Trading News with economics & probability | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Successful Trading & Determining the Probability of a Trading System 10 replies

ECN && STP && Scalping && Hedging - Broker Review - Oct 2009 24 replies

Probability, Expected Value, & High-Risk Trading 7 replies

Risk Reward +Probability [Poll & discussion] 2 replies

MM, RR Ratio, & Probability of Wins 0 replies

Trading News with economics & probability

  • #1
  • First Post: Sep 24, 2015 7:34pm Sep 24, 2015 7:34pm
Hey guys i start this thread to promote a trading system that is mostly based on the fundamentals aspect of the countries , in order to create this we need to analize all current indicators of the pairs we want to trade.

We should organize a data panel using excel or whatever to see all the previous information in order to make present and ahead decition. This panel should contain the following indicators:

-GDP
-CPI
-PMI
-Interest Rate
-Unemployment Rate
-Building Permits
-Trade Balance
&more (in general all high and medium impact news if we want to use probability)

If you think we should use a statistical model to put probabilitys on the trades comment below. I personally think we can use the Gauss bell or standard deviation alone.

ˇWelcome all !
  • #2
  • Sep 25, 2015 4:14am Sep 25, 2015 4:14am
You might find it difficult to integrate information from a data panel into a usable trading system - I used to plot my own data panels, then I started plotting little line charts of each indicator so I could see which way everything was trending, now I just use the line charts from TradingEconomics .com - it's more efficient but maybe you'll learn the material better by going through the legwork manually.

Some indicators are more important than others, and different indicators carry different weights in different economies. GDP, Unemployment and Inflation are the big three - when read together they give a good indication of the direction of the economy, and give big hints about which way future events might go. Business & consumer confidence, PPI, orders etc are genuine leading indicators, (unlike technical 'leading' indicators) - also think about how indicators affect each other - GDP growth creates jobs, more people in work means more spending which pushes GDP higher creating even more jobs, etc etc... there are many such reciprocal relationships to discover..... when you start learning how different parts of the economy drive each other, the word 'trend' kind of takes on new meaning.... there are strong causal links there that don't exist in any other kind of market analysis .....

It's hard to build an entire trading system which is entirely fundamental, because most of the time when you get the direction right the timing is almost impossible. Best to use fundamentals to find which direction you want to trade each currency, then use a simple technical system which unfolds over the same time horizon as your fundamental bias.

good luck
si hoc legere scis nimium eruditionis habes
  • #3
  • Sep 25, 2015 1:12pm Sep 25, 2015 1:12pm
I'm trying to see the link between fundamentals aspect and real life market by making an exel file which is contain previous data from CDP,CPI,PMI,interest Rate,Trade Balance and Unemployment Rate.
And I saw nothing.
market goes against my analyze 50% of time ---> random ?!?
  • #4
  • Sep 25, 2015 3:41pm Sep 25, 2015 3:41pm
Quoting ForexSpider.
Hey guys i start this thread to promote a trading system that is mostly based on the fundamentals aspect of the countries , in order to create this we need to analize all current indicators of the pairs we want to trade. We should organize a data panel using excel or whatever to see all the previous information in order to make present and ahead decition. This panel should contain the following indicators: -GDP -CPI -PMI -Interest Rate -Unemployment Rate -Building Permits -Trade Balance &more (in general all high and medium impact news if we want...
i worked on something kind of similar few months ago.

it was actually a neural network that i made in matlab, used every previous high impact economic data from X country ( i did the Euro zone and Canada)
i ended up with +20 input data over 10 years period ( i could've gotten a bigger set of 20 or 30 year, but it was already huge)
the output was one of the economical release ( i mostly focused on GDP, PMI, retail sales ) .

altho i got the direction right 95% of the time, i was off by 0.1 or 0.2 more than 50% of the time.

the thing is 0.1 is already a too big margin.
what makes the market move is the deviation from the forecast (which already gives the direction most of the time)

but i have to admit, my data may have been a bit corrupted with wrong inputs, i had to enter all the values manually, so its likely.
this doesn't change the conclusion tho.

i hope you guys will have a better luck
i lose money for a living
  • #5
  • Sep 28, 2015 9:17am Sep 28, 2015 9:17am
Quoting numbnuts
You might find it difficult to integrate information from a data panel into a usable trading system - I used to plot my own data panels, then I started plotting little line charts of each indicator so I could see which way everything was trending, now I just use the line charts from TradingEconomics .com - it's more efficient but maybe you'll learn the material better by going through the legwork manually. Some indicators are more important than others, and different indicators carry different weights in different economies. GDP, Unemployment and...
Hey numbnuts ,

Youre right is more efficient to look the trend in tradingeconomics , but as i dont have a paid subscription there im not able to see the numbers which is my mainly focus , no because im trying to find the holy grail in trading; only because i believe it could trow more accurate data to analize. I use the FED database and others to get precise information.

But the intregrate panel is for actually do what you already see ( except i still only see it when i have the numbers) , is the whole connection between the leading indicators and the trend generation between rational expectations ( i still dont know how to quantify this though). What you said about technical seems very logical because it could gives the entry / exit signals

I apreciate your knowledge sharing and advices , lets start the show in here
  • #6
  • Sep 28, 2015 9:19am Sep 28, 2015 9:19am
Quoting 7thman
I'm trying to see the link between fundamentals aspect and real life market by making an exel file which is contain previous data from CDP,CPI,PMI,interest Rate,Trade Balance and Unemployment Rate. And I saw nothing. market goes against my analyze 50% of time ---> random ?!?
Hey 7thman ,

May i know what kind of data analysis youre making ? , and if youre quantifying ?
  • #7
  • Sep 28, 2015 9:34am Sep 28, 2015 9:34am
Quoting tylerbose
{quote} i worked on something kind of similar few months ago. it was actually a neural network that i made in matlab, used every previous high impact economic data from X country ( i did the Euro zone and Canada) i ended up with +20 input data over 10 years period ( i could've gotten a bigger set of 20 or 30 year, but it was already huge) the output was one of the economical release ( i mostly focused on GDP, PMI, retail sales ) . altho i got the direction right 95% of the time, i was off by 0.1 or 0.2 more than 50% of the time. the thing is 0.1...
Hey tylerbose,

Im waiting for a friend to borrow me Eviews , meanwhile ill deal with the limitations of excel ( which is not a bad program though) , mathlab is cool too but i believe is better to stock portfolio im i wrong? . Im interest in the conclution you make , Getting a right direction 95% of the time is a great success in data analysis, your p-value must have been close to 0 is good to know one can get to those percetages,but when you mean 0.1 or 0.2 off that means the standard deviation was 10-20% in the forecast? .

We could elaborate a more accurate data panel putting an extended timeframe lets put it 20 years ( because you already make it from 10) , using only the leading indicator we all have proposed until now ( well for start we could use one pair of currency and 2 or 3 indicator) calculate the correlation between a pairs movement and the high impact news , and separetly the standard deviation ; we could have a nice system.

I apreciate you have sharing your experience
  • #8
  • Sep 28, 2015 2:07pm Sep 28, 2015 2:07pm
Quoting ForexSpider.
{quote} Hey tylerbose, Im waiting for a friend to borrow me Eviews , meanwhile ill deal with the limitations of excel ( which is not a bad program though) , mathlab is cool too but i believe is better to stock portfolio im i wrong?
never heard of Eviews, but it looks interesting.
as for matlab, its not specifically a financial related software (altho it has a financial toolbox), but its really powerful and has everything you need for advanced mathematical work.



Quoting ForexSpider.
{quote} when you mean 0.1 or 0.2 off that means the standard deviation was 10-20% in the forecast?
no, i mean that when i used to forecast GDP (which is X.X% ) i was off by 0.1 or 0.2.
for example, when i get a forecast of 0.7, the actual GDP would be 0.8 or 0.6.
its close but not close enough.

i think its possible to make it a bit more precise, but i got fed up with the amount of data i had to deal with.
i wish they were already available and the right format so i wouldn't have to input everything by hand. it was a pretty annoying job

Quoting ForexSpider.
{quote} We could elaborate a more accurate data panel putting an extended timeframe lets put it 20 years ( because you already make it from 10) , using only the leading indicator we all have proposed until now ( well for start we could use one pair of currency and 2 or 3 indicator) calculate the correlation between a pairs movement and the high impact news , and separetly the standard deviation ; we could have a nice system.
well, if you manage to make a data panel of that size i would be more than happy to help with the forecasts

Quote
well for start we could use one pair of currency and 2 or 3 indicator

i'm not fallowing, what do you mean ?

Quoting ForexSpider.
{quote} I apreciate you have sharing your experience
no problem, mate
i lose money for a living
  • #9
  • Sep 28, 2015 7:55pm Sep 28, 2015 7:55pm
Quoting tylerbose
{quote} never heard of Eviews, but it looks interesting. as for matlab, its not specifically a financial related software (altho it has a financial toolbox), but its really powerful and has everything you need for advanced mathematical work.
{quote} no, i mean that when i used to forecast GDP (which is X.X% ) i was off by 0.1 or 0.2. for example, when i get a forecast of 0.7, the actual GDP would be 0.8 or 0.6. its close but not close enough. i think its possible to make it a bit more precise, but i got fed up with the amount of data i had to deal...

Eviews is a statistical package only with econometrics tool also , i havent put my hands on it yet but when i do , ill be happy to share if it is a better program than Mathlab or not.

Yes that is what i meant with standard deviation , is like the error of the estimated value , the higher the "σ" value the more error but i still think 0.1 to 0.2 is a good number though. Im working on the data panel right know i will post here when finished and we can start making the forecasts

I meant what pair we want to calculate ? for example we grab USD/JPY or GBP/USD and 3 leading indicator example GDP Unenployment and Inflation and we see how the pair was affected by the news in the historical data

PD: dont know how too segmentate the responses like your previous post
  • #10
  • Last Post: Sep 29, 2015 10:22am Sep 29, 2015 10:22am
Quoting ForexSpider.
{quote} I meant what pair we want to calculate ? for example we grab USD/JPY or GBP/USD and 3 leading indicator example GDP Unenployment and Inflation and we see how the pair was affected by the news in the historical data
forecasting accurately major news releases is already a challenge big enough, as for how the market will react to it, thats another story, we should leave that for later i think.
i lose money for a living
0 traders viewing now