Testing real-time forecasting systems - page 43
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
2 Helex. The calculations and examples are all unfortunate (out of place). As an illustration I will tell such a story. I remember an interesting problem we were given at the institute. So a young man has two grandmothers, one of whom lives at one end of the metro line and the other at the other. He himself lives somewhere in the middle. Every morning he goes down to the metro and gets on the first train that comes up. Question - which grandmother does he go to more often? Wrong answer: 50-50. Correct answer: It depends on the timetable - if on the timetable the trains in one direction arrive with a greater/lower delay relative to the other, there will be a skew in the respective direction.
In the context of probability work in markets - it is also important to know the "schedule", i.e. to find such random processes whose probabilities are not 50/50 (as in the case of apples in the basement), but are skewed. With apples, too, there are 'patterns' - some varieties are more prone to worming than others, and knowing this we can only select apples that are less wormy de facto.
2 Helex. The calculations and examples are all unfortunate (out of place). As an illustration I will tell such a story. I remember we had an interesting problem at the institute on tervers. So a young man has two grandmothers, one of whom lives at one end of the metro line and the other at the other. He himself lives somewhere in the middle. Every morning he goes down to the metro and gets on the first train that comes up. Question - which grandmother does he go to more often? Wrong answer: 50-50. Correct answer: It depends on the timetable - if on the timetable the trains in one direction arrive with a greater/lower delay relative to the other, there will be a skew in the respective direction.
In the context of probability work in markets - it is also important to know the "schedule", i.e. to find such random processes whose probabilities are not 50/50 (as in the case of apples in the basement), but are skewed. With apples, too, there are "patterns" - some varieties are more prone to worming than others, and knowing this we can only select apples that are less wormy de facto.
That's what DC is all about - that bias just not in our direction. I am just saying that there are so many conditions of price movements that we cannot rely on the probability theory and statistics, or at least build the base of the system on them. This is utopia. If you do, I will be glad for you.
The DC is only one factor - a local one. There are plenty of more global factors than a single DC. But I see no point in arguing - we have pluralism.
But I want to know what to rely on, please enlighten us? Or profitable trading is impossible? That would be like the statement of a casino player, who got on this site because of the recent closure of the casino. ;-)
The DC is only one factor - a local one. There are plenty of more global factors than a single DC. But I see no point in arguing - we have pluralism.
But I want to know what to rely on, please enlighten us? Or profitable trading is impossible? That would be similar to the statement of a casino player, who got on this site because of the recent closure of the casino. ;-)
I can tell you that a long term MM is very difficult to write, I believe the foundation of a good system is money management. That's it in a nutshell.
Sergey. If you don't look that far ahead (the whole trading week), what has changed when you added the last two days (Monday, Tuesday), because prices have moved a figure lower than you estimated. The process itself is interesting. How far the trajectory will go with the new data.
These are statistics + self-organising stochastic systems, and in this system - "all in the mean". If you look closely, the "extra" has been added by a single bar, clearly outstanding in size. It is very difficult to predict such a strong movement on one bar, but experience teaches you not to limit yourself in your cognitive capabilities :o) As I wrote earlier the result of the forecast is a number of probable trajectories. It's all modeled by systems with random structure (i.e. according to some probability laws the movement from one model to another takes place) and I'm showing far from all results of forecast. So it's not that simple.
What's different after adding the last two days (Monday and Tuesday)? Prices went one week lower than you calculated.
There is one subtlety here, the calculation takes 1 to 5 hours. And this is on a 2 core computer with 4g of RAM. I can't go through it all, though I would like to.
Prove it mathematically, if you're so smart, thanks.
xDDD
Well yes, the esteemed Helex is proving the uselessness of probability theory with probability theory and apples. It seems more interesting to me to estimate the pressure in Daughter Cili's stomach after eating so many apples.
And this:
I'm not saying I'm so smart, I'm saying you're playing with casino odds no more
It's a "I'm not a reader, I'm a writer" kind of thing.
A slight refinement of the prediction
in general - testing
A small refinement of the forecast
in general - testing
Sergei, a short one is open and kind of a target is indicated at the bottom.
Is the stop assumed or is it at the top outside the chart?
A small refinement of the forecast
IMHO, it is desirable to give a time reference. Apparently the refined forecast is not from the beginning of the week (especially since for ~2 days), but at the initial price it looks as if it starts in the middle of the 29th, which is certainly possible, but unexpected. ;-)