Pascal
06-14-2011, 08:39 AM
We received the following question/comment by e-mail.
It is nice to have this sort of questions, because they allow us to expand on a specific subject.
So please, do not hesitate to ask questions publicly.
So, in layman's terms, Pascal's 20DMF system detected institutional accumulation during last Thursday's trading, right? And that, along with oversold conditions (as defined by Pascal), precipitated the change to a BUY signal?
If so, wasn't there any institutional distribution detected using Pascal's system on Friday's nasty selling day?
I'm guessing your answer will be somewhere along the lines of, "Once a 20DMF buy is generated, it would take time--sheer number of days--before a switch to sell would even be possible." Is that the essence of why it didn't switch to a SELL on Friday?
In other words, what's been bugging me is, Didn't the Effective Volume system detect any institutional dumping of stocks last Friday? Or put another way, wasn't enough distribution detected by Effective Volume on Fri. to warrant a switch to NEUTRAL or SELL?
Below is the latest 20DMF figure.
The upper panel is used to find shorting opportunities. The lower panel is used to find buying opportunities.
After a buy signal is issued, there is no provision inside of the 20DMF to cancel it. The only way is that the MF signal of the upper panel moves above either the average or the 0 line. From there, a move either below 0 or below the average would trigger a short signal.
There are however two "fail safe" procedures inside the 20DMF that I have deactivated at this point, as the risk management responsibility is handed over to the robots.
- The first fail safe was a 6% stop loss.
- The second one is a procedure to revert to neutral when the lower panel signal moves above -70 and then below -70 again.
The Robots are much more efficient in doing that sort of risk management work - and they almost did not improve the overal return for the past years.
So, to come back to the selling of last Friday. Yes, the 20DMF detected selling as can be seen in the upper panel.
Yesterday, we can see that the lower panel also reversed down, which is an indication of weakness.
However, even if the 20DMF is sometimes too quick to issue a buy signal, back-tests have shown that too quick is better than too slow.
So, in short, the 20DMF is not a day trading tool, although I sometimes might use it that way. It is a market direction model for position trading. The model is set to put you in quickly and keep you as far as possible in a trend.
Finally, do not forget that both the upper and lower panels of the 20DMF are based on sector based 20D price/volume oscillators. This means that these oscillators have a mathematical tendency to revert back to the 0 line.
Pascal
8793
It is nice to have this sort of questions, because they allow us to expand on a specific subject.
So please, do not hesitate to ask questions publicly.
So, in layman's terms, Pascal's 20DMF system detected institutional accumulation during last Thursday's trading, right? And that, along with oversold conditions (as defined by Pascal), precipitated the change to a BUY signal?
If so, wasn't there any institutional distribution detected using Pascal's system on Friday's nasty selling day?
I'm guessing your answer will be somewhere along the lines of, "Once a 20DMF buy is generated, it would take time--sheer number of days--before a switch to sell would even be possible." Is that the essence of why it didn't switch to a SELL on Friday?
In other words, what's been bugging me is, Didn't the Effective Volume system detect any institutional dumping of stocks last Friday? Or put another way, wasn't enough distribution detected by Effective Volume on Fri. to warrant a switch to NEUTRAL or SELL?
Below is the latest 20DMF figure.
The upper panel is used to find shorting opportunities. The lower panel is used to find buying opportunities.
After a buy signal is issued, there is no provision inside of the 20DMF to cancel it. The only way is that the MF signal of the upper panel moves above either the average or the 0 line. From there, a move either below 0 or below the average would trigger a short signal.
There are however two "fail safe" procedures inside the 20DMF that I have deactivated at this point, as the risk management responsibility is handed over to the robots.
- The first fail safe was a 6% stop loss.
- The second one is a procedure to revert to neutral when the lower panel signal moves above -70 and then below -70 again.
The Robots are much more efficient in doing that sort of risk management work - and they almost did not improve the overal return for the past years.
So, to come back to the selling of last Friday. Yes, the 20DMF detected selling as can be seen in the upper panel.
Yesterday, we can see that the lower panel also reversed down, which is an indication of weakness.
However, even if the 20DMF is sometimes too quick to issue a buy signal, back-tests have shown that too quick is better than too slow.
So, in short, the 20DMF is not a day trading tool, although I sometimes might use it that way. It is a market direction model for position trading. The model is set to put you in quickly and keep you as far as possible in a trend.
Finally, do not forget that both the upper and lower panels of the 20DMF are based on sector based 20D price/volume oscillators. This means that these oscillators have a mathematical tendency to revert back to the 0 line.
Pascal
8793