This thread is quite interesting and it looks like it has attracted some of the biggest brains on FF. I certainly don’t understand the technical issues as well as many who have posted here but it seems to me that many of these issues could be resolved by taking a different approach to the ARB conditions, so I humbly offer these thoughts to the super brains to see if it can help.
Maybe I'm wrong but I don't see the necessity to hedge between 2 brokers. In my mind it makes more sense to trade back to the means. Similar to Bolinger bands but between brokers. I think it would be much better (and more profitable) to only make one transaction with one broker instead of hedging with 2 brokers.
Here’s my thought. It is FAR more likely that only one broker is away from the mean. The chances of two brokers both being off the norm at the exact same time in opposing directions is highly unlikely. (Which means by the way, that when trading an ARB with a hedge between brokers your trading the broker with the price anomaly against ANY broker who is at the norm) I think it is much more important to determine which broker is outside of the norm by comparing many brokers to establish the norm and then looking for the anomaly. If 9 out of 10 brokers are at x and one broker is at x+ then we only want to open a position with the 1 broker and trade back to the means.
In other words, if all brokers are within 1 or 2 PiPs of one another except Joe's bucket shop, and he is 15 PiPs from the norm, then not only would you NOT need to hedge between brokers, you would now know which broker to trade with, which direction to trade, and for how many PiPs.
This would cut the spread costs in half. In addition, trading the anomaly would eliminate the necessity to open and close two positions simultaneously. The downside is that it would need to close as soon as the anomaly gap closes or you could turn a winner into a loser. You couldn’t ride the hedge to safety and then close in several hours like has been suggested, however, I still think it would be easier than trying to open and close 2 positions simultaneously.
One of the initial pitfalls I see with this is, if the market moved towards the anomaly. In other words, before Joe's bucket shop moves back to the norm, the market moves towards Joe, or against your position. I'm not sure that this would happen to often but it may require some intelligence built into the program so that it knows when the broker spread is collapsing on the market side and not the brokers side and thereby closes the trade when broker equals market. At least that could minimize your loses.
In addition, it seems to me that this should be broken down into two modules. (I think that’s where you guys are headed also.) There are certainly much better/faster platforms to develop an analysis module in than MQL. I think that a standalone module that receives the data, looks for the anomalies that meet a specific criteria, and then sends trade execution info to an EA would be much faster than trying to process the data in MQL. Moreover, by trading the anomalies back to the means with ONE broker we have eliminated the need for simultaneous execution of 2 trades. That should simplify things a great deal.
Maybe I missing something here, but could’nt this help solve some of the issues?
Maybe I'm wrong but I don't see the necessity to hedge between 2 brokers. In my mind it makes more sense to trade back to the means. Similar to Bolinger bands but between brokers. I think it would be much better (and more profitable) to only make one transaction with one broker instead of hedging with 2 brokers.
Here’s my thought. It is FAR more likely that only one broker is away from the mean. The chances of two brokers both being off the norm at the exact same time in opposing directions is highly unlikely. (Which means by the way, that when trading an ARB with a hedge between brokers your trading the broker with the price anomaly against ANY broker who is at the norm) I think it is much more important to determine which broker is outside of the norm by comparing many brokers to establish the norm and then looking for the anomaly. If 9 out of 10 brokers are at x and one broker is at x+ then we only want to open a position with the 1 broker and trade back to the means.
In other words, if all brokers are within 1 or 2 PiPs of one another except Joe's bucket shop, and he is 15 PiPs from the norm, then not only would you NOT need to hedge between brokers, you would now know which broker to trade with, which direction to trade, and for how many PiPs.
This would cut the spread costs in half. In addition, trading the anomaly would eliminate the necessity to open and close two positions simultaneously. The downside is that it would need to close as soon as the anomaly gap closes or you could turn a winner into a loser. You couldn’t ride the hedge to safety and then close in several hours like has been suggested, however, I still think it would be easier than trying to open and close 2 positions simultaneously.
One of the initial pitfalls I see with this is, if the market moved towards the anomaly. In other words, before Joe's bucket shop moves back to the norm, the market moves towards Joe, or against your position. I'm not sure that this would happen to often but it may require some intelligence built into the program so that it knows when the broker spread is collapsing on the market side and not the brokers side and thereby closes the trade when broker equals market. At least that could minimize your loses.
In addition, it seems to me that this should be broken down into two modules. (I think that’s where you guys are headed also.) There are certainly much better/faster platforms to develop an analysis module in than MQL. I think that a standalone module that receives the data, looks for the anomalies that meet a specific criteria, and then sends trade execution info to an EA would be much faster than trying to process the data in MQL. Moreover, by trading the anomalies back to the means with ONE broker we have eliminated the need for simultaneous execution of 2 trades. That should simplify things a great deal.
Maybe I missing something here, but could’nt this help solve some of the issues?