Testing real-time forecasting systems - page 39
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
Can I recalculate the forecast!? As long as the levels are worked out.
I can't now. For technical reasons. :о(
Exploring new horizons:
Not very practical yet, but pleasing to the eye :-)
The blue ones are possible implementations of the process, and the bold line is kind of average?
The blue ones are possible implementations of the process, and the bold line is sort of a middle ground?
Accidentally deleted the post :-)
It's like you said :-) Only one average won't be enough, I need several, but how many and how to calculate them I haven't thought of yet :-(
Accidentally deleted the post :-)
It's like you said :-) >> Only one middle won't be enough, I need several, but how many and how to calculate it I have not invented yet :-(
There is a strong suspicion that the average will always be near-horizontal. And what does "Just one average won't be enough, you need several" mean? What do you mean by several averages? But there's always one average, isn't there an average, or am I missing something?
Yes, it is, the bigger the sample, the more the forecast approaches the horizontal line. I think we could somehow normalise the curve to the average volatility, it would be more realistic.
Yes, a few averages, of course they will no longer be averages, but lines drawn through the areas of maximum process concentration.
Accidentally deleted the post :-)
All as you said :-) Only one average won't be enough, I need several, but how many and how to calculate them I have not thought of yet :-(
I think in this case we need to pay more attention to the kinks in the "middle" line, how much they are executed...
Yes, it is, the larger the sample, the more the forecast approaches the horizontal line. I think we can somehow normalise the curve to the average volatility, it will become more realistic.
Yes, a few averages, of course they will no longer be averages, but lines drawn through the areas of maximum process concentration.
Here https://forum.mql4.com/ru/17052 I described a very simple approach to refining pivot points for my ZZ. Try to keep only local extrema from the trajectories and correlate them with the occurrence frequencies of historical price levels, something like that (the link describes it in more detail):
Maybe we'll get something. At least there will be some selection criteria. In addition, take a look at the statistics on the completion of one wave of a ZZ. Decide on the parameters of a zone at some point.
I don't understand why the X-axis (which in the source material seems to correspond to the duration of the segment in time) is an inverted density? IMHO, there should be a blurred peak too.
that is the time.