My idea for broker arbitrage | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Broker which allow Arbitrage trading? 71 replies

Arbitrage with no-interest broker 147 replies

eur/usd arbitrage broker? 4 replies

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

Broker arbitrage BIG TIME 4 replies

Attachments: My idea for broker arbitrage
Exit Attachments

My idea for broker arbitrage

  • #1
  • First Post: Edited 12:51pm Apr 6, 2008 11:28am | Edited 12:51pm
We all know that broker's feeds are not synchronized. Compare the feeds of two ECNs or brokers, and you will see different prices (varying by a few pips) at any time.

The idea is to

1) Go long and short on a pair with 2 or more brokers after swap is paid (long/short combos are always negative, so you'd want to avoid it)

2) Close one of the short or long positions when the difference is greater than the spread (if the outlier price is higher, close the outlier long, otherwise the outlier short,) then close all 3 other positions when the price equalizes again

You need to go long and short because you don't know if the price discrepancy will be above or below the current price, and you want your floating equity to be 0 until it happens. IE) Only taking longs or shorts will give negative equity if the price goes against the position.

An EA that takes live price feeds from multiple ECNs and MT4 platforms is optimal. Orders need to be filled at quoted prices to make profit.

EDIT) This does not guarantee profit, as price can move against you before it equalizes, negatively affecting balance. I hadn't thought of this until now. Any thoughts?
  • #2
  • Apr 6, 2008 2:21pm Apr 6, 2008 2:21pm
I found out that the feed from CMS Forex is way faster than any MT4 brokers feed. Especially during high volume times you could see the CMS Forex price drop 20 pips and it took MT4 about 1 second do catch up.
I find the idea of arbitrage quite tempting. To do it professionally
you'd have to calculate a matrix with quotes etc. I think that's how it's done by hedge funds. I'd be very carefull trying to trade that manually.
  • #3
  • Apr 6, 2008 2:24pm Apr 6, 2008 2:24pm
If CMS offers a DDE, you could do it very easily with a spread sheet for both the CMS and the MT4
  • #4
  • Apr 6, 2008 2:25pm Apr 6, 2008 2:25pm
What is DDE?
I once hooked up VTTrader with an Excel Sheet but it was way to slow.
  • #5
  • Apr 6, 2008 2:56pm Apr 6, 2008 2:56pm
Quoting Tjpld
What is DDE?
I once hooked up VTTrader with an Excel Sheet but it was way to slow.
DDe Digital Data Exchange It is a way to bring price feeds into excel at a extremely fast real time rate, from both brokers or as many as you want for that matter. Here is any example
Attached File(s)
File Type: xls DDE-Sample.xls   16 KB | 936 downloads
  • #6
  • Apr 6, 2008 6:43pm Apr 6, 2008 6:43pm
How would you exploit an arbitrage between brokers? What's the buy/sell sequence? Is there any meat on this bone?
  • #7
  • Apr 6, 2008 6:45pm Apr 6, 2008 6:45pm
Quoting tdion
How would you exploit an arbitrage between brokers? What's the buy/sell sequence? Is there any meat on this bone?
Ouch! Did I push some button? sorry about that, thought I was offering discussion. You told me to remove my vouch that I gave you once already...

Now I am really confused... I'll just keep my idea on this to myself. Don't want to cause any grief.

Edit Oh I see you removed the question why I don't vouch for you, Saying I have vouched for everyone else.

No problem, I will unsubscribe this thread.
  • #8
  • Apr 6, 2008 10:11pm Apr 6, 2008 10:11pm
No No No

I meant it as a jest, but then after I read it I saw it could be taken the wrong way.

Of course your input is always welcome.... You're insight is keen.

EDIT) I removed it immediately, but you spotted it quick. Caveat Empatur.
  • #9
  • Apr 6, 2008 10:11pm Apr 6, 2008 10:11pm
have fun having your account band.. And money lost...
  • #10
  • Apr 6, 2008 10:51pm Apr 6, 2008 10:51pm
I understand the general idea of what you're trying to do, but your description below is a little vague. What do you mean "higher" and "lower"..etc? What is the frame of reference? Also, note that you don't have to do this kind of thing on the same pair.. think GU and UJ on one broker (your reference) and GJ on another(the lagging outlier).

I wouldn't want to discourage you but the broker himself probably has robots in place that suck this fountain dry. You may get a few pips in the end, but as you realized, there is no guarantee for that crucial little movement either, as with all things in life. Furthermore, you may well be attacked for scalping. Beware the wrath of ZeusFX, almighty god of the brokerages!


Quoting tdion
We all know that broker's feeds are not synchronized. Compare the feeds of two ECNs or brokers, and you will see different prices (varying by a few pips) at any time.

The idea is to

1) Go long and short on a pair with 2 or more brokers after swap is paid (long/short combos are always negative, so you'd want to avoid it)

2) Close one of the short or long positions when the difference is greater than the spread (if the outlier price is higher, close the outlier long, otherwise the outlier short,) then close all 3 other positions when the price equalizes again

You need to go long and short because you don't know if the price discrepancy will be above or below the current price, and you want your floating equity to be 0 until it happens. IE) Only taking longs or shorts will give negative equity if the price goes against the position.

An EA that takes live price feeds from multiple ECNs and MT4 platforms is optimal. Orders need to be filled at quoted prices to make profit.

EDIT) This does not guarantee profit, as price can move against you before it equalizes, negatively affecting balance. I hadn't thought of this until now. Any thoughts?
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #11
  • Apr 6, 2008 10:59pm Apr 6, 2008 10:59pm
Well this is one of the few inefficiencies in forex.

Furthermore, you can pool together 40 brokers if you had the guts to open account with all of them.

Imagine the permutations available with 40 crossing ALL PAIRS! It'd be hundreds of millions at least.

In fact, if you really let your imagination run wild, you might be able to set up triangular arbitrages across several pairs

EX) EURUSD -> USDCHF -> CHFEUR (this is just 3, but strings of 10 or more are possible)

Then there would be billions of combinations. A Dijkstra greedy algorithm (that I already wrote) uses heuristics to cut the processing down to under a second. I am just not sure about how you would do the buying and the selling.

The question is

1) What is the correct theory to apply?

2) Is it feasible?

I think I've put out an idea that many have thought of, but few have posted.

PS) Plutonite I always enjoy your posts. Thx for stopping by.
  • #12
  • Apr 6, 2008 11:54pm Apr 6, 2008 11:54pm
OK, let's start with a simple example (single pair). You seem to want to open several positions all at once for no prior discrepancy reason, so I'm a little confused with your intention, but here's what I would do:

1) Given say, EURUSD feeds from 5 brokers, you "notice" that 1 out of the 5 has a lower ask/buy price than the ask prices of the other 4. Let's say the discrepancy is 5 pips.

2) Enter long immediately on the lagging (cheaper) EU broker. Let's call him (A).

3) The bid price with (A) should hopefully reach that of the others within a very short amount of time, assuming (A) is lagging the general market and arbitrage at an internal/high level will cause him to come to his senses. You can now close your single position with 5 - 2(spread) pips in profit. If (A) continues to lag, you may choose to close anyway, or wait for greater profit. This is the good scenario.

4) In the bad scenario, your broker (A) was actually "correct" and the ask prices on the other brokers moves down instead. You have lost 2 pips in spread immediately, and the story continues depending on what your entry is like.


As you can see your only hope on a non-centralized market like this where you can't throw money around BETWEEN the brokers, is that you manage to get wind of how one broker price will change based on some others. The derivative pair thing is similar but more complex, due to the fact that there may be discrepancies between both each of the base pairs, which can actually cancel out on the derived (say GBPJPY) pair that you are betting on.

I say stick to normal trading :
Virtue finds and chooses the mean. Aristotle, Ethica Nichomachea
  • #13
  • Apr 7, 2008 2:02am Apr 7, 2008 2:02am
Quoting tdion
Well this is one of the few inefficiencies in forex.

Furthermore, you can pool together 40 brokers if you had the guts to open account with all of them.

Imagine the permutations available with 40 crossing ALL PAIRS! It'd be hundreds of millions at least.

In fact, if you really let your imagination run wild, you might be able to set up triangular arbitrages across several pairs

EX) EURUSD -> USDCHF -> CHFEUR (this is just 3, but strings of 10 or more are possible)

Then there would be billions of combinations. A Dijkstra greedy algorithm (that I already wrote) uses heuristics to cut the processing down to under a second. I am just not sure about how you would do the buying and the selling.

The question is

1) What is the correct theory to apply?

2) Is it feasible?

I think I've put out an idea that many have thought of, but few have posted.

PS) Plutonite I always enjoy your posts. Thx for stopping by.
What you are talking about is TriArb.
There is an excelent thread on http://kreslik.com/forums/viewtopic.php?t=307.

Check it out.
  • #14
  • Last Post: Jun 3, 2008 11:57pm Jun 3, 2008 11:57pm
smjones,

how can you load data from more than one broker into Excel? Your example just had it from one.

Thanks very much!




Quoting smjones
DDe Digital Data Exchange It is a way to bring price feeds into excel at a extremely fast real time rate, from both brokers or as many as you want for that matter. Here is any example
0 traders viewing now