Market prediction based on macroeconomic indicators - 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
There is economic data in these minutes, I understand you can get it. If not, we need to assess these meetings in three ways - +1/-1/0 - the information for the assessment can be taken from the media as an option.
Try it. No economic data new to the market there
As I recall the Soros attack on the pound a long time ago, the market reacted to the frequency of the head of the English central bank blinking during his speech
Try it. No economic data new to the market there
As I recall, during the Soros attack on the pound a long time ago, the market reacted to the frequency of the head of the English central bank blinking during his speech
There are some questions about the problem statement:
Well, the answer to the question about stationarity (imho): the number series is divided into two symmetric hemispheres +- (with mirror symmetry at zero), which in turn have two symmetric parts with central symmetry at points 1 and -1, respectively. As long as the studied series does not leave the boundaries [-1 : 1 ], the series is stationary, as soon as the value leaves these boundaries we cannot answer precisely to which region the value belongs, and whether it is merely a mapping of some point of the stationary series. Except that the transformation (if we just bring the series |x|>1 to stationary via 1/x) is not transitive, which is very bad, because once we get a prediction we cannot use it.
There are some questions about the problem statement:
1. it does. Not to all, but there is.
2. Why? The vast majority of macroeconomic indicators are forecasted and quite successfully by experienced analysts and they put their forecasts out to the public a week or two in advance.
3. the market reacts to both the forecast and possible deviation
1. there is. Not for everything, but there is.
2. Why? The vast majority of macroeconomic indicators are forecasted and quite successfully by experienced analysts and put their forecasts out to the public a week or two in advance.
3. the market reacts to both the forecast and possible deviation
Great, then what is the model that takes into account the reaction of the market to both the forecast and the forecast error in the output?
It is obvious (imho) that these are different models, in the first case we deal with data from stationary series, in the second case we deal with data from symmetric region to stationary series.
Great, then what is the model that takes into account the market response to both the forecast and the forecast error on exit?
It is obvious (imho) that they are different models, in the first case we deal with data from a stationary series, in the second case we deal with data from a symmetric region of a stationary series.
Ideally, the independent variable(macroeconomic indicator) will change twice within the same time period - at the moment of the forecast and at the moment of the indicator output. Or once, if the forecast turns out to be correct.
Then the one quarter step as in this model is impossible. Only one day step
Ideally, the independent variable(macroeconomic indicator) will change twice in the same time period - when the forecast is released and when the indicator itself is released. Or once, if the forecast turns out to be correct.
Then the one quarter step as in this model is impossible. Only one day step.
I see some misunderstanding, let me clarify:
We do not know which forecast the market has considered, because this forecast is a result of thousands of traders' and analysts' thoughts, although of course we can take some forecasts as a reference, but in fact the first part of the Marmazon ballet consists in calculating the forecast the market has reacted to while waiting for news to be released.
Having the forecast indicator, we can already calculate the market reaction to the forecast error upon release. We have g(f(x)-n)-(x+1), where f is the function transforming the quotes into forecast indicator, n is the value of indicator at the exit, g is the inverse transformation of indicator into quotes, x are the quotes before the news release, x+1 are the quotes after the news release.
I see some misunderstanding, let me explain:
We do not know which forecast the market has considered, because this forecast is the result of thousands of traders' and analysts' thoughts, though of course we can take some forecasts as a reference, but in fact the first part of the Marmazon ballet is to calculate the forecast the market has reacted to while waiting for the news to be released.
Having the forecast indicator, we can already calculate the market reaction to the forecast error upon release. We have g(f(x)-n)-(x+1), where f is the function transforming the indicator quotes into forecast, n is the indicator value at the exit, g is the inverse transformation of the indicator into quotes, x are the quotes before news release, x+1 are the quotes after news release.
The publicly available indicator forecasts are more or less adequate to the general market perceptions.
Having these indicators, we CAN calculate the market reaction, but NOT EVERYWHERE and NOT accurately:
1. the correlation coefficient for macro-indicator quotes changes over time and very dramatically, up to the sign.
2. the problem is the presence of unformalised news - rumours, verbal information, political information, etc.
3. several indicators can be released per day + p.1 + p.2 = non-stationarity.