Digital ACSTrend - page 10
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Hi, I wish to thank you for all the good work you have done by compiling so much and putting all this info for freshers to this field like me ...
Additional tuning
Please Mr. Tools explain a little more how you do your fine tuning.
I was not able to understand very well. In fact I understood something but I am not sure to get it right. Of course if you have time.
And thank you for the help.
I think that now when we have more powerful computers we can use the Richcap libraries in full power.
I will explain what I did. I watched the R-MESA-Cut-off frequencies. I saw that recently we had really very low periods. Look at the bottom line. On the other hand we had also very big periods.
However our adaptive indicator do not take into account them in its standard settings: Its min_period is set at 20, the max_period is set to 120. So we miss information from the both sides.
Look the minimum period is set to 20. However the R-MESA-Cut-off frequencies show us that the minimum period should be 7.
So I did it in the R-FTLM adaptive indicator. The result is completely different.
The same is for the bigger periods R-MESA-Cut-off frequencies identifies a valley at 240 for example. The maximum used period as default is 120. So I changed it to 380 to match those higher periods (on the chart it went to 380). Again the result is completely different. We have a completely different picture.
The Slow filters give us completely different pictures. The adapted "right" is predicting a possible low swing (and telling us that the movement of this morning is kind of a misguiding sucker move LOL from long term perspective of course), the default is going for upswing.
We will see!
Low Phase Space Singularity: Highway to Hell
Here is another example of Low Phase Space Singularity.
We had:
1.High fractal dimension
2.Fractal break - out
3.Persistent movement sequence followed
If we have the conditions and if the market makes strong persistent monowaves:
DO NOT MAKE ANY ANY ANY DISCRETIONARY COUNTER TREND TRADES. BECAUSE IF YOU DO SO YOU WILL FEEL THE HIGH WAY TO HELL. FOLLOW YOUR INSTRUMENTS (objective technical analysis), FORGET ABOUT THE DAILY RANGE, ANYTHING CAN HAPPEN.
Once again the theory explains very well what happens on the market.
So we have cycles in the fractal dimesnion charcateristics. A perdion of high fractal dimension is followed by a perdion of law fractal dimension.
This is a fundamental feature of the market.
The original ASTRend is designed to catch and exploit the market trend.
Here on TSD a Top gun approach has been developped for low time frames. This is achieved
by using a complex set of filters specially designed for different time frames. This
works pretty well under the current market conditions.
The Digital TSD ASCtrend uses different approach.
Still we have two approaches:
1. Trend mode. Well in this case forget about 5 m time frame, go to 1 h time frame.
2. Fractal cycle mode. In this case we can use lower time frames.
We should use directionnal signals only when we have a fractal dimension lower than 1.5 (0.5 iVAR reading).
That means that we have a mathematical probability that the following movement will be in the same direction as the previous one. In those marekt conditions we use our directionnal algorythms according to their design.On the opposite case it makes no sense to use any directional algorythm and different market approach should be used.
The idea is that the Digital TSD should use adaptive algorithms.The adaptive algorythms are not only reading the market language but adapt to it. The first mods use Cicle period of Ehler to adapt digital filters with jurik smoothing.
We had hybrid technology both digital smoothing and cycle period adaptation.
The latest mods use MESA (Maximum Entropy Spectral Analysis) to adapt jurik digital filters.
Again this is another hybrid technology mod.
However those mods use icustom function at the limits of Metatrader you need a powerfull computer. I was kidding to call
them the dual core mods (a quad is even better) it is not far from the truth.
As I am not a programmer I do just concept mods and competent programmers can improve.
As additional filters I proposed some hybrid indicators. As SSAsqueeze which uses SSA of sufficient amount of lag.
For the SSA the REAL TRICK to make it "casual" it is to block the reading for the past and to input every bar one by one.
I used a neural net to predict next price movements but still recalculates. Imagine that you run for the past and you put an
"imaginery" neural net hitting 100 % right it will still recalculate. Moreover in the serious scientific literature this problem is not discussed. That is just a TRICK nothing more. In fact when you will get the non repainting cycles they sometimes will be difference from the reading bar by bar and the reading when you apply the algorithm on the data set.
The other is the Richcap MESA mods (really nice).
The best time frame to trade
Here I will give you a kind of a contrarian opinion about the best time frame to trade.
My ideas is that the best time frame to trade depends on the circumstances.
I have an objective approach about this. The best time frame is the time frame where we have the biggest Hurst exponent (the lowest fractal dimension) and we prefer for signal a lower time frame.
Today on another forum we discussed what is the forecast for the Eur/Usd. We had Elliot waves counts, traditional and according to Glenn Neely, Traditional patterns, Gann analysis, Harmonic patterns etc.
However, you have my opinion. Please take a look at the shot. We had some nice opportunities, the fractal dimension was favorable.
In fact when we use low time frames we have lower risk, moreover the signal can be clearer. Sometimes when we have big bars the signal may be effective but we do not have a signal until the end of the bar.
For example if we have a really long trust bar on a higher time frame. That is not a good place to enter at the open of the next bar, because the market may be ready for a correction.
(There is an indicator from the Corona charts which is names S/N ratio, it is very useful).
There are a lot of waves of exploration, think about tick charts, renko charts, range charts.
Here I will give you a kind of a contrarian opinion about the best time frame to trade.
My ideas is that the best time frame to trade depends on the circumstances.
I have an objective approach about this. The best time frame is the time frame where we have the biggest Hurst exponent (the lowest fractal dimension) and we prefer for signal a lower time frame.
Today on another forum we discussed what is the forecast for the Eur/Usd. We had Elliot waves counts, traditional and according to Glenn Neely, Traditional patterns, Gann analysis, Harmonic patterns etc.
However, you have my opinion. Please take a look at the shot. We had some nice opportunities, the fractal dimension was favorable.
In fact when we use low time frames we have lower risk, moreover the signal can be clearer. Sometimes when we have big bars the signal may be effective but we do not have a signal until the end of the bar.
For example if we have a really long trust bar on a higher time frame. That is not a good place to enter at the open of the next bar, because the market may be ready for a correction.
(There is an indicator from the Corona charts which is names S/N ratio, it is very useful).
There are a lot of waves of exploration, think about tick charts, renko charts, range charts.Hello John,
Thanks for the screenshot and thanks for the asctrend mesa, on my renko charts it's one of my favorite, truly effective.
Sorry to disturb you, but can i ask you what indicators you use ( on the chart window ) and the settings in this screenshot, because i can't reproduce these signals.
Thanks !
Is cycle period changed at these red vertical line?
Thanks John,
R-ftlm-stlm indicator is good. But, put my computer slow down a bit when load or start even my computer is quad core.
But, I confused these red vertical line on chart. Is cycle periods changed at these red line?
Best Reg.
Vertical lines
Well this is Metatrader there and it is a bit different to calculate a MACD and Mesa adaptive digital filter. Metatrader is not Mathlab and has never been intended as a number crunching software. We are using it at its limits really.
Wait a little until it calculates after that it is OK.
The vertical lines show you when the filter is adapted according to the cutoff-frequencies. See my previous post.
This parameter corresponds to filter persistence bars parameter. And by default is set to 100. This is not enough for me, I use every 20 bars.
There is another important parameter "rectify". As the frequencies change in an abrupt way that produces strange movements in the indicator, kind of a bonds from one level to another. To avoid this you can put rectify to true. However take care if you want to pick tops and bottoms with that because it puts some fictive values to obtain smooth curves.
On the previous posts on the chart I used the MESA mod with default settings. For the stops I used my jurik ASctrend 2 mod, I think you have it in the original ASCTrend thread. This mod is useful when the fractal dimension is clearly low and we want a low lag filter.
In the screen shot you mark the iVAR, and the iVAR has nothing to do with the MESA.
However if the iVAR is above 0.5 the time series get too noisy, and that a good premice for a range. Well with enough experience every trader can see when we have a nice trend or nice large range to trade and we have noisy, shitty market where we loose money.
However the edge with the use of iVAR is to see the transition from one state to the other. And that is the point of using it. For example of the reading is below 0.5 and we have a nice trend when the reading gets higher above 0.5 we know that the odds favor ranging patterns. This is especially true when you use together with the daily volatility (not every hour has the same volatility, this is 101 but I am sure there are folks here who do not know it).
Hey, when I set my maximum period to 400 and my minimum to 5 I get completely different readings. The ASCtrebd MESA mod is also completely different from the others. Those indicators are not easy to use especially when you can try to make custom periods based on manual spectrum analysis. Of course you need all the jjma files and include files you have it previously on this thread.
follow up above
I also put PriceSeries.mqh in include foler.
Use 4 digits broker.
For example on FXOpen it works on Alpari it does not.
By the way use more smoothing because the default is set too low. The idea is not to use a cross over of JJMA but its displacement with shift.
For example 18 period and 18 period shifted. You can do manually do so.