Best Thread FXCM Discussion

I need to get back to installing some recently designed upgrades to my system. If I catch it again, I'll post a video or some more pics of the problem. It has happened several times since my first post, but I was too business on the development end to bother with it.

When I get the system upgraded, I'll be monitoring everything in real-time.
 
Jason,

With system upgrades completed and the Asian session starting-up with historical volume, I would expect to see spreads that were commensurate with FXCM's marketing campaign. Yet, this is all I see:

2ebbogi.jpg


These are not the holistically sarcastic widened spreads that I reported earlier, but these spreads are somewhat gleefully clinging to the absurdity that somehow by using the FXCM platform, I get spreads that approximate Truth in Pricing. Audi calls it Truth in Engineering.

I would like to see lower spreads closer to planet earth while testing on this platform. Or, at least absent the massive amounts of element called (He).

What say you?
 
You are talking about the difference between Retail and Institutional. My belief is that if any retail FCM gets into the business of Retail FX, then they ought to provide pricing that is commensurate with interbank, or at least Fair & Balanced.

Yeah, they have to keep the spreads wide enough to make a profit from their business model, but the buck needs to stop right there. All of this jerking around on the spreads at moments when the market is either about the move, is completely unacceptable - not to mention rude.

I use a set of density probability calculations to protect me from too many losing trades, but the optimization of my system is lowered by stunts like this being played all the time by FXCM. Striking targets routinely are important to long-term profitability. When these guys start hitting the market with double-digit 20 pip spreads for no good reason whatsoever, right near a TP level that was predicted with a high level of precision - that's just wrong.

I use OANDA and they pull a similar stunt. While both of us realise these guys have to maintain profitability and protect themselves from risk exposure, I can't help feeling they use these news and data events to beef up their take. It's unreasonable to imagine they get quotes from their liquidity providers which enable them to quote us retail players at 1.4 pips at one point and then 15 or 20 the next - I just don't believe there is that much volatility to factor in. And regardless of volatility there are just as many percentage-wise taking each side of any given pair - all the brokers have to do is hedge the delta.
 
I use OANDA and they pull a similar stunt. While both of us realise these guys have to maintain profitability and protect themselves from risk exposure, I can't help feeling they use these news and data events to beef up their take. It's unreasonable to imagine they get quotes from their liquidity providers which enable them to quote us retail players at 1.4 pips at one point and then 15 or 20 the next - I just don't believe there is that much volatility to factor in. And regardless of volatility there are just as many percentage-wise taking each side of any given pair - all the brokers have to do is hedge the delta.

Is it not true that if a ECN model is used you should be able to get the price you want despite volitility?There is always i belive a small amount of slippage but not to the extent that you are quoting 20 pips? this is the company widening the spread and Jason can bleat as much as he likes thats a fact.If you are not happy report them to the is it the FCU now ,not the FSA i,m not sure .They were by the American authorities hit with large fines for i think much the same thing although i could be wrong there however they were caught cheating customers and i belive they still are.Thats my opinion for what its worth. Mike.
 
Here we go again.

I was doing some maintenance on my panel when it happened. This pic pretty much shows the matter quite clearly:

r931oi.jpg


20.1 pips on the Chief. 16.1 on Jeppy. 9.9 on Caddy. 16.1 on Nzed and 11.2 on Audi. The only barely sane participant was the U.S. at this point. Time is shown on the panel in upper right corner and is MT4 chart time for FXCM.

Hi Everyone,

It's important to keep in mind that spreads are a factor of market liquidity. Just as spreads widen in equities and futures markets during illiquid periods, spreads can widen in the forex market. The most typical times for this to happen are when trading opens for the new week on Sunday, around economic announcements, and during trade rollover.

The reason why you see spreads widen on our platform around 21:00 GMT is because that coincides with 5pm New York time, which marks the end of one 24-hour trading day and the start of the next. During that time the largest banks in the world including our liquidity providers widen their spreads as part of a process called trade rollover. Trades are "rolled over" from one trading day to the next, and rollover interest is applied to these positions.

You must realise when you use FXCM or similar, you are not trading the underlying market, but a synthetic derivative of that market as determined beneficial to the company providing that derivative.

Is it not true that if a ECN model is used you should be able to get the price you want despite volitility?There is always i belive a small amount of slippage but not to the extent that you are quoting 20 pips? this is the company widening the spread and Jason can bleat as much as he likes thats a fact.

Trade rollover is an illiquid time in the market as banks refresh their servers for the new trading day. Some brokers stop quoting prices altogether for a few minutes before and few minutes after rollover. However, FXCM does continue to quote prices, and it's precisely because our prices reflect the quotes we receive from our liquidity providers that you will see our spreads widen when theirs do.

The spreads widened on FXCM just before that spike seen here (Nzed chart):

290tuzs.jpg


This is pretty typical on JPY for FXCM. Spreads seem to take off into no-mans land just before a move of some sort which typically positions me out of a decent trade entry. It happens too many times to be considered acceptable.

BTW - that blue dashed line that you see in the chart - yeah, that one. That one with the FXCM spike straight through the heart. Well, that was precisely the Target that was calculated by my system a couple of hours ago from the top of the move.

That means that when FXCM widened the spread, it moved the Ask price higher during that same move. That means that my Short position never reached the Target, when there should have been room to clear the exit on the move.

It wasn't that case that you missed out on an opportunity to take profit because of the spreads widening. Rather it was the widening of the spread during trade rollover which created the illusion that your take profit level had been reached. I hope the tick chart below (which you can also view on your Trading Station) can provide you with clarification. It shows both the bid price (thin blue line) and the ask price (thin purple line) for the time in question. I've plotted your take profit order to buy at 81.027 (blue dotted line). As you can see, the NZD/JPY spread began to widen at 20:58:00 GMT (red dotted line) because of trade rollover. The spreads returned to normal at 21:03:59 (green dotted line).

ah70.png

In between those times, the midpoint between the bid price and the ask price never went below 81.115 even as the spread widened. As you know, it's the ask price which determines when you take profit order on a short position can be executed. So while it may have seemed like you missed out on an opportunity to exit, when looking at your chart of the bid price in isolation, in actuality the market never went down to your take profit level. Furthermore, the only reason the bid price touched 81.027 was because spreads widened.

Please let me know if you have further questions.

Thanks,
Jason
 
Last edited:
...It's unreasonable to imagine they get quotes from their liquidity providers which enable them to quote us retail players at 1.4 pips at one point and then 15 or 20 the next - I just don't believe there is that much volatility to factor in.

And, that really is the point I'm trying to make with Jason. I'm trying get them to explain why we should take such a huge leap of faith and start believing that such high levels of volatility exist in any business that transacts nearly 5 Trillion per day. The very nature of that kind of cashflow should denote a high degree of stability in volume - relative to the equities markets and the like.

I've read the Triennial Bank of International Settlements Report. When you read those reports every three years, it becomes crystal clear the FX is growing, not shrinking in volumes of all kinds. These FCM's are trying to convince people that somehow, FX volume behavior is the same or similar to equity market volume behavior and nothing (and I mean nothing) could be further from the truth of the matter - especially when you are talking about the USD, EUR and the JPY.

There is tons of cash in those denominations flying all over the world 24/7/365. Telling me that those base/counter pairings somehow suffer such a drop in liquidity to warrant up to a 1000% or more drop in market participation, is like trying to sell me the Brooklyn Bridge with an eBay Merchant Account.


And regardless of volatility there are just as many percentage-wise taking each side of any given pair - all the brokers have to do is hedge the delta.

There is certainly more than one way for them to skin their cat and keep the slot machine numbered to their favor. I'm sure they take advantage of many such structural "coincidences." ;)

They should bring the retail spreads to within at least a reasonable range of sanity. They've been getting away with this for years.
 
Is it not true that if a ECN model is used you should be able to get the price you want despite volitility?

Not necessarily. These retail FCMs are taking their liquidity pool prices and running them through a so-called "pricing engine" before they get to your desktop via TS-II and/or MT4. That basically means that the prices you trade through outfits like FXCM are being manipulated through an algorithm designed to keep their profits in sink with their revenue requirements, based on what their "pool" publishes to the stream. They are basically Retailing a Wholesale price out to the general public.

I'm not against them making a profit - that is of course, the nature of doing any business. All businesses have costs and FXCM is no different - they have to pay their own bills. Office space in New York near Wall Street can't be cheap. My problem is the extreme to which these guys go to manipulate the price feed.

No "real" institution trading on Interbank executes anything at 2+ pips per transaction. Those guys are trading inside the pip in most cases and trading Flat Exchanges (no spread transactions) is not uncommon. Most people here are far from that level of institutional account status, but that should not translate into being ripped a new galactic hole the size of 20+ pips.

Supposed you clicked on a trade just at the point where they jerked the spread up to 20 pips? You are now the proud owner of an Albatross and you need 20 pips just to reach BE.

There is always i belive a small amount of slippage but not to the extent that you are quoting 20 pips?

The spreads seen on my system are taken directly from an API connection through Excel to MT4's price feed and that comes directly from FXCM. So, when my system tells me that a pair drew a 20.1 pip spread, you can rest assured that it came directly from the trading platform.


...this is the company widening the spread and Jason can bleat as much as he likes thats a fact.If you are not happy report them to the is it the FCU now ,not the FSA i,m not sure .They were by the American authorities hit with large fines for i think much the same thing although i could be wrong there however they were caught cheating customers and i belive they still are.Thats my opinion for what its worth. Mike.

It is not just the spread widening that FXCM does on a far too routine basis, but it is the Maximally Poor Fills that I routinely see across the board. They always seem to fill me at the worst end of the extreme spread range during the time it takes for my transaction request to leave my machine and hit their back-end. I realize that there is a delay caused by the routing on the network, but to never get filled at the long end of the stick at least sometimes, is a pretty good indicator of what's going on.

See 3.2. Filled at 3.5. See 3.5. Filled at 3.8. See 3.8. Filled at 4.1. This kind of things goes on all the time on TS-II in the vast majority of cases. It is rare that I see 3.2 and get filled at 3.0, or 2.8. Network latency is one thing - but that should cancel out as the aggregate of all fills over time should balance out to include those that do get filled at better than seen on the platform. That very rarely happens on my FXCM platform - if at all.
 
Trades are "rolled over" from one trading day to the next, and rollover interest is applied to these positions.

I'm not talking about just interbank swap. And, by the way - this did not use to be a problem in years past. You did not use to see gaps in bars of data on the M1 chart during this time of day and the spreads did not widen to absurd levels either. Something happened a few years ago and this is now common place. So, it does not have to be this way.

I get the swap, but I'm also referring to periods where the spreads seem to intentionally widen just before a significant move in the market. This happened plenty of times on the FXCM platform and just because I did not video my desktop each time, does not mean that others have not noticed the exact same thing. But, that's not the most interesting part. In my mind, the most interesting part is the direction of the widening relative to the move that follows. Typically, if the move is to the upside, I notice the Ask move to some absurd level just before the move, where the Bid pretty much remains where it was. This means that if you wanted to get Long before the move, your net effective cost is significantly higher. The same seems to hold true for significant moves to the downside where the Bid stretches downward and the Ask remains steady, forcing a net/net higher cost of entry when the spreads snap-back into a more normative range.

Like I said before, I've been doing a lot of development work on my newest trading system and I have not been recording my desktop or paying that much attention in a while. I've leaving the panel alone for a while so I can dial-in its usage protocols. That should give me more time to pay attention to the spreads. I'll post what I can when I can.

In the meantime, I would appreciate an explanation for why a 1000% widening of the spread makes any sense, even during the interbank swap periods.

Things look fairly stable for now, but I would say that even these spreads are too high in general terms. This is the JPY basket of majors for goodness sakes. There are tons of transactions going on out there around the world and around the clock on these pairs. Abusing the volatility and swap excuse is just not acceptable with a highly transacted "coin" like the JPY. Everybody is using the Yen - all over the place. Especially, when paired up with the rest of the majors.

This cannot be a liquidity problem - please stop selling me liquidity issues at FXCM. We are talking about the Yen, crossed with USD, GBP, EUR, CHF, AUD, NZD and CAD. Those are crossed made in heaven as far as liquidity is concerned.

We are not talking about NOKJPY, for crying out loud, Jason. Nor, are we talking about ancient Roman Quadrans -vs- the ancient Persian Siglos.

349b76e.jpg
 
Jason (FXCM) I'll take the lack of response on the other thread as a nod.

Back to the business in hand here though - what would FXCM have to do with its spreads to offer a genuinely fixed spread service to those who wished to have it?

In place of hitting your clients with massively widened spreads at key news times, what would you have to increase your spreads to at all times - in terms of percentage increase - across all pairs - to offset that risk?

Seriously, if you have a 2 pip spread on aaa/bbb at normal times and average number of trades during normal times is N, there has to be a relatively simple calculation that you can apply that says "If we made the spread 2 + X, we could offer a fixed spread at all times and cover our exposure on increased spread from our liquidity supplier at volatile times on the net or delta positions".

If any broker were to offer a two tier system and it was reasonable - I'd jump ship and join them in an instant.

Trading, even for retail players, is a business. In business you can often choose to go for variable cost or fixed costs for many services and supplies. Fixed costs make planning a lot easier. It's also normally more expensive nominally, but cheaper in the long run because you don't get any nasty surprises that require short term gap financing.

Even with something as simple as electricity supply we get to choose a fixed rate for 2 years or 5 years at a higher rate than the variable - but it locks costs in to a known amount.

Why can't FXCM lead the way and offer a facility such as this?
 
Isn't that what Prospreads effectively offers?
I'm none too keen to get on board a company which was a Gibraltar based subsidiary of a London holding group but is now in the process of being sold to a privately owned company in Gibraltar. Location and regulation apart, change of ownership and more importantly, class of ownership it always a risk for any company.

I am also concerned with their operating P&L and financial bona fides.

http://www.stockmarketwire.com/article/4667876/ProSpreads-to-be-sold-by-London-Capital-Group.html
 
I've read the Triennial Bank of International Settlements Report. When you read those reports every three years, it becomes crystal clear the FX is growing, not shrinking in volumes of all kinds. These FCM's are trying to convince people that somehow, FX volume behavior is the same or similar to equity market volume behavior and nothing (and I mean nothing) could be further from the truth of the matter - especially when you are talking about the USD, EUR and the JPY.

There is tons of cash in those denominations flying all over the world 24/7/365. Telling me that those base/counter pairings somehow suffer such a drop in liquidity to warrant up to a 1000% or more drop in market participation, is like trying to sell me the Brooklyn Bridge with an eBay Merchant Account.

You're right when you say the forex market is growing. While it's true that the average daily volume of over $5 trillion dwarfs that of any other financial market, that does not mean this trading activity is evenly distributed throughout the day. The table below shows data from the BIS survey you referenced.

h1p3.jpg

The UK makes up 41% of the daily volume, followed by the US with 19%. Singapore has a 5.7% share, followed by Japan’s 5.6% and Hong Kong’s 4.1%. That's indicative of how much more trading is done during the European session than either the US or Asian sessions. Also, trade roll over occurs during the gap between the US and Asian sessions.

Banks specifically chose this time to perform trade rollover because of the low volume. It's a fact that banks will widen their spreads at this time. Furthermore, even at other times, banks can widen spreads because of market conditions. Each bank decides what prices they want to quote at any given time. The advantage of our No Dealing Desk execution is that banks have to compete with each other to get the order volume from our clients. Only the liquidity providers quoting the highest bid or the lowest ask price will get the order flow.

It is not just the spread widening that FXCM does on a far too routine basis, but it is the Maximally Poor Fills that I routinely see across the board. They always seem to fill me at the worst end of the extreme spread range during the time it takes for my transaction request to leave my machine and hit their back-end. I realize that there is a delay caused by the routing on the network, but to never get filled at the long end of the stick at least sometimes, is a pretty good indicator of what's going on.

You don't seem to be taking into account how liquidity works. For your market order to be filled at a certain price, there has to be a liquidity provider willing to take the other side at that price. If the market price changes before your order can be filled, then your order will be filled at the next available price. This is called slippage.

Slippage is one of the risks you assume when trading. The risk of slippage is greatest during volatile markets such as around a news announcement like NFP, or when holding trades open over the weekend. Bringing up NFP is a great example. Take a look at the below chart which shows the GBP/USD tick chart right at 8:30am ET during the March 8, 2013 NFP release.

gbpusd44201332723pm.png

GBP/USD dropped from 1.50361 to 1.49972 in one tick. Yes one tick! That means there were no quotes quoted by the liquidity providers in between those prices. If you had an order to sell GBP/USD at 1.5020, your order would not have been filled at 1.5020 because there was no quote at that price from a liquidity provider offering to buy when you wanted to sell. You likely would have been filled at the next available price around 1.49972. FXCM isn't deciding to fill or not fill at certain prices, your orders are filled based on liquidity and where there is liquidity on the other side of the transaction to fill your order.

Also, keep in mind that slippage works both ways. If you have a limit or take profit order, and the price gaps in favor of your trade, then you will get filled at the more favorable price which is trading in the market. This is called positive slippage or price improvement. The stats below that FXCM clients receive price improvements on limit orders just as frequently as they receive negative slippage on stop orders. In fact, we even provide tips on our website to help clients minimize negative slippage and maximize price improvements.

fxcmpriceimprovementspo.png

I get the swap, but I'm also referring to periods where the spreads seem to intentionally widen just before a significant move in the market. This happened plenty of times on the FXCM platform and just because I did not video my desktop each time, does not mean that others have not noticed the exact same thing. But, that's not the most interesting part. In my mind, the most interesting part is the direction of the widening relative to the move that follows. Typically, if the move is to the upside, I notice the Ask move to some absurd level just before the move, where the Bid pretty much remains where it was. This means that if you wanted to get Long before the move, your net effective cost is significantly higher. The same seems to hold true for significant moves to the downside where the Bid stretches downward and the Ask remains steady, forcing a net/net higher cost of entry when the spreads snap-back into a more normative range.

What you've described only proves that banks have to manage their risk just like any other market participant. They can't predict the future, but if there is a pending news announcement that has the potential to cause prices to gap higher, they don't want to caught on the wrong side of it by setting their ask prices too low. In such situations all the liquidity providers would raise their ask prices as a precaution. As an FXCM client you would still benefit from the NDD model because the buy price you see would reflect the lowest ask price being quoted from our liquidity providers plus our fixed pip markup.
 
Last edited:
what would FXCM have to do with its spreads to offer a genuinely fixed spread service to those who wished to have it?

In place of hitting your clients with massively widened spreads at key news times, what would you have to increase your spreads to at all times - in terms of percentage increase - across all pairs - to offset that risk?

Seriously, if you have a 2 pip spread on aaa/bbb at normal times and average number of trades during normal times is N, there has to be a relatively simple calculation that you can apply that says "If we made the spread 2 + X, we could offer a fixed spread at all times and cover our exposure on increased spread from our liquidity supplier at volatile times on the net or delta positions".

If any broker were to offer a two tier system and it was reasonable - I'd jump ship and join them in an instant.

Trading, even for retail players, is a business. In business you can often choose to go for variable cost or fixed costs for many services and supplies. Fixed costs make planning a lot easier. It's also normally more expensive nominally, but cheaper in the long run because you don't get any nasty surprises that require short term gap financing.

Even with something as simple as electricity supply we get to choose a fixed rate for 2 years or 5 years at a higher rate than the variable - but it locks costs in to a known amount.

Why can't FXCM lead the way and offer a facility such as this?

Actually, what you're describing is a fixed-spread model which was fairly common in the past and is still used by some dealing desk brokers. The reason so few fixed-spread brokers have lasted is because of the challenges of this business model. Banks providing liquidity to brokers widen their spreads as market conditions dictate. If there's a major news announcement pending and all banks have all widened their spreads to 10 pips, but the broker has promised a fixed spread of 3 pips, then the broker's dealing desk has to take on all that market risk for the orders their clients execute that cannot be offset immediately.

If after the news announcement the market ends up moving against their clients, the dealing desk can offset these orders for a profit, but if the market ends up moving in favor of their clients the dealing desk could end up losing a lot, even more than they made by charging a premium on spreads during less volatile times. This is when a dealing desk broker might resort to dealer intervention practices such as re-quoting orders, delaying execution, skewing prices, or widening spreads. The fixed-spread model is very difficult to maintain, some fixed-spread brokers have gone bankrupt trying this model.

I feel that FXCM's size and financial strength and the diminishing number of fixed-spread dealing desk brokers is a reflection of the popularity of our variable spread No Dealing Desk model. There are no requotes on our platform, and the spreads reflect the best prices we receive from 10+ liquidity providers plus our fixed pip markup. FXCM's deep and diverse pool of liquidity providers helps to ensure that NDD prices are reliable - not set by a single provider - and that they do reflect the broader forex market. Also, through competition, the NDD model ensures that prices are market-driven and fair. FXCM rewards liquidity providers with order flow when they provide the best bid or ask prices. The more advantageous their prices are the more order flow the liquidity provider will receive. The less advantageous, the less order flow they will receive. This efficient selection method keeps a single liquidity provider from adversely affecting your price.
 
Attention CFD traders

HKG33 trading will be closed on October 1st for Hong Kong's National Day and on the 14th for the Chung Yeung Festival.

chung-yeung-festival2.jpg
 
Actually, what you're describing is a fixed-spread model which was fairly common in the past and is still used by some dealing desk brokers. The reason so few fixed-spread brokers have lasted is because of the challenges of this business model. Banks providing liquidity to brokers widen their spreads as market conditions dictate. If there's a major news announcement pending and all banks have all widened their spreads to 10 pips, but the broker has promised a fixed spread of 3 pips, then the broker's dealing desk has to take on all that market risk for the orders their clients execute that cannot be offset immediately.

If after the news announcement the market ends up moving against their clients, the dealing desk can offset these orders for a profit, but if the market ends up moving in favor of their clients the dealing desk could end up losing a lot, even more than they made by charging a premium on spreads during less volatile times. This is when a dealing desk broker might resort to dealer intervention practices such as re-quoting orders, delaying execution, skewing prices, or widening spreads. The fixed-spread model is very difficult to maintain, some fixed-spread brokers have gone bankrupt trying this model.

I feel that FXCM's size and financial strength and the diminishing number of fixed-spread dealing desk brokers is a reflection of the popularity of our variable spread No Dealing Desk model. There are no requotes on our platform, and the spreads reflect the best prices we receive from 10+ liquidity providers plus our fixed pip markup. FXCM's deep and diverse pool of liquidity providers helps to ensure that NDD prices are reliable - not set by a single provider - and that they do reflect the broader forex market. Also, through competition, the NDD model ensures that prices are market-driven and fair. FXCM rewards liquidity providers with order flow when they provide the best bid or ask prices. The more advantageous their prices are the more order flow the liquidity provider will receive. The less advantageous, the less order flow they will receive. This efficient selection method keeps a single liquidity provider from adversely affecting your price.
Thanks for your response Jason. I can see why variable spreads are more popular with brokers and as a client I wouldn't be too pleased with re-quotes/slippage if that was the only protection the brokers had against providing fixed spreads.

I just assumed it was a simple modelling exercise to work out what size of fixed spread would offer the same overall protection against risk for the broker.

e,g, eur/usd 5 pips at all times versus 1.4 most of the time and 20 for 90 seconds every other day. Obviously not quite as simple as that in reality.

Is the volatility at these times a function of expectation of transaction volumes or in response to it?

Do you at FXCM see your liquidity providers widen their spreads ahead of increased action or is there increased action on the back of the spread play?
 
The UK makes up 41% of the daily volume, followed by the US with 19%. Singapore has a 5.7% share, followed by Japan’s 5.6% and Hong Kong’s 4.1%. That's indicative of how much more trading is done during the European session than either the US or Asian sessions...

No - it is not. You are equating Location with Volume and that is not correct. Not by a long shot. It misses the entire point. The BIS report does not conclude anything based on a longitude/latitude transactional model of some sort. That's a very simplistic look at what the graph actually means. It is a Currency Type specific graph that you just posted, not a Logistical implication of where the transactions are actually taking place, who the counter-party to each transaction happens to be and in what manner each transaction is being initiated. Again, you can't look at that graph and conclude anything specific about Retail Forex transactions -vs- Time of Day they occurred. I'm very surprised that you did not understand that (shocked).

Banks specifically chose this time to perform trade rollover because of the low volume. It's a fact that banks will widen their spreads at this time. Furthermore, even at other times, banks can widen spreads because of market conditions.


Wait a minute, slow down. You are conflating facts. My conclusion was that based on my observations of the FXCM platform, I was noticing plenty of times where spreads were widened just before a significant move and that such widening did not always take place during the daily swap period. That is a distinction with a definite difference and not the other way around.

I just posted yet another pic showing that outside of the swap period, the FXCM spreads were holistically silly in terms of their wideness. If you check this intermediaries FX prices Monday through Friday (not weekends), you will not a major difference in the spreads as compared to FXCM. Sure, they charge a commission - but the commission as compared to the actual Pip Cost derived from the higher spreads that FXCM "charges," proves to be a rather small concession.

Real Time Forex Spreads.



Only the liquidity providers quoting the highest bid or the lowest ask price will get the order flow.

And, I'm 100% certain that each bank in the FXCM liquidity pool knows absolutely nothing about what each other is quoting - right? :rolleyes:


You don't seem to be taking into account how liquidity works. For your market order to be filled at a certain price, there has to be a liquidity provider willing to take the other side at that price. If the market price changes before your order can be filled, then your order will be filled at the next available price. This is called slippage.

So, I don't know how this business works? You describe Encoded Slippage to me and somehow, I'm the one not getting it? :rolleyes: Did you not read my post. I specifically said that I never get the off-set coming the other way. It never is the case that my filled level is an improvement over the quoted price on my screen. I specifically said that IF there was a true fairness in the price stream, then the instances of both Negative and Positive Slippage would return straight back to the mean of Total Slippage.

I would not continuously see Negative Slippage in the aggregate to a level that dwarfs Positive Slippage in the aggregate. Unless you are prepared to tell me that Negative Slippage on the FXCM platform has left sided skew coupled to positive kurtosis, then a return to the sample mean should be exactly what Total Slippage does in the long run.



Slippage is one of the risks you assume when trading.

Positive Slippage is also one of the risks that you assume equally. You can't have it both ways. If the pricing is fair, than slippage should moments of both left and right sided skew and it should establish a mode that is readily identifiable through statistical analysis. There cannot possibly be a case where Negative Slippage sustains perpetual positive kurtosis in a fairly priced market. Either you are playing me for a fool, or you don't understand your own business half as well as you think you do.

As positive kurtosis in the sample set increases it would do so to an historical peak that should be readily identifiable. That peak would be part of a sub-set of historical Bid/Ask spread data that would likewise have a density probability associated with it. You should be able to calculate the probability for both the Location of increased spreads about its Mean, as well as its Frequency - or Rate of Occurrence over a specified time frame.

This is not rocket science and the data simply does not lie. I'm seeing not only highly skewed instances of negative slippage, but sustained highly skewed instances of negative slippage and that my dear friend is supposed to be statistically impossible in a fairly/freely priced market. Stats 101.


FXCM isn't deciding to fill or not fill at certain prices, your orders are filled based on liquidity and where there is liquidity on the other side of the transaction to fill your order.

FXCM controls both the Bid and the Ask price that are seen on my platform! FXCM does this through a computed algorithm that determines what the spread should be based on its already determined per transaction revenue model. FXCM is not charging a "stated commission" but the inverse is exactly the same net/net result. FXCM makes its money by artificially widening the spread and pocketing the differential - Wholesale to Retail Pricing.

I'm not blaming FXCM for doing this - that's not the point. You keep missing the point. The point is the resultant absurdity of the spreads and the perpetually left sided skew coupled with endless positive kurtosis in the Total Slippage. That's my point.


The stats below that FXCM clients receive price improvements on limit orders just as frequently as they receive negative slippage on stop orders.

Those states for Positive Slippage don't come anywhere near my stats for Positive Slippage on the FXCM platform. Not even remotely that close. I'm seeing virtually 99% negative slippage and that is not supposed to be possible in a fairly/freely priced market.


As an FXCM client you would still benefit from the NDD model because the buy price you see would reflect the lowest ask price being quoted from our liquidity providers plus our fixed pip markup.

If they are seeing what each other is quoting (and they do) then much of that NDD gain is negated by the fact that they can "fix" prices within a range and never have to speak a single word to each other to do it.

To prove all of this - why not simply switch the FXCM model to a pure Commission basis and then let the providers fight it out over who gets the Retail business by truly keeping their spreads somewhere closer to the surface of the earth, as opposed to the other-worldly type of spread scenarios I encounter far too often on the FXCM platform. Encode a true Depth of Market window into the platform (not the fake one used for marketing hype on the HF platform that FXCM offers). This would give Retail traders an opportunity to trade inside the pip on far more occasions. If a couple banks decided to widen their spreads to epic levels, then the Retail market (through a true DOM) could simply decide to go where the pricing reflects sanity. Or, encode a Spread Monitor window into the platform that showed the actual spreads of each provider (even better). That would take care of DOM. In fact, it would create a significant portion of the DOM.

Sit back. Collect the reasonably priced commission and let the Free Markets take care of themselves. Stop manipulating the thing to death. Show me ALL the spreads. I will go where I see the best spreads and the level of liquidity that meets my trading needs. That will do more to create competition among your providers than anything else. After all - you implied that creating competition among your providers is something that FXCM is concerned about for its customers sake. Well, prove it. Go commission basis, expose the spreads and the DOM and let the chips fall where they might.

Still think I don't know anything about this business? ;)
 
No - it is not. You are equating Location with Volume and that is not correct. Not by a long shot. It misses the entire point. The BIS report does not conclude anything based on a longitude/latitude transactional model of some sort. That's a very simplistic look at what the graph actually means. It is a Currency Type specific graph that you just posted, not a Logistical implication of where the transactions are actually taking place, who the counter-party to each transaction happens to be and in what manner each transaction is being initiated. Again, you can't look at that graph and conclude anything specific about Retail Forex transactions -vs- Time of Day they occurred. I'm very surprised that you did not understand that (shocked).




Wait a minute, slow down. You are conflating facts. My conclusion was that based on my observations of the FXCM platform, I was noticing plenty of times where spreads were widened just before a significant move and that such widening did not always take place during the daily swap period. That is a distinction with a definite difference and not the other way around.

I just posted yet another pic showing that outside of the swap period, the FXCM spreads were holistically silly in terms of their wideness. If you check this intermediaries FX prices Monday through Friday (not weekends), you will not a major difference in the spreads as compared to FXCM. Sure, they charge a commission - but the commission as compared to the actual Pip Cost derived from the higher spreads that FXCM "charges," proves to be a rather small concession.

Real Time Forex Spreads.





And, I'm 100% certain that each bank in the FXCM liquidity pool knows absolutely nothing about what each other is quoting - right? :rolleyes:




So, I don't know how this business works? You describe Encoded Slippage to me and somehow, I'm the one not getting it? :rolleyes: Did you not read my post. I specifically said that I never get the off-set coming the other way. It never is the case that my filled level is an improvement over the quoted price on my screen. I specifically said that IF there was a true fairness in the price stream, then the instances of both Negative and Positive Slippage would return straight back to the mean of Total Slippage.

I would not continuously see Negative Slippage in the aggregate to a level that dwarfs Positive Slippage in the aggregate. Unless you are prepared to tell me that Negative Slippage on the FXCM platform has left sided skew coupled to positive kurtosis, then a return to the sample mean should be exactly what Total Slippage does in the long run.





Positive Slippage is also one of the risks that you assume equally. You can't have it both ways. If the pricing is fair, than slippage should moments of both left and right sided skew and it should establish a mode that is readily identifiable through statistical analysis. There cannot possibly be a case where Negative Slippage sustains perpetual positive kurtosis in a fairly priced market. Either you are playing me for a fool, or you don't understand your own business half as well as you think you do.

As positive kurtosis in the sample set increases it would do so to an historical peak that should be readily identifiable. That peak would be part of a sub-set of historical Bid/Ask spread data that would likewise have a density probability associated with it. You should be able to calculate the probability for both the Location of increased spreads about its Mean, as well as its Frequency - or Rate of Occurrence over a specified time frame.

This is not rocket science and the data simply does not lie. I'm seeing not only highly skewed instances of negative slippage, but sustained highly skewed instances of negative slippage and that my dear friend is supposed to be statistically impossible in a fairly/freely priced market. Stats 101.




FXCM controls both the Bid and the Ask price that are seen on my platform! FXCM does this through a computed algorithm that determines what the spread should be based on its already determined per transaction revenue model. FXCM is not charging a "stated commission" but the inverse is exactly the same net/net result. FXCM makes its money by artificially widening the spread and pocketing the differential - Wholesale to Retail Pricing.

I'm not blaming FXCM for doing this - that's not the point. You keep missing the point. The point is the resultant absurdity of the spreads and the perpetually left sided skew coupled with endless positive kurtosis in the Total Slippage. That's my point.




Those states for Positive Slippage don't come anywhere near my stats for Positive Slippage on the FXCM platform. Not even remotely that close. I'm seeing virtually 99% negative slippage and that is not supposed to be possible in a fairly/freely priced market.




If they are seeing what each other is quoting (and they do) then much of that NDD gain is negated by the fact that they can "fix" prices within a range and never have to speak a single word to each other to do it.

To prove all of this - why not simply switch the FXCM model to a pure Commission basis and then let the providers fight it out over who gets the Retail business by truly keeping their spreads somewhere closer to the surface of the earth, as opposed to the other-worldly type of spread scenarios I encounter far too often on the FXCM platform. Encode a true Depth of Market window into the platform (not the fake one used for marketing hype on the HF platform that FXCM offers). This would give Retail traders an opportunity to trade inside the pip on far more occasions. If a couple banks decided to widen their spreads to epic levels, then the Retail market (through a true DOM) could simply decide to go where the pricing reflects sanity. Or, encode a Spread Monitor window into the platform that showed the actual spreads of each provider (even better). That would take care of DOM. In fact, it would create a significant portion of the DOM.

Sit back. Collect the reasonably priced commission and let the Free Markets take care of themselves. Stop manipulating the thing to death. Show me ALL the spreads. I will go where I see the best spreads and the level of liquidity that meets my trading needs. That will do more to create competition among your providers than anything else. After all - you implied that creating competition among your providers is something that FXCM is concerned about for its customers sake. Well, prove it. Go commission basis, expose the spreads and the DOM and let the chips fall where they might.

Still think I don't know anything about this business? ;)

Good post, but I think you may be wasting your breath or typing fingers if you expect a reasoned reply from FXCM.:)
 
Thanks for your response Jason. I can see why variable spreads are more popular with brokers and as a client I wouldn't be too pleased with re-quotes/slippage if that was the only protection the brokers had against providing fixed spreads.

I just assumed it was a simple modelling exercise to work out what size of fixed spread would offer the same overall protection against risk for the broker.

e,g, eur/usd 5 pips at all times versus 1.4 most of the time and 20 for 90 seconds every other day. Obviously not quite as simple as that in reality.

Hi PB,

Not simple indeed. The problem for this hypothetical fixed-spread broker would be arbitrage. During normal market conditions, hardly anyone would trade on their 5-pip spreads, because FXCM's variable spreads would be tighter. Traders would only opt for the fixed-spread broker's 5-pip spread during market conditions that caused FXCM's variable spreads to widen beyond 5 pips. However, on our NDD model, that would only happen if our 10+ liquidity providers widened their spreads due to a news event, trade rollover or some other period of illiquidity. Given such market conditions, it would be difficult for the fixed-spread broker to offset client orders without taking a loss, even with a 5-pip spread, unless they requote their clients.

Is the volatility at these times a function of expectation of transaction volumes or in response to it?

Both. It's a chicken or egg scenario. Traders expect the market to move based on a key news event, so they plan to place trades based on the news. Since traders plan to place trades based on the news, this causes the market to move during a key news event.

Do you at FXCM see your liquidity providers widen their spreads ahead of increased action or is there increased action on the back of the spread play?

Liquidity providers prepare for times of potential volatility or low liquidity by widening spreads in advance. In this way, they are in essentially asking for more compensation to account for the additional risk they take in offsetting orders at these times. That's because even banks find it harder to offset their risk with other banks when spreads are wide and liquidity dries up.

Jason
 
Top