My Own Broker Arbitrage | Page 13 | Forex Factory

Options

Go to
Permalinks
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

  • #241
  • Jan 20, 2012 11:07am Jan 20, 2012 11:07am
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?
  • #242
  • Jan 20, 2012 12:18pm Jan 20, 2012 12:18pm
Quoting Ronald Raygun
If by 'REAL work' you really mean:

Reading
Elaborate
And
Long-
Winded
Or
Ruined
Kode

Then yes.
Yes well, I'm mangling some code right now so touche'.
  • #243
  • Jan 20, 2012 1:16pm Jan 20, 2012 1:16pm
Quoting PIP-Squeak
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...
By executing two trades properly, we can hold them open indefinitely--pending margin requirements. That's the largest advantage over trying to figure out whether a broker is leading, or lagging the market.

Keep those ideas coming. The beauty of FF is the diversity of creativity that exists.
  • #244
  • Jan 20, 2012 2:51pm Jan 20, 2012 2:51pm
Quoting Ronald Raygun
By executing two trades properly, we can hold them open indefinitely--pending margin requirements. That's the largest advantage over trying to figure out whether a broker is leading, or lagging the market.

Keep those ideas coming. The beauty of FF is the diversity of creativity that exists.
Just had some brainstorming about the leading/lagging part:

Either the lagging vs. leading is 'momentarly' (meaning the brokers are more or less leading or lagging the same amount, but all of a sudden one becomes much more leading), or one is leading or lagging for a longer amount of time (which is usually the case I guess)

Whenever is momentarly - or whatever would be a good word here - it's going to be hard to determine which is lagging and which is leading. You should wait for another datastreams to confirm that one is leading, in that time the arb will most likely be gone.

Whenever one broker is clearly lagging behind another broker, it should be more easy to recognise. There should be something like an array with all the ticks from the last minute, and some really nifty algoritm to determine the time offset between those two. That algoritm is going to be a headache anyway, as broker will post different amount of ticks in one minute.

Wouldn't it be usefull to store all tickdata on one server, allow all the calculations to be done there, and return the trade info? Might be easier to orginaze, although it's going to cost you valuable milliseconds.

I did start up such a project (using mysql & vb.net), but had to stop saving data due to extremly big db sizes (150 mb a day for 4 (this method should have >10) brokers & 12 pairs, my vps wasn't build for that )

Anyway, just some brainstorming, might help
  • #245
  • Jan 20, 2012 3:31pm Jan 20, 2012 3:31pm
Not that hard.

But read earlier on this thread we talk about the "lagging" type of arb. That's not what we're trying to do hear.

Quoting esquire
Just had some brainstorming about the leading/lagging part:

Either the lagging vs. leading is 'momentarly' (meaning the brokers are more or less leading or lagging the same amount, but all of a sudden one becomes much more leading), or one is leading or lagging for a longer amount of time (which is usually the case I guess)

Whenever is momentarly - or whatever would be a good word here - it's going to be hard to determine which is lagging and which is leading. You should wait for another datastreams to confirm that one is leading, in that...
  • #246
  • Jan 20, 2012 4:25pm Jan 20, 2012 4:25pm
Quoting PIP-Squeak

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.

Maybe I missing something here, but could’nt this help solve some of the issues?
You need to hedge your position to make sure you can make profit. Lets say Joes is off 3 pips. We are opening the trade waiting for him to come back to the mean price. But until this happen the price could have always changed 30 pips against us and Joe's price is still 3 pips off. You really need the hedge.
If Joe's price is higher than it is supposed to be it doesn't mean that the overall pair can still move up, just think about it. The market moves randomly at that point and we don't have any tools to suggest where it is going to move so this would be just guessing.
  • #247
  • Jan 21, 2012 9:59am Jan 21, 2012 9:59am
Quoting Kilian19
We are opening the trade waiting for him to come back to the mean price. But until this happen the price could have always changed 30 pips against us and Joe's price is still 3 pips off. You really need the hedge.
Yes I suppose that is correct. It would be interesting to see how often that occurs versus how often Joe spikes away from the mean and then back.

We have been running Jason's ARB program in demo mode just to see how often it records an ARB. I realized from watching it that there is no way to take the ARB at it's peak. Since we don't know how long the ARB will last and how large it will become before it collapses we need to trigger our hedge as soon as it meets our criteria. (i.e X number of PiPs). Therefore, you would never place a trade larger than what your criteria for opening a position was. (No 30 PiP ARBs if your criteria is set to take a trade when an ARB hits 10 PiPs) AND since we don't know how long the ARB condition will last until AFTER it has occurred, we would end up taking all ARBs that meet our criteria even if it collapsed instantaneously. (Since Jason's record of how long the ARB was present and how large it was shows after it has occurred, it's worthless for trading purposes. All it does is show you the opportunity you just missed!)

Anyway, how are these two issues addressed in what is being developed here? Or is it even an issue? Not that I wouldn't take 10 PiP hits everyday, but it seems to me that the reality is your never going to hit a 30 PiP ARB unless that is your criteria for taking a position, and that speed of execution is the only solution to trying to take a position before the ARB collapses.
  • #248
  • Jan 21, 2012 11:54am Jan 21, 2012 11:54am
Quoting PIP-Squeak
Since we don't know how long the ARB will last and how large it will become before it collapses we need to trigger our hedge as soon as it meets our criteria. (i.e X number of PiPs). Therefore, you would never place a trade larger than what your criteria for opening a position was. (No 30 PiP ARBs if your criteria is set to take a trade when an ARB hits 10 PiPs) AND since we don't know how long the ARB condition will last until AFTER it has occurred, we would end up taking all ARBs that meet our criteria even if it collapsed instantaneously....
You could very possibly get a fill that is much higher arb than your set criteria during news time. Volatility is your friend and slippage is your enemy, assuming you are using a VPS with the least latency to data center.
  • #249
  • Jan 21, 2012 12:50pm Jan 21, 2012 12:50pm
Quoting PIP-Squeak
Yes I suppose that is correct. It would be interesting to see how often that occurs versus how often Joe spikes away from the mean and then back.

We have been running Jason's ARB program in demo mode just to see how often it records an ARB. I realized from watching it that there is no way to take the ARB at it's peak. Since we don't know how long the ARB will last and how large it will become before it collapses we need to trigger our hedge as soon as it meets our criteria. (i.e X number of PiPs). Therefore, you would never place...
You might want to start running the thing on Live very soon. Arbs always work in demo. You can try Live with tiny lots to see if they really work. You dont want to run out of your 60 day trial period to return the product.
  • #250
  • Jan 21, 2012 5:16pm Jan 21, 2012 5:16pm
You need to remember, this is a "scalping" type system. Small profits. If you're looking for something of that size on a regular basis, this isn't for you.

Quoting PIP-Squeak
Yes I suppose that is correct. It would be interesting to see how often that occurs versus how often Joe spikes away from the mean and then back.

We have been running Jason's ARB program in demo mode just to see how often it records an ARB. I realized from watching it that there is no way to take the ARB at it's peak. Since we don't know how long the ARB will last and how large it will become before it collapses we need to trigger our hedge as soon as it meets our criteria. (i.e X number of PiPs). Therefore, you would never place a trade...
  • #251
  • Jan 21, 2012 7:18pm Jan 21, 2012 7:18pm
Quoting PIP-Squeak
Not that I wouldn't take 10 PiP hits everyday, but it seems to me that the reality is your never going to hit a 30 PiP ARB unless that is your criteria for taking a position, and that speed of execution is the only solution to trying to take a position before the ARB collapses.
I can assure you now that live 10 pips arbs are non-existent except DURING a news event or intervention, in which case you are as likely to get 20-100 pips slippage as 10 pips of arb. Please don't base your live expectations on performance in demo. Of all strategies for trading an arb has the least correlation between demo and live of any other strategy .. period

If they are hittable at all they will be no more than 1 or 2 pips in excess of spread/commission. 10 Pips arbs on demo system are simply because the demo system isn't configured to track live properly. You need to manage your expectations - arbs only work by building a decent profit from a large number of trades, each making 1-2 pips max profit.

Arbs in the interbank world are usually in the fraction of a pip range, relying on high volumes and trade sizes
  • #252
  • Jan 21, 2012 9:18pm Jan 21, 2012 9:18pm
I can't say this is typical, but happened quite often. A live trade example.

http://i162.photobucket.com/albums/t...oesarbwork.gif
  • #253
  • Jan 22, 2012 11:03pm Jan 22, 2012 11:03pm
Quoting FxSwordfish
You could very possibly get a fill that is much higher arb than your set criteria during news time. Volatility is your friend and slippage is your enemy, assuming you are using a VPS with the least latency to data center.
Hummmm? It seems to me that the chances of slippage go up exponentially during volatility. Based on the fact that speed of execution is an extremely critical part of this strategy I would think that trying to get fills at the speed of light during a fundy would drop dramatically. In addition, the chances for re-quotes also goes up exponentially during volatility. I would think a re-quote on one side of the hedge and a fill on the other side would be killer for this strategy.

I guess I would have to agree with ianj1
“I can assure you now that live 10 pips arbs are non-existent except DURING a news event or intervention, in which case you are as likely to get 20-100 pips slippage as 10 pips of arb.


Quoting bravester
You might want to start running the thing on Live very soon. Arbs always work in demo. You can try Live with tiny lots to see if they really work. You dont want to run out of your 60 day trial period to return the product.
I can’t! I wasn’t one of the lucky ones who got to pay the 2k to purchase the system before the offer period expired. I’ve yet to see any LIVE trading results from anyone. Jason presented NO valid 3rd party evidence of success. Maybe we can get some of the people that shelled out the 2k to provide the evidence of success before their 60 day trial period runs out.


Quoting nondisclosure00
You need to remember, this is a "scalping" type system. Small profits. If you're looking for something of that size on a regular basis, this isn't for you.
I think you misunderstood my question. 10 PiPs was just some arbitrary number I picked for illistration. My question was more along the lines of how do you capture the largest ARB possible? I don’t see any way to take an ARB at it's peak. (if there even is one). Since we don't know how long the ARB will last and how large it will become before it collapses we need to trigger our hedge as soon as we identify the ARB, or as soon as the ARB meets our criteria, say 2 PiPs over spread cost. Correct?

If that’s true we would need to take EVERY trade that meets our ARB + Spread criteria as soon as it meets our requirements. Therefore we would never get an ARB greater than our ARB/Spread criteria even if it did occur. Consequently ALL of our trades would be at our ARB/Spread criteria and no greater (baring a lucky, or UN-lucky fill). Correct?

Anyway, I guess I was wondering how or if there was a way to take the largest ARB possible if our trigger is at x over spread.
  • #254
  • Jan 23, 2012 1:20am Jan 23, 2012 1:20am
Pip-Squeak

After what... a month there is still no positive data on forexarb. Either they are all rich or too ashamed to post they've been scammed.
Truth be told I don't think very many shelled out the 2k since there was no 3rd party verified statements showing it worked.

Now I've been watching the one Fxswordfish has for a week now. It's not cheap but at least he's up like $40 bucks in the first week. That EA developer at least shows real live trades and its closer to what these programing experts on this thread say.
And that is, there are not many arbs over a 1/10 of a pip and the system has to be automated or you won't catch them.
  • #255
  • Edited 1:57am Jan 23, 2012 1:43am | Edited 1:57am
Has anybody tried this?

www.trademakersys.com
from an e-mail:
"
Hi,
TradeMaker arbitrage is design for retail traders to do automatic risk
free trading. It can be work with upto 10 mt4 brokers same time. The
price is for the system is 2000 Usd but we have launching promotion with
discount of 40%. We give you
All you need is to purchase the license, hire a vps/dedicated server
and we configure it for you. We give you a demonstration of video
created.

"
  • #256
  • Edited 5:22am Jan 23, 2012 3:58am | Edited 5:22am
Looks like the $2000 mark is kind of the market price for such product.

There are 2 things which I fail to understand:

1. if their program is sooooo good why they don't make millions of $$$ for themselves rather than sharing it with anyone... after all if it is

2. Assuming that there is a reason that it makes better sense to sell such product to others why charge a once off fee ... my business logic tells that it would make more sense to charge a monthly fee of $100-$200, which will attract a lot more clients, becuase they could afford the monthly sub and they will gladly pay a monthly fee from profits they make and eventually such relationship will become win-win forever. The application writer will make more money long term and the clients will happily pay that.

It therefore makes me wonder why all of them without exception prefer to go for a once off fee ... it can only be one of two reasons ... a) the program doesn’t really work or there are issues with that which only they know. b) they are simply desperate for the cash and are not business minded people and therefore looking for the short term gratification rather than the long term service provider client relationship.
Quoting joancb
Has anybody tried this?

www.trademakersys.com
from an e-mail:
"
Hi,
TradeMaker arbitrage is design for retail traders to do automatic risk
free trading. It can be work with upto 10 mt4 brokers same time. The
price is for the system is 2000 Usd but we have launching promotion with
discount of 40%. We give you
All you need is to purchase the license, hire a vps/dedicated server
and we configure it for you. We give you a demonstration of video
created.

"
  • #257
  • Edited 7:13am Jan 23, 2012 6:51am | Edited 7:13am
Quoting PIP-Squeak
Hummmm? It seems to me that the chances of slippage go up exponentially during volatility. Based on the fact that speed of execution is an extremely critical part of this strategy I would think that trying to get fills at the speed of light during a fundy would drop dramatically. In addition, the chances for re-quotes also goes up exponentially during volatility. I would think a re-quote on one side of the hedge and a fill on the other side would be killer for this strategy.

I guess I would have to agree with ianj1
“I can assure you now that...
I do not disagree with you both. I am new in this arb trading. Here was one trade happening on Jan 17 when German economic sentiment was released at 5:00 AM eastern standard time. Sell price on top of buy price. Compare "trigger" and "fill price" for slippage. Guess what, net slippage was in my favor and the arb was 12.7 pips. So, I would agree with you more than I disagree with you. It is probably better not to trade the news because the reverse could happen.


http://i162.photobucket.com/albums/t...eNewsTrade.gif
  • #258
  • Jan 23, 2012 6:58am Jan 23, 2012 6:58am
Fxswordfish, you know quite well that it can go 12 pips your way the one day and 30 pips against you the next day... not sure what are the stats on that, but usually u r going to lose in my view if you trade during volotile times (news etc...)
  • #259
  • Jan 23, 2012 7:36am Jan 23, 2012 7:36am
Quoting FxSwordfish
Compare "trigger" and "fill price" for slippage. Guess what, net slippage was in my favor and the arb was 12.7 pips. So, I would agree with you more than I disagree with you. It is probably better not to trade the news because the reverse could happen.


http://i162.photobucket.com/albums/t...eNewsTrade.gif
That looks to me like a -12.7 pip 'arb', or about 2.4 pips the other way, but then i don't entirely understand what i am looking at (is it the feed prices at the time of the arb ?)
  • #260
  • Edited 11:59am Jan 23, 2012 7:50am | Edited 11:59am
Quoting ianj1
That looks to me like a -12.7 pip 'arb', or about 2.4 pips the other way, but then i don't entirely understand what i am looking at (is it the feed prices at the time of the arb ?)
Feed price is the trigger price. Fill price is the transaction price. Arb is defined as (Ask - Bid) between 2 brokers. The arb criteria for the trade is set for 1.9 pips (or -0.00019). As you can see, the trigger price met the criteria (ask 1.27730 - bid 1.27764 = -0.00034). As you know, you always buy ask and sell bid. Normally Ask is higher than Bid. So if Ask - Bid is negative, it is negative "cross-spread" between two brokers. The order came back (ask1.27740 - bid1.27867 = -0.00127). The absolute value is the arb, thus 12.7 pips. Slippage is within the same broker when you compare the market price at the time and the fill price afterwards. Clear as mud?
0 traders viewing now