Hi Fp
For Mov(E) the ratio is 2/(N+1) where N is Periods. For example, 0.33 recurring is 1/3 which is equal to 2/(5+1), so the Periods for this EMA is 5. The 0.33 tells us that 33% of the new data array value will be added to 67% of the old data to form the new EMA value. For Wilders() the ratio is more direct, being 1/N, i.e. 1/3 or 3 Periods for the first calculation and 5Periods for the second. Another way of relating Wilders() to Mov(E) is to double the periods and subtract 1. A 10-period Wilders is the same as a 19 period Mov(E) and a 20-period Wilders is the same as a 39-period Mov(E). Notice that an even number of periods in Wilders() cannot be matched by Mov(E) because MetaStock does not allow partial/fractional periods in any function, not any that I can think of anyway. So at least in the case of Mov(E) it's possible to fake half periods by switching to Wilders(). Any time you see PREV code as shown below, where the sum of the multipliers is 1, (with no other elements thrown in) you can be reasonably sure that what you have is an exponential moving average. If the ratio isn't a whole number (e.g. 0.6 recurring to 0.3 recurring) then the expression cannot be transferred exactly to Mov(E), and possibly not to Wilders() either. Hope this helps. Roy _____ From: [email protected] [mailto:[email protected]] Sent: Friday, 11 July 2014 4:59 p.m. To: [email protected] Subject: RE: [EquisMetaStock Group] Swingtrd 1 Modified w/HT Do you have a table of matching moving averages with I guess percentage or fractional value. Like MACD 12,26,9 = To some decimal or fractional value. What is another way of expressing that. I had some table years ago. example: 0.33 * value2 + (0.67 * PREV) = Mov(value2,5,E); 0.2 * value3 + ( 0.8 * PREV) = Mov(value3,9,E)
