you know electric this could work if every pair was broken down scientifically kind of like we did with gbpusd, the 10 best non correlated pairs were chosen, and the API was automated on oanda the other option which remains which would need a large account though is what I always maintained, the only cost effective way to hedge this and trade at higher leverage would be to buy 6 month calls and puts on the pairs VERY VERY far out of the money (which would protect against just the kinda of devastating move I talked about earlier, where GBPUSD moved 3500 pips within 6 months)
jasonjm, you are a trader genius...I mean that. Can i do this manuelly while holding down a full time job (working from home and replacing trades in between phone calls)? Or... Do I need the API Michael B. P.S. Folks this is exactly the interaction I dreamed of. I am just an average trader like most here. This is not an ego thing. This is what ET should be about. I hope this Journal gets archived and many traders can visit it for years to come. P.S.S. The API at Oanda costs $650.00/mo. (high volume = $0)
the way to calculate how good a pair is for this system would depend on 3 things in my opinion 1) what is the yearly or 10y range 2) what is the spread cost on oanda? 3) what is the average daily amount of pips you can hope to capture trading long / short (average realized profit per day in pips) once an equation is calculated to work out a number using those 3 inputs, then we can compare pairs directly and figure which look like the best pairs It would be great if 10 pairs could be chosen where no single currency appears more than 3 times - that would reduce risk a lot the only irritating thing is that the USD pairs are generally the very low spread pairs (with a few exceptions) as for trading this manually? I could never, i would crack from the pressure within a few days - I dont see how anyone could trade this manually, you have to rake in every pip you can from range trading, so what happens when you sleep? what happens when you go out? what happens when you just feel tired? this would need a large account to offset the cost of having a serious coder paid to implement the API, and a large account to reduce the API fees per month or at least generate enough profit (once you start trading millions per month the API costs fall) then we come to the question, is all this work worth the result? is there no better trading idea? Another out of the box way of looking at this system may be as follows : basically you are betting on the fact that the pairs range more than they trend Perhaps an idea would be to just sell calls and puts, and close them at break even (if they expire in the money), pocketing the premiums..... if the pairs really range more than they trend, this should also yield money with a lot less work, and a hell of a lot fewer trades. Maybe I will look into that....
Ok, I am going into MaTAF.NET..(some say the the correl function in the spreadsheet works too). I will pick some pairs first...lets see if we concur. I will post first. Michael B.
Spreads Rank order: (they can variate however) EUR/USD 1.5 EUR/GBP 1.8 AUD/USD 2.0 EUR/CHF 2.0 EUR/DKK 2.0 AUD/JPY 3.0 USD/JPY 3.0 EUR/JPY 3.0 USD/CHF 3.0 CHF/JPY 3.0 NZD/USD 4.0 GBP/USD 4.0 USD/CAD 4.0 USD/HKD 5.0 USD/SGD 5.0 USD/SAR 5.0 USD/THB 5.0 GBP/JPY 6.0 GBP/CHF 6.0 EUR/AUD 10.0 AUD/NZD 10.0 EUR/HUF 25.0 USD/DKK 30.0 EUR/PLN 35.0 USD/NOK 50.0 EUR/NOK 50.0 EUR/SEK 50.00 USD/ZAR 125.0 USD/MXN 175.0 EUR/CZK 350.0 100 day Correlation: EUR/USD GBP/JPY 01- EUR/GBP GBP/USD 05+ AUD/USD USD/JPY 07+ EUR/CHF USD/CAD 03- EUR/DKK USD/CAD 02+ AUD/JPY EUR/CHF 08+ USD/JPY GBP/USD 04- EUR/JPY EUR GBP 01+ USD/CHF GBP/JPY 02- CHF/JPY USD/CAD 29- NZD/USD USD/JPY 15+ GBP/USD USD/JPY 04- USD/CAD CHF/JPY 29- USD/HKD GBP/USD 16- USD/SGD EUR/JPY 04+ USD/SAR USD/THB GBP/USD 07- GBP/JPY EUR/PLN 11+ GBP/CHF EUR/JPY 07- EUR/AUD EUR/JPY 10+ AUD/NZD EUR/HUF USD/DKK EUR/PLN GBP/JPY USD/NOK EUR/NOK EUR/SEK USD/ZAR USD/MXN EUR/CZK EUR/CHF 08- This is what I got so far...about the ranges I will put what I have.. Michael B.
Why would it matter if the pairs are uncorrelated? The only thing that matters is that the pairs are uncorrelated when trending, i.e. one pair trends while another pair is rangebound. I don't think that type of correlation is what is calculated. Also, knowing how far some of the pairs have trended in the past, trading in the current is almost pure gamble, trying to grab realized before the trends starts Another thing, the realized from one pair can't be used to offset unrealized from another pair. You have to treat each pair individually, since any of the can trend at any given time. IMO, much much much better to trade only in the direction of the underlying trend Also, without backtesting this to see exactly how many pips you would have captured eg. from 2002-2003 while accumating unrealized you would be trading blind in the current Don't hope the realized will cover you, know it
10 Y Ranges EUR/USD 1.4462-0.8225 = 0.6237 EUR/GBP 0.9019-0.5671 = 0.3348 AUD/USD EUR/CHF EUR/DKK AUD/JPY USD/JPY 121.80-116.78 = 005.02 EUR/JPY 162.48-88.77 = 73.71 USD/CHF 1.8311-1.1279= 0.7032 CHF/JPY NZD/USD GBP/USD 1.9561-1.3680 = 0.5881 USD/CAD USD/HKD USD/SGD USD/SAR USD/THB GBP/JPY 241.09-131.24 = 109.85 GBP/CHF EUR/AUD AUD/NZD EUR/HUF USD/DKK EUR/PLN USD/NOK EUR/NOK EUR/SEK USD/ZAR USD/MXN USD/CZK
There are three terms that I know of: Correlated Anti-Correlated Un-correlated Here is Matafs description (I do not have the algorithm): The following tables represents the currencies correlation. The correlation coefficient show the moves similarity between pairs. If the correlation is high (above 80) and positive then the currencies move in the same way. If the correlation is high (above 80) and negative then the currencies move in the opposite way. If the correlation is low (below 60) then the currencies don't move in the same way. As long as flow>capacity by a certain amount compared to "sized trade ticket range", then there will be enough time to prevail without a margin call, to collect the realized. Mogul, This code can be cracked, it's not gambling. Let's get out of the box and try. No disrespect intended, just trying to extend your brainpower. Your input is needed. Not knowing what infinity is CAN be the all encompassing factor that we can use as an excuse. The unknown can be dealt with as long as flow>capacity. How would trend direction be measured? Wouldn't that introduce directional trading back into this? What happens when the trend changes?
I think it is a fundamental assumption of this system that the sum of the back and forth movements is always greater than the sum of the straight line moves. Even if a market went 4500 pips in one direction, I would be willing to bet that it did more than 9000 pips of back and forth movement along the way, before or after (though if it were sudden it could cause a margin call before you had a chance to recover).. I haven't studied the price series you refer to so maybe I'm offbase. Here is an interesting thread on this subject: http://www2.oanda.com/cgi-bin/msgboard/ultimatebb.cgi?ubb=get_topic;f=15;t=002782#000000
I don't see it. I am so stupid...Why is this so? Another thing, the realized from one pair can't be used to offset unrealized from another pair. You have to treat each pair individually, since any of the can trend at any given time.