshing to see so much NEW stuff ... and to hear of the plan for many things still to come. The SS platform, and BackScanner in particular, is a great step forward.
It would be nice to get an email or PDF or something from you that has all of the info above, including active links, separate from this forum. Is that possible? If so, please send it to me ... you know my contact address.
Jim Dean
|
Bruce -
Thanks, once again, for maintaining support on these "monster" formulae that I originally posted before SnapSheets was available.
Much of this is very practical to do in SnapSheets. I presume that soon SS will include Scan/Filter watchlist capabilities as well, which would put the icing on the cake.
Even tho SS is still a work in progress, I'd encourage anyone who is seriously considering using these massive ADX/DMI formulae (or other long ones) to move their operations over to SS as soon as possible.
Jim Dean
|
It would have about 180 days of lookback, to have good accuracy (<.01%). I think it would probably be too slow for most people.
SnapSheets would offer a better, faster alternative.
Jim Dean
|
Thanks for your help, Bruce. I'm pretty swamped right now.
Jim
|
|