Discussion of article "Regression Analysis of the Influence of Macroeconomic Data on Currency Prices Fluctuation" - page 3
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
The idea is clear, but the model is not formulated correctly.
To be more precise, it is not correct.
When you have such a model and there is news on one of the economic indicators included in the model, your task is to enter a new value of the indicator into the model and get an increment of the asset price y(0)=(p(0)-p(-1)). That is, in fact, the model finds today's price p(0) at the known yesterday's price p(-1).
But you cannot get the increment you are looking for because the two independent factors in the model are increments x(5) and x(10), which also include p(0).
You are actually forecasting today's price based on today's price.
...
You're actually predicting today's price based on today's price.
Not even today's price, but the future price! And he's happy to see it come true. Good man.
Not even today, but in the future! And he's happy that the forecast's coming true. Good man.
I do not insist that this method is manna from heaven, I just offer you an additional tool that can make your work easier and save you a lot of time.
Or on the contrary, it will be a waste of time.
The trouble with regression models is that they give acceptable results only on information with inertia. That is, when trends have a high probability of continuation.
But financial markets, unfortunately, are not always inclined to continuation of tendencies, which is eloquently evidenced by crises that occur in spite of all models built with the help of statistics.
However, other methods do not allow to predict economic crises with acceptable accuracy, because they are rare events and it is difficult to generalise them. Each crisis is unique.
Earthquakes are also unique because the tectonic plates are arranged differently in different places. After the tectonic plate shift that caused the earthquake has occurred, the underground structure has changed once and for all, and irreversibly to its previous state. Next time, if there is a shift, it will most likely be in a different place.
Approximately the same thing happens in the economy. Each crisis radically changes the economy, because all those who adhered to the previous models were bankrupt. And those who stayed afloat were just more cautious or just lucky. Therefore, it is useless to build previous crises into the models, because the probability of repeating their previous signs is extremely low. Where it was thin before, it has already torn. Next time it will tear in another thin place. Therefore, it is better to look for thin places rather than hoping for a recurrence of previous trends. But unfortunately, regression analysis will not help here due to its conservativeness with respect to previous trends.
An Expert Advisor was created on the basis of the article. Trading on gold. You can see the results of its work here: ...
An Expert Advisor was created on the basis of the article. Trading on gold. You can see the results of its work here: https://www.mql5.com/en/signals/61321.
Profit results are nothing. You should look at the ratio of losing and profitable trades, and it is approximately equal! This is pure haggle. The model does not give statistical advantages.
PS.
The profit shown by you is either the result of stronger movements, which fell on profitable trades, or on the efficiency of the money management you use.
Profit results are nothing. You should look at the ratio of losing and profitable trades, and it is approximately equal!
Oh, don't talk about things that you are a nobleman in
Oh, please don't talk about things you're a nerd about.
Can you be more specific, as a person who boom-booms?
simple correlation of the number of profitable trades with the number of losing trades is meaningless, as it is not flipping a coin with 2 variants of the result.