As ever. A tool is a tool. Give someone a loaded rifle and they might shoot themselves in the foot. Give the same tool to a different person and he might take out a sparrow's eye from 50 paces.
In other words the tool is only as good as the person using it and there is more than one way to use any tool.
In other words the tool is only as good as the person using it and there is more than one way to use any tool.
In Trading Systems That Work by Stridsman, there are a few chapters on optimising a mechanical MA system with TradeStation. Yes, I reckon you can get profit from it, and if sitting around while your computer trades for you for mediocre profits are for you then fine. So I say YES it's possible to get an edge. The trouble however is that you must optimise the system every now and then ( more backtesting) to find the correct MA's as markets change as we all know.
The advantage of the discretionary trader is that he sees the move well before the crossover, and can also determine whether to take the trade or not (are we trending or not). I believe this is far more profitable.
The advantage of the discretionary trader is that he sees the move well before the crossover, and can also determine whether to take the trade or not (are we trending or not). I believe this is far more profitable.
TheBramble
Legendary member
- Messages
- 8,394
- Likes
- 1,170
Don't use them for X-over, but as a trend tool, thy're quite useful.
Grey1
Senior member
- Messages
- 2,190
- Likes
- 204
The usefulness of any indicator comes from its power to reveal some thing about the market..
If I tell you to go long when the current price exceed previous 5 bars then quite rightly you would ask me what is the significance of 5 previous bars.. My answer then would be Duhhhhhh... Not quite sure ,, or as some people would say " it has worked for me for the last past 50 years " to escape the challenge of finding answers to straight forward questions ...
Now look at this scenario,
Go long if the average of last 5 bars is greater than the average of last 10 bars... Are not going to ask me where on earth am I getting these numbers from ?
If you did ask then I would say Duhhhh..
A great amount of work has been done by Top rocket scientists using FFT techniques to drag an objective and meaningful settings for MA's from cycles within the price action but they have all sort of failed..
IMHO , simple MA cross overs don't reveal anything what so ever about market future's move..
They are kind of a joke. I mean no disrespect for those who still use them ..
If I tell you to go long when the current price exceed previous 5 bars then quite rightly you would ask me what is the significance of 5 previous bars.. My answer then would be Duhhhhhh... Not quite sure ,, or as some people would say " it has worked for me for the last past 50 years " to escape the challenge of finding answers to straight forward questions ...
Now look at this scenario,
Go long if the average of last 5 bars is greater than the average of last 10 bars... Are not going to ask me where on earth am I getting these numbers from ?
If you did ask then I would say Duhhhh..
A great amount of work has been done by Top rocket scientists using FFT techniques to drag an objective and meaningful settings for MA's from cycles within the price action but they have all sort of failed..
IMHO , simple MA cross overs don't reveal anything what so ever about market future's move..
They are kind of a joke. I mean no disrespect for those who still use them ..
In my humble opinion moving averages are of *SOME* use when trading assets with a marked degree of serial correlation. To this end I would conduct such tests on the historic data before considering their use.
Obviously this does not mean that they will necessarily be of use going forward as degrees of serial correlation are notoriously unstable.
My advice would be to disregard them entirely.
As an aside I attended a Technical Analysis seminar done by Updata about 18 months ago and we were shown their optimisation function for the use of moving averages. This struck me as about the most bizaar concept I had come across. A blatant case of curve fitting.
Just because, for example, 47 days is the 'optimal' moving average to have used this has zero predictive power. It is a total fluke.
Attached is a screen shot from a program a friend built for me which shows the dispersal of the % return on EUR/USD using a simple moving average rule to buy and sell (long and short).
Ignore the fact that all returns would have been negative (this is not the point as they may as well have been positive) what matters is the instability. It is a spiky function with no constant distribution over any area.
If any one trade had been very positive for whatever reason this would weigh the results in favour of the moving average that would have predicted that as if it happened as a consequence of that. A dangerous tool!
Obviously this does not mean that they will necessarily be of use going forward as degrees of serial correlation are notoriously unstable.
My advice would be to disregard them entirely.
As an aside I attended a Technical Analysis seminar done by Updata about 18 months ago and we were shown their optimisation function for the use of moving averages. This struck me as about the most bizaar concept I had come across. A blatant case of curve fitting.
Just because, for example, 47 days is the 'optimal' moving average to have used this has zero predictive power. It is a total fluke.
Attached is a screen shot from a program a friend built for me which shows the dispersal of the % return on EUR/USD using a simple moving average rule to buy and sell (long and short).
Ignore the fact that all returns would have been negative (this is not the point as they may as well have been positive) what matters is the instability. It is a spiky function with no constant distribution over any area.
If any one trade had been very positive for whatever reason this would weigh the results in favour of the moving average that would have predicted that as if it happened as a consequence of that. A dangerous tool!
Similar threads
- Article
- Article
