User Functions

Most users will never need to write a user function, so you can just skip this section if you want to.
The purpose of user written functions is to allow the extension of a program beyond what the original author supplied.
Your extension may be something from a magazine or an entirely new idea of your own.
Don't be afraid to experiment - you can't hurt anything!
Don't let the idea of a programming language scare you.
SpiffyCharts User Functions are written in a rather simple language that is very similar to many others that you may see in magazines such as Stocks & Commodities. Our manual gives a detailed explanation and a number of examples.

There are a number of user functions supplied with SpiffyCharts. You can copy one and edit it to see what happens.

Simple Example
; The following is an "advanced" User Function. A Simple User Function can be Really simple!
; Anything after a semicolon in a user function is a comment.
; The first few lines of a user function are usually an explanation of the function.
; The Kalman function, originally for missle tracking.
; Some smoothing, very little lag, but results can be wild.
; adapted from Stocks & Commodites, page 163, October 1985
; Xm of magazine is input, Price1
; Kone of article is Data1
; NB smoothing factor must be less than one
; Newcol = Kalman(oldCol, 0.04) is an example of worksheet input.

Require 1 columns, 1 numbers; The first real line of the function - allows us to check for proper input.
Columns; You must have a "Columns"definition line, even if empty. The Output column is always defined by default.
Variables Ktwo Xp Xdel dX Xs; Variables and columns are initially set to zero

; blank lines are OK. The first three lines below set variables.
Ktwo= sqrt(Data1*2.0);
Xp = Price1;    XP is the predicted price, we are setting an initial value
Xdel = 0;    Since the default value is zero we could have left this out!

Procedure; This word Tells SpiffyCharts to loop through your data and do what follows on each day.
Xdel = (Price1 - Xp); Xdel is diff from actual price to predicted price.
Dx=DX + Xdel*Data1;
Xs = Xp + Xdel*Ktwo;
Xp = Xs + 9*DX;
Output = Xp; Output is the value predicted for today

Here's an example chart showing the result of using the Kalman filter.
Kalman stock chart