|
PureBytes Links
Trading Reference Links
|
To clarify the previous posts and hopefully avoid confusion among people who don't understand the chat about the MS-11 Adaptive Indicators and the zag functions, selecting the volatility option (#2) eliminates all use of the zag function so the indicators are usable as indicators that change with volatility.
Selecting the volatility option means the adaptiveness of the indicator is based on volatility only and not cycle count.
I've never found cycle count to be worth much so I don't bother with it. There are indicator packages available from Equis that use DLLs without the zag function to create adaptive indicators for the John Ehlers indicators. If someone is interested in cycle counts, Sine Waves and adaptive indicators, the Adaptive Cycle Toolkit does that well.
While I've read, programmed and tested everything Ehlers has developed over the years, I personally don't find it provides me much of any benefit so I don't use anything from Ehlers. He is particularly popular with the engineering/traders, so there are some very loyal devotees who live and die on his work.
If anyone wants to use cycles as a method of indicator adaptive function, then you may want to consider basing the code on the Hilbert cycle formulas which don't need a zag function. I've seen that done previously but I have not tested anything using Hilbert so I don't know how well they work.
One small point about the zig zag functions. If they are applied to stock prices as a way to find tops and bottoms, the dynamic last leg creates phantom signals that no one has found away around--at least not yet. So I agree with what other people have said, as long as they are talking about applying the zig zag functions to price data.
However, there are legitimate programming uses for the zig and zag functions as well as peak and trough functions. When these are applied to things other than the price function, the degree to which the dynamic last leg interferes with or causes problems in that equation differ considerably from equation to equation. That means that sometimes the dynamic nature of the functions fail to influence the outcome enough to make the use of them problematic. Each situation has to be taken on a case by case basis and the extent of the problem quantified for that specific use. There may be cases where the dynamic nature of peak and trough, for example, only create a misleading signal one time out of every 1000 signals. It may turn out that the misleading signal creates almost no change in the results.
I've written articles on testing previously where I have stated that test results are never accurate. At best they are only relative. In other words, almost every systems test has problems related to survivorship bias, data mining, improper application of statistics, over optimization and a host of other things. Proper testing is one of the hardest and most expensive things to do, and even the experienced testers have trouble figuring out how much some of the problems I mentioned have influenced their test results.
In general, when I run test results I know they are not accurate and will never approximate future trading. However, I also know that I can compare them on a relative basis. So if one test has triple the return with one third of the drawdown of another test, and I used exactly the same testing procedures and data sets in each test, then I can count on being better off using the strategy from test 2 rather than test 1. If I'm expecting my live trading results to duplicate or even parallel the tests, then I'm going to be in for a shock, so personally I don't bother with those expectations.
I test everything I do extensively so I can see how it has changed over time, what kind of market conditions it works best in and what kind of volatility and variance in the results I should be looking out for.
The bottom line based on what I've seen so far is the MS 11 volatility based adaptive indicators work as any indicator based on volatility would be expected to work. The MS 11 adaptive indicators based on cycle count have many problems, the least of which appears to be the zig function, so stay away from those.
Super
------------------------------------
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/equismetastock/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/equismetastock/join
(Yahoo! ID required)
<*> To change settings via email:
equismetastock-digest@xxxxxxxxxxxxxxx
equismetastock-fullfeatured@xxxxxxxxxxxxxxx
<*> To unsubscribe from this group, send an email to:
equismetastock-unsubscribe@xxxxxxxxxxxxxxx
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|