- This topic has 68 replies, 20 voices, and was last updated 10 years, 9 months ago by
.
- Posts
Ok, It’s action time. Rough system overview: 1. Historical data filtering. Every decent statistical study “filter” its data and exclude abnormal “spikes”. Using filtered data will also make it easier for any learning algorithm and reduce the “error” margin. 2. Filtering Real-time data feed. Shouls use same methods as #1. 3. Channel of probable future price range. Create an extended channel into the “future” – all predictions should fall withinn that channel range. 4. Rate of change (ROC). When we get a prediction, we need to know WHEN it will get to that price prediction. Enough for now… G.
Dear G.,
A few weeks ago, I tried hard and finally were able to get a NARX net feeding itself on pure GBPUSD data. I was able to get a number around 1.54 on April 27th -where GU was around 1.52-. I found the result nonsense and moved on. Now I suspect whether it was luck or did that hold any potential…
I had 4 hidden layers.
Do you think this is the right track to follow?
Best regards…
Nothing has ever motivated me more than this...
Ok, It’s action time. Rough system overview: 1. Historical data filtering. Every decent statistical study “filter” its data and exclude abnormal “spikes”. Using filtered data will also make it easier for any learning algorithm and reduce the “error” margin. 2. Filtering Real-time data feed. Shouls use same methods as #1. 3. Channel of probable future price range. Create an extended channel into the “future” – all predictions should fall withinn that channel range. 4. Rate of change (ROC). When we get a prediction, we need to know WHEN it will get to that price prediction. Enough for now… G.
Dear G., A few weeks ago, I tried hard and finally were able to get a NARX net feeding itself on pure GBPUSD data. I was able to get a number around 1.54 on April 27th -where GU was around 1.52-. I found the result nonsense and moved on. Now I suspect whether it was luck or did that hold any potential… I had 4 hidden layers. Do you think this is the right track to follow? Best regards…
When you “judge” some prediction as “nonsense” – you are actually saying that YOU are predicting the opposite….
You should ask yourself, in those cases, how did you come to that blind, “definite” prediction of yours… and what is your historical blind “predictions” track record…
G.
My answer to a question raised about high probability in latest Skype session:
If the market opened in the lower 20% of its daily range and closed in the upper 80% of its daily range,
a SELL setup is of high probability for the next day (vice versa for BUY).
[Night data (my TimeZone – NY Close till EU Open, i.e. Asian Session) is ignored. The range should be created from day-sessions only.]
G.
This reply has been reported for inappropriate content.
i use alpari broker[/url] they always provide good support, among others, providing daily trading analysis and online customer support. Alpari is one of the world’s largest Forex brokers. If you not interest to trading you can invest in Alpari PAMM[/url] . The PAMM Accounts service allows investors to earn on Forex without trading on their own and managers to get additional income through efficient management of investors’ funds.”
This reply has been reported for inappropriate content.
i use alpari brokerthey always provide good support, among others, providing daily trading analysis and online customer support. Alpari is one of the world’s largest Forex brokers. If you not interest to trading you can invest in Alpari PAMM . The PAMM Accounts service allows investors to earn on Forex without trading on their own and managers to get additional income through efficient management of investors’ funds.”
Alpari UK went bankrupt after the SNB stop it’s support of EURCHF @1.20 …. ;-}
yeppp!
If you not interest to trading you can invest in Alpari PAMM . The PAMM Accounts service allows investors to earn on Forex without trading on their own and managers to get additional income through efficient management of investors’ funds.
Hmmm – depending on where you live you should be careful with trading for others. Over here in Germany it’s absolutely illegal to do so without a formal permission by the Supervisory Authority ‘BaFin’. And those regulations are tough! I know about a guy who ended up with a 3 years 9 months sentence after trying so without a permission. So take care!
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready.
It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#.
Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good.
BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW.
Any other suggestions?
Thanks.
G.
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready. It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#. Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good. BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW. Any other suggestions? Thanks. G.
SP500 is better or Russell2000, Treasury US 10 YR T-Note. Dollar Index too but it’s very similar to eurusd.
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready. It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#. Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good. BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW. Any other suggestions? Thanks. G.
SP500 is better or Russell2000, Treasury US 10 YR T-Note. Dollar Index too but it’s very similar to eurusd.
How is S&p or Russel correlated to Forex? USDx sure is. I’ll try it. Dow is ~inverse correlated to EURUSD.
Thanks.
G.
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready. It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#. Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good. BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW. Any other suggestions? Thanks. G.
SP500 is better or Russell2000, Treasury US 10 YR T-Note. Dollar Index too but it’s very similar to eurusd.
How is S&p or Russel correlated to Forex? USDx sure is. I’ll try it. Dow is ~inverse correlated to EURUSD. Thanks. G.
Talking about forex indeed everything is correlated: currencies, equities, commodities, treasury. For every move in last 3 specific markets you can find a reaction in forex pairs.In the specific all equities index are correlated, (sp dow russell nasdaq dax footsie topix ans so on …), so it’s my opinion that’s better to use an index based on 500 large companies than an index based only on 30 big companies (that are too big to fail…). And yes, now equities are generally inverse correlated with EURUSD but years ago the exact opposite happens so … correlation changes over time. I don’t understand exactly why you needs an external index, maybe to anchor something, take care if you don’t know already that DXY have 57 % of is weight based on EUR so EURUSD and DXY are very very similar for that reason, but inverted.
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready. It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#. Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good. BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW. Any other suggestions? Thanks. G.
SP500 is better or Russell2000, Treasury US 10 YR T-Note. Dollar Index too but it’s very similar to eurusd.
How is S&p or Russel correlated to Forex? USDx sure is. I’ll try it. Dow is ~inverse correlated to EURUSD. Thanks. G.
Talking about forex indeed everything is correlated: currencies, equities, commodities, treasury. For every move in last 3 specific markets you can find a reaction in forex pairs.In the specific all equities index are correlated, (sp dow russell nasdaq dax footsie topix ans so on …), so it’s my opinion that’s better to use an index based on 500 large companies than an index based only on 30 big companies (that are too big to fail…). And yes, now equities are generally inverse correlated with EURUSD but years ago the exact opposite happens so … correlation changes over time. I don’t understand exactly why you needs an external index, maybe to anchor something, take care if you don’t know already that DXY have 57 % of is weight based on EUR so EURUSD and DXY are very very similar for that reason, but inverted.
I planned to use it as an extra Factor/Weight that correlate or influence the market.
It’s true that S&P include 500 companies, but that’s where the problem is. There are different industries that many times offset each other.
Initially I used Interest rate, but that’s hardly change, and meant nothing in the end result.
G.
After a loooong silence here: the skeleton (or Alpha ver.) of an NN is ready. It’s based on ENCOG NN framework (thanks SaverO…) and .Net C#. Currently testing daily prediction of EURUSD, GBPUSD, AUDUSD. Results looks good. BUT: I think I need at least one more daily data, related to Forex, but outside from that market. I thought about the DOW. Any other suggestions? Thanks. G.
SP500 is better or Russell2000, Treasury US 10 YR T-Note. Dollar Index too but it’s very similar to eurusd.
How is S&p or Russel correlated to Forex? USDx sure is. I’ll try it. Dow is ~inverse correlated to EURUSD. Thanks. G.
Talking about forex indeed everything is correlated: currencies, equities, commodities, treasury. For every move in last 3 specific markets you can find a reaction in forex pairs.In the specific all equities index are correlated, (sp dow russell nasdaq dax footsie topix ans so on …), so it’s my opinion that’s better to use an index based on 500 large companies than an index based only on 30 big companies (that are too big to fail…). And yes, now equities are generally inverse correlated with EURUSD but years ago the exact opposite happens so … correlation changes over time. I don’t understand exactly why you needs an external index, maybe to anchor something, take care if you don’t know already that DXY have 57 % of is weight based on EUR so EURUSD and DXY are very very similar for that reason, but inverted.
I planned to use it as an extra Factor/Weight that correlate or influence the market. It’s true that S&P include 500 companies, but that’s where the problem is. There are different industries that many times offset each other. Initially I used Interest rate, but that’s hardly change, and meant nothing in the end result. G.
Ok, got it. Well try Dow and then SP but I think you get very similar results.
As said each market influence the forex market and in turn it affects the others, is the so-called intermarket analysis and is not easy to model. In a hypothetical model you should put into gold, oil, equities and treasury. The most important, in my opinion, is the bond market. Money flows from the equities (risk on) to bonds (risk off) and vice versa, and the mountain of money is managed by the bond desk. In such a model you absolutely must take account of them because they are the ones that move the world.
Ok, thanks. Bonds are good idea.
Now I have to scan which one to use, because I have only one variable left…
Thanks again.
G.
Ok, thanks. Bonds are good idea. Now I have to scan which one to use, because I have only one variable left… Thanks again. G.
The benchmark is the 10 year US T-note.
The yield index trades only few hours:
http://www.cboe.com/DelayedQuote/advchart.aspx?ticker=TNX
The future tracks price 23 hours:
http://www.cmegroup.com/trading/interest-rates/us-treasury/10-year-us-treasury-note.htmlOne is the inverse of the other, so if yields rise treasury prices fall. Institutionals looks primarily at yields and moves accordingly.
Ok, thanks. Bonds are good idea. Now I have to scan which one to use, because I have only one variable left… Thanks again. G.
The benchmark is the 10 year US T-note. The yield index trades only few hours: http://www.cboe.com/DelayedQuote/advchart.aspx?ticker=TNX The future tracks price 23 hours: http://www.cmegroup.com/trading/interest-rates/us-treasury/10-year-us-treasury-note.html One is the inverse of the other, so if yields rise treasury prices fall. Institutionals looks primarily at yields and moves accordingly.
Thanks again for that. Going to “train” my NN with that.
G.
Initially I used Interest rate, but that’s hardly change, and meant nothing in the end result.
Noted only now this quoted.
You can use the Eurodollar future, it’s based on US Libor, it’s a big and liquid future, basically if rates go up then ED futures go down.
http://www.cmegroup.com/trading/interest-rates/stir/eurodollar.html, look at dec mar jun sept.
- You must be logged in to reply to this topic.