Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Arbitrage with no-interest broker 147 replies

eur/usd arbitrage broker? 4 replies

does you broker ban arbitrage on manipulated price quotes? 2 replies

My idea for broker arbitrage 13 replies

Broker arbitrage BIG TIME 4 replies

Attachments: My Own Broker Arbitrage
Exit Attachments

My Own Broker Arbitrage

Quoting Iceworld79
My software will try to close the one-leg order every 30 seconds (5 times maximum, if failed, it will beep and send u an email). I leave it run all day long without any problem. The thing is just slippage which kills the profit. Technically, my software should be responsible fast enough to the market. The best weapon of broker against us is slippage, which does work. I feel it is painful if they gave you 4-5 pips slippage every time. It is just the brute truth, I even got 10 pips slippage. since we are only targeting 1~2 pips every trade, the slippage...
Iceworld79,

Your software use MT4 or FIX/Java API ?
Quoting jeuro
Therefore, if most have same suppliers, when a real arbitrage occurs ( up in the levels we do not have access to)
Institutional arbitrage on liquid pairs is rarely more than a few fractions of a pip, and rarely more than a few 10's of milliseconds - liquid pair arbitrage is removed in co-location in a few ms at most - what remains is low liquidity pairs, matching across geographically distributed matching centres (EBS has 1 in US, 1 in UK and 1 in Asia - Reuters has 1 in UK), and client flow in 2nd tier brokers (non EBS/Reuters rates) . By the time they reach the retail market they have been submerged by the distribution 'artifacts' - ie what you are seeing are likely not the arbs from the institutional market at all - they are likely artifacts of the retail price distribution hierarchy including poorly performing retail trading systems - they are likely NEW arbitrage, not sourced from the top level.

They are therefore possibly fresh opportunities - to be taken very carefully

Quoting jeuro
Recently, as per request of one of my passive investor I tested one commercial software between Alpari UK and Fxdd.
....

One of my accounts loss some $500. and the other won $512 using 0.10 lot. That means if I did it with one full lot. Would have drained $5000 from one account and make $120. That hardly covers the wire fees to replenish from one broker to the other. Conclusion and advise to my investor. Needs at least 15K, per broker, and at least 4 of them to alternate. Recovering cost of the software 3 to 6 months.
I mentioned long ago developing an 'equity pump' or equity management strategy where they are opened for long enough to 'move' equity from one broker to another. You can't guarantee that you can leave the equity in correct broker each time, but over time you should be able to 'steer' the equity sufficiently to balance the broker in most cases

In the institutional world prime brokers are used to net positions on brokerage accounts - for a fee of course. These fees are comparably with brokers fees
Quoting Iceworld79
Anyone has new idea on how to deal with the slippage?
A real limit order (on a non MT4 system) will give you a guaranteed price, but not a guaranteed fill. ie no slippage - but maybe no fill at all

Very few MT4 systems offer real limit orders though.

(Ill have to eat this comment - its out of date with the thread as the disadvantages have been highlighted earlier)


Quoting FxSwordfish
Thank you, Ianj1. How about giving me those names so I can test them live and report back to the group? RR's endeavor needs a boost of encouragement from live trading. I don't mind doing it - the real work
I would guess Oanda are honest , perhaps not arbable, but likely honest

Ok - How about a new idea - only execute an arb with a limit order so that it can only slip one way (you cant guarantee it wont slip both ways) but try and ensure that the short term trend is supporting it so more often than not slippage will be covered by price momentum

Of course this is not pure arb - its arb entry supported by short term momentum
Quoting FXEZ
It doesn't even take 100 transactions in a single day to alert the broker. A consistent pattern of (small) profits over a couple of weeks will do it given even 5-15 trades per day, particularly if those trades are small or the holding period is small. But if you're making $$$, you're on the radar regardless of the strategy employed.

Given how easy it is to create a report and to sort that report by equity change over a given period of time, don't you think that any broker who has a part time intern programmer has a monthly, weekly and daily profits...
Then make sure you LOSE on the dodgy brokers, and not by similar small amounts each time - just a steady net loss. take yourself off that radar if you can. You can still be detected but you are gonna be further down that list. You have the power to vary the win/loss profile - its just more work to do it
Quoting surfeur
Iceworld79,

Your software use MT4 or FIX/Java API ?
MT4 for now. I have an communication adapter architecture, it can be easily changed to Fix/API if someone can show me it does work. I would like to invest the time and efforts on it.
Quoting Iceworld79
MT4 for now. I have an communication adapter architecture, it can be easily changed to Fix/API if someone can show me it does work. I would like to invest the time and efforts on it.
It works very well at the prime brokerage level. The problem is that's where a number of HFT firms are trading. These are the guys who measure in micro/nanoseconds.
Quoting Ronald Raygun
It works very well at the prime brokerage level. The problem is that's where a number of HFT firms are trading. These are the guys who measure in micro/nanoseconds.
My program is build on ANSI C and use C# to build the GUI. From the appearance of arb opportunity to the order sending, my program should be quick enough (~1ms). The thing is many HFTs use custom hardware, such as FPGA to accelerate (which is my expertise in some way). I know for sure that my program can not compete with FPGA with the speed. If we go that way, we are in a different battle.
Quoting ianj1
Then make sure you LOSE on the dodgy brokers, and not by similar small amounts each time - just a steady net loss. take yourself off that radar if you can. You can still be detected but you are gonna be further down that list. You have the power to vary the win/loss profile - its just more work to do it
Or avoid the dodgy brokers but change the underlying strategy so you accept some directional risk with the potential for greater reward than just a pip or two. Perhaps I'm still an idealist but sites like Forex Factory should be filled with threads teaching useful information. Instead of technical analysis or price action, threads on how to design a strategy off the arbitrage trade. If I had started with this it would have saved me a lot of time. This thread and a few (very few) others on this site are on the right track.
Quoting Iceworld79
My program is build on ANSI C and use C# to build the GUI. From the appearance of arb opportunity to the order sending, my program should be quick enough (~1ms). The thing is many HFTs use custom hardware, such as FPGA to accelerate (which is my expertise in some way). I know for sure that my program can not compete with FPGA with the speed. If we go that way, we are in a different battle.
That's largely the reason why I'm trying to do this at the MT4 broker level. The opportunities are still large enough to work for us, but too small to be profitable for the HFT funds.
Quoting ianj1
Then make sure you LOSE on the dodgy brokers, and not by similar small amounts each time - just a steady net loss. take yourself off that radar if you can. You can still be detected but you are gonna be further down that list. You have the power to vary the win/loss profile - its just more work to do it
If I knew how to consistently lose on the dodgy brokers, I really don't need to hedge.

The arb swings in either direction in a random fashion. That alone should help you stay below the radar; you end up win and lose 50% of the time with any one broker. In addition, trading with a large pool of brokers should dilute their attention.
Quoting FxSwordfish
If I knew how to consistently lose on the dodgy brokers, I really don't need to hedge.

The arb swings in either direction in a random fashion. That alone should help you stay below the radar; you end up win and lose 50% of the time with any one broker. In addition, trading with a large pool of brokers should dilute their attention.
In a perfect world where re-balancing accounts could be done quickly and without costs (wire fees, etc..) then this would be ideal.

In general tough, this is a game best played by HFTs, and any advantage found now might not last long.
FXGears.com
THANK YOU! I've been saying that for a while now!

Quoting FxSwordfish
trading with a large pool of brokers should dilute their attention.
More like we are the dwarves getting trampled by the giants.

Tom Brady felt that. Oh, sorry Pats fans. That was too soon.

Quoting Jack_Larkin
In a perfect world where re-balancing accounts could be done quickly and without costs (wire fees, etc..) then this would be ideal.

In general tough, this is a game best played by HFTs, and any advantage found now might not last long.
What exactly are you looking for? I've got a communication adapter to between MT4 and pretty much anything else you can dream up (as long as it can read from unmanaged code).

Now, the fix protocol is something I've been dieing to get the time to play with.

Quoting Iceworld79
MT4 for now. I have an communication adapter architecture, it can be easily changed to Fix/API if someone can show me it does work. I would like to invest the time and efforts on it.
Quoting nondisclosure00
What exactly are you looking for? I've got a communication adapter to between MT4 and pretty much anything else you can dream up (as long as it can read from unmanaged code).

Now, the fix protocol is something I've been dieing to get the time to play with.
What I mean adapter is that I only need some client code to translate fix message, no need to change server side. my software is currently using similar thing to communicate with MT4s. I think it is broker personality which matters. Technically, not hard. but to get real profit, I now have some doubts. although sad.
Although sad, I found this route might be a dead-end. From this week, 80% percent of my trades turn out to be loss. Same brokers, same software, at the beginning, things are much better. Apparently, brokers has some weapons to defend themselves against arb. You can get something at the beginning, soon the situation will change. Any thoughts?
I'm not sure why brokers should prevent this arb? 50 % of your trades at one broker will loose thats a fact.
Quoting Iceworld79
Although sad, I found this route might be a dead-end. From this week, 80% percent of my trades turn out to be loss. Same brokers, same software, at the beginning, things are much better. Apparently, brokers has some weapons to defend themselves against arb. You can get something at the beginning, soon the situation will change. Any thoughts?
How many brokers are you trading with? You must have drawn their attention. Now they launch the "automated individual pricing" on you by adding more slippage to your orders.
Quoting Kilian19
I'm not sure why brokers should prevent this arb? 50 % of your trades at one broker will loose thats a fact.
Well, in my opinion, it is pretty easy to detect, but I'd rather not reveal it in public
Quoting FxSwordfish
How many brokers are you trading with? You must have drawn their attention. Now they launch the "automated individual pricing" on you by adding more slippage to your orders.
I used 4 brokers, one of them is ECN. eventually, all brokers became against me. I guess what you were talking is what they were doing. I've stopped my live testing until some new thoughts come in.
0 traders viewing now