MT4 Broker Watch: Community Project to monitor MT4 Brokers

Hello MT4 traders,

Is it time to question your datafeed?

I have found consistent and substantial differences between the streaming (real time) quotes and refreshed quotes on my live MT4 datafeed from FXCM. e.g., while data is streaming in real time, bar #1 Close would be reported as 1.2345. However, a few hours later, when the bar becomes, say, bar #36, it might be reported as 1.2346.

While price differences (between streaming/refreshed quotes) can get substantial. Volume differences can easily reach 200%. FXCM claims these differences stem from synchronization issues in the MT4 platform/bridge. It may or may not be the case, but it has the interesting side effect of making any system tuning you might want to do - irrelevant.

Although my main interest is these intra-broker differences (streaming vs. refreshed quotes) it got me going on comparing data feeds between a few brokers (inter-broker). I have started to compare Close prices and Volumes for a series of about 2000 EURUSD M5 data points between a few brokers and found some very interesting deviations.

I believe it makes sense to monitor more brokers, data feeds and TFs (let’s focus on live accounts). This is certainly more than what one person can handle but if there is interest we can do it together. I have attached an indicator that continuously monitors any symbol/TF.

Any insights, thoughts, volunteers?

Additional results (on both topics) and live accounts are more than welcome.

In the next days I will post some results from my running monitors.

Thanks

interesting idea. we are working on something similar. it will gather statistics from subscribers who volunteer to load our 'EA' in their terminal. and our latency page is going to be updated in real-time with statistics.

maybe we can leverage them together...??

ps: our project is only measuring broker issues...latency, requotes, dropped connections, etc..

There is no FX volume reported, volume you see is tick volume, of course that will be completely different at each broker. Charts/candles will be different too based upon a whole multitude of reasons: variance in spread, weekly open/close time, liquidity providers, etc. That's simply the nature of the business. Welcome to OTC FX, enjoy your stay.

Hey BarrySDCA,
I like it. I can see the page title reading "Broker Watch" with entries for all of the above, plus sections like "Stop Hunts from hell" and "Bad Ticks like you could not believe".

I can certainly see leverage for a combined and systematic monitoring. The bottom line is that one strength that we have as a community of retail traders is our mass.

Let me bring some data in so we can get a handle of the best way to organize such a project.

Cheers
Marathon

Edit: Just changed the thread name... (Barry man, its your "inception" :-).

Hey Trader KGB,

Thanks for your input. This would be a good opportunity to take the V word out of the system.

OTC FX volume reported is indeed "tick volume", not "traded volume". Still, as every experienced trader probably knows it caries valuable information in it. At any rate this has no relevance to the main issue: "is it time to question the datafeed we are getting as MT4 traders?".

As an FYI, my initial tests show correlation of ~85% on volume sequences between different brokers - but these indeed represent different OTCounters. The main problem with volume reporting is intra-broker. That is, when a broker refreshes its own datafeed quotes post factum.

To make sure we are on the same page - these on-the-fly-changes also include price changes. So, regardless if you personally use volume or not - it should be your interest to get a consistent datatream that does not change under your feet.

Cheers,
Marathon

Here's an alarming catch of the MT4 datafeed monitoring tool from the first two hours of trading of the current week (Monday Aug-16, 2010 Asia session).

As I mentioned, I'm trading live with FXCM, so it easiest for me to monitor their live feed.

The first image shows a snapshot of GBPUSD. Note the impressive 20 minutes 50+ pips spike.


http://www.forexfactory.com/attachme...1&d=1282105562

For comparison purposes I used a non-MT4 datafeed from another major broker (see next image). Sure enough there are no spikes at all.


http://www.forexfactory.com/attachme...1&d=1282105562


Based on these snapshots seems like FXCM MT4 clients were taken on some 50+ pips field trips (!). A technical malfunction? A market liquidity issue (mind you this is GBPUSD)? A careless stop hunt? Who knows, but what is sure - many fellow MT4 traders lost money in these out-of-no-where ‘spikes’. According to FXCM manager of the Active Traders Group Client Services desk - this is a ‘bad tick’. Duh?

To complete the picture for the same session – here’s a couple of the bad-tick's brothers - AUDUSD and AUDJPY:

http://www.forexfactory.com/attachme...1&d=1282105562



http://www.forexfactory.com/attachme...1&d=1282105562


Needless to say, none of these two additional spikes (AUDUSD, AUDJPY) appear on the other broker's feed.

Guys, this is no joke. This is really happening. Unless we start monitoring, systematically comparing and sharing this information – it will continue like this.

Next post would be about intra-broker datafeed fluctuations.


Cheers,
Marathon

I'm not seeing the screen capture pics.

It's been well documented on the FXCM Discussion thread that their MT4 bridge suffers from bad ticks, which cause the spikes.

I agree it's in every trader's best interest to utilize multiple data sources to ensure that their executions are not out of line with the general interbank quote. Having only traded with ECNs, I rarely encounter this, but I imagine the risk is much greater for those trading with MMs.

@ Trader KGB,

Reloaded the images. Hope its coming through now.

Well, what is a 'bad tick'? What we see above is going on for more than 20 minutes on at least three pairs. This is like 25,000 bad ticks! This is severely wrong.

Isn't FXCM claiming to be a sort of an ECN?

A 'bad tick' would be any LP sending an erroneous quote. Typically this happens at or around the 5pm rollover, but can happen anytime during the trading day.

FXCM is purportedly NDD/STP, not ECN. They use BostonTech for their MT4 bridge, and are trying to wean clients over to StrategyTrader, partially b/c of the data issues you're seeing I imagine.

Hello Marathon
Wath hapenned with your project?

What happen with the spike, did you talke to FXCM about it? I am very curious.

Great Topic.

Rule Number 1 in Retail FX:

Its a rigged game, not in your favor, but in favor of the intermediary.

Rule Number 2 in Retail FX:

Avoid trading with a Board Certified and Serial Registered Bucket Shop. Board certification and serial registration numbers can be found encoded in the "name" of the intermediary you use. For example:

FXCM = Bucket Shop Number 0001-0001A, Certified by the CFTC.
Oanda = Bucket Shop Number 0002-0001A, Certified by the CFTC.

[notice how both numbers are verrrrrry similar in appearance]

You get the idea - I'll move on.


FX Retail is OTC on steroids, but that does not mean impossible to trade. The reality is that even in Retail FX, there is a ton of structure in the data. You just have to know where and how to view the data. Will there be variances in the data? Yes, it is OTC, after all. And, its Global OTC, to make things even more interesting. Having said that, you can run [for example] multiple disparate Retail FX platforms on the same machine and still see relative synchronicity in real-time pricing. But, that is not what the OP started this thread to see.

The OP clearly stated that there are too many discrepancies in a data point's historical value as differentiated from its real-time value and that is a horse of another color. That's called repainting the bar and I've seen it happen, too. This is one reason why using any back-tester on Retail FX data, can present some problems on future real-time trading, using a Bot or EA. If your trading system and/or style, requires tight ranges and precision pip clearances in order to work, then you might run into some problems, when you start to compare your back-test results to your real-time results.

This is why Forward Testing is just as important, even more important, than your Back Testing. Do them both and compare results. You can Scalp Retail FX, contrary to the belief of some - it can be done successfully. However, you will need to increase the Scalp Range, so that discrepancies built into disparate data sources that make-up your transactional data in Retail FX, can be absorbed and thus, mitigated. So, the trade simply becomes a little larger in pip size.

For some people, 3 pips is classified as a "scalp" trade. Yet, for other people, 30 pips is considered a "scalp" trade, too. The definition of a "scalp" trade, depends entirely upon the largest bar used to encapsulate the trade logic that generates the scalp signal/trigger. So, in order for the research done in this thread to have real world meaning for traders, it needs to first define the types of trades that such discrepancies in data might adversely impact. If I'm a weekly or monthly Swing trader, then I may not care too much about small variances in real-time or historically repainted variances between data points on different platforms, because my trade style might absorb those anomalies.

As all intelligent traders already know, there is no Central Clearing in Retail FX. There is no "Time & Sales" in RTFX. You literally (emphasis on the root: literal) trade on a Platform hosted on a Server, running at the site where your intermediary sets up shop. This is "online trading" in its truest form - literally.

Every RTFX intermediary in the world (banks included) are providing you with access to their proprietary consortium of market liquidity providers. None of them are committed by law and/or regulation to provide their market makers with consistent pricing streams. It is a *free* market, where prices are set by the largest sharks in the tank. Period.

What "forces" credibility and data quality in the pricing streams that hit your platform, are the names and reputations that over the years have become associated with RTFX in positive ways. Therefore, when you trade, using a platform that gives you access to [for example] Currenex, you are getting some of the best "credibility" and "quality" in the streaming data available right now, to the Retail FX segment of the Forex market. Counter clockwise, when you trade using something like serial number 0001-0001A above, you get access to a lack of credibility and a lack of quality in the streaming data.

The absolute best (my favorite) approach for traders that do not yet have millions to open up an account on something like Autobahn, is FXall. You can set-up a commercial or light-weight institutional account and trade on a platform that offers you credibility and quality in the data streamed to your platform. The other route and also a good one, would be BARX FX, provided that you can start somewhere in the $300k - $500k range (talk with them directly, they will work out an arrangement with you, if you are sincere). On that platform, you will be getting access to a much more "serious" version of Interbank, than you ever would on serial number 0001-0001A's platform.

So, if you really want to test this out, then include some Currenex platforms in your analysis, so that you can get a better picture of just how different things truly are out there.

Retail FX, is Retail FX. There's no getting around that fact. You have to learn how to trade really well in FX in general. Those that make it out of Retail, only do so, once they have learned how to trade well. Getting out of Retail FX, is like being awarded a badge of honor - like a graduation ceremony.

You KNOW you are a good trader, once you get out of Retail Jail. It is a prison, but you can break out. Once you do, head for the hills and get on FXall. For now, FXall is the preeminent Middle Layer between full blown Retail and full blown Institutional and they one of the widest and deepest Middle Layer liquidity pools in the industry. You simply set-up a PrimeBroker account and get set-up with FXall.

FXall, is simple the place to be: http://altair.fxall.com/activetraders.html. You can build a sizable business on FXall, without all the legal headaches typically associated with getting onto something like Autobahn, and you will get access to liquidity that is similar - not exactly the same, but darn good enough.

I trade on MT4 as well, for other purposes. So, I'll be following this thread to see how it turns out. I use a bank, as I do not have enough confidence in non-bank intermediaries for FX. I have RTFX accounts for "other" purposes, as well. I've lost complete confidence in any non-bank FX intermediary entity, with only a handful of exceptions and most of them are not in the United States. Most of the U.S. based FX Retail business, has become a sick joke.

If you are going to do FX Retail at all, then you really need to do your homework when it comes to selecting an Intermediary, with whom you can place your confidence. Then graduate as soon as possible out of Retail FX, where things get a LOT more sane and rational.

For those of you with the cash to do so, skip Retail FX entirely - you don't need the headache. Go PrimeBroker and then set-up with FXall or Currenex, or go straight to the Bank - with something like Barclays, BNY Mellon, UBS, HSBC, Standard Chartered, etc., but cut your teeth on lots of demo trading. Ideally, if you have the cash to start here, then you should not be jumping into anything right away as a matter of strict protocol. You probably already have some equity, options, commodity and/or futures experience anyway - but still, do lots of FX demo testing as a sanity check.

Nice thread - I'll keep an eye out. People should participate.

@Golum

Sorry for the late response.
Actually not much.

This is an opportunity for the MT4 community to better its situation but I'm not sure a critical mass of people realizes how it affects them. Or, they simply do not have the capacity to monitor their own feed.

Bottom line, as Jack Shepherd says in LOST - "If we can't live together we're going to die alone."

People need to start monitoring their datafeed and share their results using the tools I have provided.

On my own accounts, I took substantial measures to limit my exposure to the FXCM MT4 feed. Among other things, I monitor (in real time) every thing that moves in it.

Cheers, Marathon

@forexnewbe,
Well, I did talk with them - but only as a reference to the general quality of the feed.
You can only apply for reimbursement if you were IN a trade that was affected by the problem (needless to say, you have to be able to identify the problem in the first place..).
In other words, your only chance of getting reimbursed is if you monitor your feed and able to show there was a problem.

Cheers, Marathon

@WaveTop, thanks for your thorough post.
I tend to agree with practically all of your analysis - except that in my eyes, bogus datafeed (for whatever reason) is one step too far.

Can you share which bank you are using with MT4?

Cheers, Marathon

There was a time when I would agree with this approach, but really IB handles this level of trader quite well. I can't see many/any benefits of going with FXall or BARX over IB. I've been down the Currenex & Hotspot road before and stuck with IB. Spreads, liquidity & execution are largely the same on all platforms. Given that, and the benefits of multi-market global access, no volume negotiations/requirements, etc and IB simply comes out ahead in my book.

Best Marathon,

You are right in your analyse and I thank you. I try to trade on demo with fxpro at the moment. I have almost 3 years trying to develop an ea that is profitable. So far, my backtesting went finally well and I got the typical eureka experience last weeks, but in real time trading went indeed different. Because I use volume as key to my signials I noticed that like you describe volume bars went repainted after some time. I'm agree with you that this is not evil but just criminal. I'm agree with some writers here that maybe bad ticks, differences in data feed between brokers are part of this game but if even the same broker repaint his history all my work on backtesting an ea etc is just a waste of time and from my eureka experience i came in a very bad feeling experiences about all this trading at all. I want to remark that I try to trade oil futures and not the fx. Reason more volume numbers should be somehow decent but they are not...they differe incredible between brokers and not only in value but are at some moments not even correlated. BUT LIKE YOU MENTION REPAINTING THEIRE OWN DATAFEED IS REALLY CRAP....why ?...like I said, what is this here for forum, endless discussions about systems, ea's , backtesting, all that is totally bogus and looks a waste of time for me...I spend enless hours programming and trying to find something that is near profitable (which is a topic on itself) but if you can not backtest, and even your forward test is different because your data is changed while you are watching it, it looks to me just gambling with dishonest casino's. I think then the online casino is a more fair play than this. You are right, the minimum you can expect is that their own datafeed is not repainting , not in ticks and not in volume.
I hope this tread can make it clear to pêople because some people just don't understand what Marathon write here, but I saw it with fxpro and this broker is overwhelmed with good critics. Maybe if I chat with the moderators they give me explanation but I bet they will say that this happens only with demo servers. Like you describe Marathon it happens also on the life server? Anyway a demo server should have same quality like a life , why it is otherwise good to demo anyway.
Ok, this topic is important but also sad and hope I can learn something quick new to help this problem or I must quite the whole idea to trade completely.
Trading like a pro with milions, I'm sorry I thought this could be a start on low budget and if you have a good system money would grow anyway. But if it is all true what i read here then trading looks only good for the millionairs but why trading if you are rich anyway.
Hope someone is reading this thread because the topic put all the other topics in this forum to the waste.
Thanx for reading.

Okay, I do the same thing and I compare the pivot results of several MT4 platforms and if they are within 2 or 3 pips of each other I'm not too alarmed. If the data feeds for ask/bid is withing 2 or 3 pips of each, I'm not that concerned. Anything more I do not trade the pair.

That is an indication that the data feed with one or more of the MT4 platforms is out of sync or the market for that pair is out of sync. I keep watching and comparing. It might not be the day to trade the euraud or eurcad and another day might be best.

But FXCM is not your best choice. Keep trying and doing your comparison with other MT4 provide in the US and abroad and it will come out right.