Strategic foresight systems - page 10

 
alsu:
Read the top of the home page: "mechanical trading systems forum". He just feels unwanted here, so he shits in the technicians' threads...
I'm preparing the material (will take some time), there are questions and some misunderstanding. About 2 years ago I tried to find a coherent pattern with macroeconomics - nothing worked. And the idea of identifying a model based on external data remained. So I want to try to answer this question ... once again.
 
strangerr:

Yes, and after that a correction or trend change is possible.
seems to be going up so far... (although gambling is a bad quality)
 
In general, it would be interesting to draw up a table of some sort. In the left-hand column are the different ways of analysis. Next, for each, set a selection criterion. Then for each selection criterion set a criterion for defining parameters. For example: Select autoregressive model because and because, select model order 3 because and because, select amount of data for the model because and because, determine the forecast range because and because, determine the forecast confidence limits because and because, make the forecast, or: select moving average model because and because, etc. And have such a tree at least for the most common models. However, I don't think such a task is feasible, because it takes a lot of time, education, research and people, and it's hard to do it alone.
 
Farnsworth:
Yesterday it (growth) was over in a couple of hours. By the way, it went up a lot, just in the morning. Besides, I predict Open D1 - from the point of view of regulations it's convenient, but from the point of view of determining SL and TP it's a problem fraught with huge drawdowns. I'll only be able to recalculate my probabilities in the evening.


Hello!

I understood you. My bad. For me the completion of the move is a complete set of internal structure. On the M5 chart (left upper corner) it was said about the sinking, that had to happen at 5-6 PM MSK, and it happened, but not by a hundred pips, as I estimated, but by a little over 40 pips. After 6HMSC it went up (that was expected) - this is the H1 chart - red dashed line in the black square. I had to put up a new chart at 6 o'clock (see the red dotted line in the black square). I should have put a new chart and show my understanding, then the previous charts would have been clearer, but I was sleeping ...))

 
-Aleksey-:
In general, it would be interesting to draw up a table of some sort. In the left-hand column are the different ways of analysis. Next, for each, set a selection criterion. Then for each selection criterion set a criterion for defining parameters. For example: Select autoregressive model because and because, select model order 3 because and because, select amount of data for the model because and because, determine the forecast range because and because, determine the forecast confidence limits because and because, make the forecast, or: select moving average model because and because, etc. And have such a tree at least for the most common models. However, I don't think such a task is feasible, because it takes a lot of time, education, research and people, and it's hard to do it alone.
I am going to conduct such quality control for my system, but for other systems it is probably unlikely, you are right there.
 
Zet:


Hello!

I understand you. My bad. For me the completion of the move is a complete set of internal structure. The M5 chart (left upper corner) showed the decline that was to happen at 5-6 PM MSK, and it did happen, but not by a hundred pips, as I estimated, but by a little over 40 pips. After 6HMSC the pair went up (that was expected) - this is the H1 chart, red dotted line in the black square. I had to put up a new chart at 6 o'clock (see the red dotted line in the black square). I had to display a new chart and show my understanding, then the previous charts would have been clearer, but I was asleep ...))

Exactly, it's all about scale. Could you make a forecast on a higher scale, so that the minimum scale would be a day? Is this possible or do you have a specific methodology?
 
Farnsworth:
Exactly, it's all about scale. Could you make a forecast on a higher scale, so that the minimum scale would be a day? Is it possible or do you have a specific methodology?

Yes, but it's not interesting...)) There are plenty of waveforms. Everyone from the masters of VA to.... etc.
 
Zet:

It's easy, but it's not interesting...)) Lots of wave markings. Everyone from the masters of VA to.... etc.
don't really get it, the days are full of markings and it's not interesting, but there's a shortage of them on the clock? We'd have to match the scale. How about retraining for time? :о)
 
Farnsworth:
preparing material (will take some time), there are questions and some misunderstandings. About 2 years ago I tried to find clear patterns with macroeconomics - nothing worked. And the idea of identifying a model based on external data remained. So I want to try to answer this question ... once again.
If you need help, you can discuss in private.
 
alsu:
If you need help, you can discuss in private.
OK. I just need to formulate a problem statement. And that's not so easy.