FX Thoughts

Every now and then something either in my trading, or a comment from another trader compels me to write down my thoughts on a particular trading subject. Much of the recent impetus for this has come from spirited online discussions hosted by OANDA, where I go by the handle ~chaffcombe.

On re-reading some of these submissions it is very easy to wish that I had put a point a different way, or to be overcome with desire to correct the style or grammar. However, what was writ, was wrote. Online discussions have always been spontaneous and informal; so I'll just leave things as they were.

I'll kick this off with just a small sampling of discussions that have come up over the last couple of years and will retrieve some more in the near future. I encourage those reading to take a look at the full discussions for a more complete view than just my own, and even to contribute themselves.

Absolute Beginners

Since starting this website I've received a number of emails from newcomers that seem to be struggling with taking the first steps. Anyway, the following is a reply I gave to one person. Hopefully it will also be of help to others.

(BTW the reason I particularly recommend OANDA to newcomers is because of their incredibly low minimum trade size (1 unit) ). The fact that I also like OANDA's trade platform, their API, and them as a company, is incidental.

I'm purely a trader; not a product salesman or an educator - so as such there's not a lot I can do for you until you have more specific questions arising from your own trading experience.

What I'd suggest right now is that you open an FXGame account at www.oanda.com and start learning the mechanics of trading fx. Read everything on the OANDA site and forums, going back as far as you can. When you feel you have some understanding of the process open a small FXTrade account (say $1000) in parallel , and see what trading real money is like. Keep doing the above for as long as it takes!

While you are doing that, systematically analyse every idea you can think off. Trade what works; just analyse what doesn't. You'll find that what works varies over time.

Don't expect easy answers. I've been a trader for over 30 years (with some time off for good behaviour), and I'm still doing 12 hour days when I have new ideas (right now it's the api). If you haven't a passion for this you will not succeed.

API

This topic is a bit more specialised than the others in this section, and will only be of interest to a few.

OANDA has a api (application program interface) that enables direct computer communications between client computers and their servers. Earlier this month (July 2004) I started developing my own systems to use this interface, and recorded my first impressions on a private forum.

The following is an extract:

.........yes, I’ve just taken on board OANDA’s api, and am doing the programming (in Java) myself. xxxxxx is also interested in this project, and we have discussed this quite a bit privately.

From my perspective am extremely happy with it. However it is not for the faint hearted. You may have seen posts of mine where I’ve said I prefer Wealth-lab over TradeStation, because you have more control. Same thing here; what you get is an application library of system calls (class objects actually), that enable you to interface with your account on their servers. You only get the building blocks (plus an example program), and the rest is up to you. If you wanted too you could write the thing we know as the FXTrade platform with the library as it covers all the low level functionality, providing you put it together correctly – ie login, getting sub-accounts, getting balances, getting rates, getting recent price history, etc, etc. It also has all the calls necessary to place and manage orders and trades. The only thing the api doesn't include is a call to your formal account history.

As I've said, I’m doing all my own development. Although I'm new to Java; I've done a lot of programming, and I really enjoy it. Now I'm through the first week's learning curve, I'm finding it very straightforward. Quite honestly, I would not feel comfortable running an api such as this, if I hadn't written the code myself. Get it wrong, and it could clean your account!

So far my biggest effort has been to make the api programs bullet proof. I'm writing them so they can run unmanaged all week 24/7. That means they have to handle communication and session failures, and restart in a logical fashion, without intervention (and do the job!). I think I'm there now, but I'm still testing. Remember I have a lot of sub accounts, and that is also an issue. Note also, I said 'api programs'. The api supports multiple concurrent sessions! By that I mean you can write multiple api's, and log them all in at once! How good is that?

I’m not sure how to answer your question – how you interface with FXGame … The key to success is knowing WHY you want the api in the first place. If you have a WHY, you should then know immediately by just your familiarity with FXTrade/Game whether what you want to do is possible. If you can do it manually in FXTrade, you can do it with the api – just far more frequently and far more accurately.

Do you have to develop a GUI front-end? Not really. The thing you are probably building is an execution engine, and/or a data collector. Certainly the former and quite possibly the latter are perfectly suited to console operations. Earlier I said I want my programs to run 24/7 unmanaged. You don’t need a GUI front-end for that. Later perhaps, but very low priority. Also you don't need to write anything that FXTrade already does, as you can run the platform in parallel.

Can trades be triggered by time? Sure: Java is a threaded application and you can control time to a millisecond! It’s pretty much straight forward programming.

I’d just add that was surprised at my bandwidth usage. My Internet account used to use about 1GB per month. Now I’m using about 250 MB a day and I’ve only just started. Beware if you plan to move to the sticks!

In closing I’d say that I’m extremely excited about the whole thing, but you MUST know why you want it, before you even consider it.

Hope the above helps. If you can ask specific questions, I’ll try to answer them.

Malcolm

For those who want more information on OANDA's api:

Asset Allocation

If you've looked at my website you'll get an idea of the approach I use, which is to basically divide my capital up amongst various ideas/methods/pairs etc. Then I apply sound money management methods against each individually. (By sound, I mean something like a FFP approach). That way, my good methods prosper; and my bad, wilt.

Providing you keep track of what's going on, you can then slowly add additional funds to the methods that are doing well.

Additionally, when you get a new idea, it can be implemented easily with a small capital allocation, and if it does well, simply move more funds to it from less successful methods. By using something like a FFP money management, nothing else changes - just the trade size.

I separate things by using multiple sub-accounts, and assign a separate Excel worksheet for each method to keep track of performance and trade information (prior to automation, this was principally to manage my trade sizing algorithms). I also have an additional Excel sheet that adds everything up to give me my consolidated figures. Yard is correct, I feed the stars, and starve the dogs; that's to say capital flows to what's working, and providing my consolidated 'ride' is within acceptable limits, I don't worry too much about the volatility of the components.

Interbank FX Trading

The following thread probably deserves to be read in full by those who are interested in how bank traders manage their positions, and how the Interbank market differs from retail fx trading. The thread is actually about leverage, and started off with the somewhat erroneous notion that leverage was a 'loan' from a market maker, and went on from there.

Here's an extract:

You are right when talking about physical delivery. We would have to borrow one side of the transaction, and deposit the other. However the fx market only physically delivers a fraction on the volume traded and we, of course, NEVER get to physical settlement. All we are doing is going through the motions, and settling the net difference in p/l.

Even in the broader interbank market where there is ultimately physical settlement, players with positions rollover these positions daily to avoid the capital expense of holding their long or short view. The only physical deliveries that need funding tend to be for commercial requirements only, not for speculation. All the rest are netted out. We, by contrast, have continuous interest payments that factor in rollover costs. (btw rollover costs are calculated from interest rate differentials).

Basically, as long as the physical delivery of an outright position is always deferred, by whatever means, no real loan needs to take place. It's all virtual, and can be settled net.

I'll stop now, but I can see this possibly getting involved! We'll see!

and it did! ....

Oh dear! I knew I would be digging a hole for myself! Your questions are perfectly valid, but a little more difficult to explain.

I didn't actually say the fx market never goes to physical settlement. What I said was that WE never get to physical settlement, and that the only settlements that require funding in the interbank market tend to be commercially oriented, and not from speculative trades.

To understand this we have to understand how speculative trades don't require funding even though the core fx market works on a 2 day settlement basis and speculative trades represent the majority of fx market volume.

At the heart of the fx market are the Banks. Now say a bank buys and sells EURUSD all day long, and wants to run a long position of EURUSD 1,000,000 for say a week.

Ok - the bank balances it's trades so that it is net long EURUSD 1,000,000 by the end of the day. Maybe they have bought EURUSD 50,000,000 and sold EURUSD 49,000,000. Now even a Bank does not want to fund the remaining EUR 1,000,000 so it rolls the position over ‘Spot Next’. That's to say it Sells Spot (2 business days out) EURUSD 1,000,000 and Buys back the next day (3 business days out). Now when settlement comes they have a completely balanced book, even though they are running a EURUSD 1,000,000 position. The only ‘funding’ that would need to take place is in their own bank account (known as a Nostro account), but this isn’t necessary because they are balanced on the day! EURUSD 50,000,000 went in, EURUSD 50,000,000 went out.

Now all that needs to be done is to continue to roll this position over for as long as it's required. Rollovers can be done Cash (today/tomorrow), TomNext (tomorrow nextday), Spotnext, or in the forwards market.

If you can grasp this concept, and understand that OANDA and others plug into this system, you'll understand that providing delivery is always deferred, there is no need to raise the full capital for a transaction, we just have to honor day-to-day price fluctuations. Yes there needs to be capital adequacy, but nothing like the full amount of the value of the underlying trades.

How does this work for us, since we never settle? Say every trader at OANDA was Long AUDUSD. OANDA would offset this in the interbank market becoming Long AUD/USD with their banking counterparties. They would also rollover this position each day with the Banks so they didn’t actually take delivery of the AUD (and have to come up with the corresponding USD). OANDA screens us from this process, but passes on the costs of these transactions by means of continuous interest payments (both debit and credit). Our capital adequacy is of course our underlying margin. That's all we need.

and....

The spot dealer is generally responsible for the size of his position (within his limits) and balancing his settlement accounts in the currencies under his control on settlement day (which is everyday on a rolling basis). Backoffice certainly confirms settlement activity, so there is a double check.

If a dealer fails to do this the Bank will be up for a lot of interest, or worse, risk failing on payments.

Generally the dealer will make his adjustments ( ie balance his settlement accounts and defer any outright delivery) by trading spot next, or at the latest tom next. He could also swap for longer - 1 week, 1 month .. etc., if he so wished by trading forwards.

Clear as mud?

Let me just add that there are two distinct kinds of transactions in interbank fx. They are outrights and swaps.

Outrights are a Buy or Sell for a fixed settlement date, normally spot. We trade a pseudo version of spot at OANDA.

The other, 'swaps', are all double transactions - Cash (today,tomorrow), Tom Next, Spot Next, 1 week (Spot, 1week) , 1 month (Spot,1 month), etc. With all these transactions there are two dates, and the pair Bought and Sold, or Sold and Bought for the price of the swap.

By using a combination of spot and swaps, a dealer can create all kinds of positions as well as deferring delivery indefinitely, if necessary. It's the mechanism that makes this whole circus go around.

Confused now?

~chaffcombe

and finally...

If we wanted to duplicate our OANDA trading as a Bank dealer, we would establish our positions by trading spot, and roll them over each day trading Spot Next or Tom Next for the life of our transaction.

Backtesting

I've absolutely no idea how some people can trade without first backtesting, however that's just the way I'm wired! For those interest in the subject the following discussion may be of interest. I notice that I made a number of contributions (~chaffcombe)

Carry Trading

Although I do not carry trade in my published accounts here, I maintain a carry trading basket in another set of accounts. To my mind, it's a very valid trading method, and I have been rewarded with a return of about 600% over the last year, which isn't too shabby, particularly since I only monitor it weekly! Want to know more? You are in luck - I've set out exactly how I do it in the thread below. You just need to read the whole thread.

I've described how I manage my carry basket in a past thread (I don't know which one), but basically I've made the decision that I want to run my carry basket indefinitely.

In order to do that I allocate a fixed percentage risk - let's say 15% of the account, calculated somehow. Each week, I review the pairs I'm holding for interest rate changes, and adjust my exposure to each pair, depending on my current NAV. If I've made money during the week, I buy a little bit more; if I've lost money during the week, I adjust my holdings down. I therefore always maintain a fixed risk, no matter how much money I have in my carry account.

With the above approach it doesn't matter how many pairs are in your basket. You simply divide your fixed risk by the number of pairs you wish to carry. If the number changes, so does the trade size of the remainder.

Believe it or not, trading once a week works! Although, as I've commented before, carry trading like this is not necessarily for the faint hearted.

All things being equal, people will invest in high yielding currencies, at the expense of low yielding currencies, for obvious reasons – yield. This demand causes the price to change, again all things being equal, in favour of higher yielding currencies.

In carry trading it's this tendency for rate appreciation that normally attributes most of the gains. It's also why I like to see at least a 3% differential, to encourage this significant phenomenon to take place,

Don't forget this is a broad generalization that only suits a long-term strategy. Day-to-day anything can happen.

To see this working in action you only have to look at the relative strength of the AUD and NZD. Also, the USD was in freefall until they started raising rates. The freefall first stopped, and then reversed.

Finally, 99% of market moves following economic announcements are based upon perceptions on how the information may affect interest rates. If it’s bullish for rates, the price goes up; if it’s bearish for rates, the price goes down.

~chaffcombe

Equity Curves

It is one of my contentions that one of the most important things a trader can do is accurately plot his or her equity curve. The process takes little to no effort once you have set yourself up (I use Excel). Again, IMHO, I believe this should be in percentages, not dollars. Looking at our trading accounts in dollar terms reinforces a whole load of psychological problems that will obstruct us from ever trading decent size, so I advocate tracking your 'cumulative daily percent' as I do in my monthly reports, published here. Obviously tracking return on original investment is simple, but once you start making capital adjustments, you will need slightly more complex calculations. For those interested, I explained how I do it here (posts 40 and 43).

Hedging

As with anything, there may be some occasions when you may want to do go against the pervasive logic, but these should be few and far between - for example my home currency is AUD, so I have a permanent hedge in place to protect my AUD investment at OANDA, which is held in USD. Likewise you may find it convenient to have opposing trades when you trade both long term and short term systems in the same currency pair, or different trading ideas for the same pair, but you pay for this luxury, with decreased overall exposure and increased costs. It's far more cost effective to net your trades, but that takes a little more work.

A manufacturer can hedge their overseas currency exposure, but as f/x traders it's generally counterproductive, because that's how we get paid! At any given moment we have to decide the total amount of exposure we wish to have, and produce that position as efficiently as possible. Consider, a net position of zero produces no P/L no matter how much it costs to set up! Even Options are not a panacea, but let's not get into that!

In a nutshell, we are in a risk business, and we manage that risk by determining the size of our net exposure, relative to our account size, while keeping costs to a minimum.

The following is part of my first post on the OANDA discussion forums, mainly in response to the popular notion that it's somehow beneficial to hold both long and short trades in the same pair at the same time. It also raised the point that with any trading platform there will be pros and cons that the client needs to be consider.

It must be remembered that OANDA is a Market Maker, not a Broker. They make their money from the spread. If you want to earn the spread yourself you must deal through a BROKER, not a market maker (There are alternative platforms out there that will allow you to do this, BUT you then pay a fee).

Ultimately this is about deciding on which trading platform suits your trading style the best. Personally I think that for most people OANDA is way ahead of the rest, because of their 'no minimum trade size' policy. It means that you can scale your trades to your account balance, and still deal at competitive rates. If you do everything by percentages it means that if you can manage a $1,000 account well, there's no reason why you cannot manage a lot more (and have the basis for a life-long source of income). With the other platforms, you have to risk a fortune just to learn the ropes. Not a good way to go.

Hedging your domestic currency

If USD or EUR's are not your domestic currency, and yet your domestic currency is one traded at OANDA, the most conservative option is to open a 2nd linked account and sell your USD (or EUR) account balance against your domestic currency. Then just leave it alone, except for periodic adjustments, as your account changes in value.

For example if your domestic currency is AUD, and you have sent OANDA A$5,000 to open an account. Make your first trade Buy AUD$5,000 against USD, and stick the trade in a second account. DON'T TRADE IT - as it's a permanent hedge.

If you later have a view on AUD/USD treat this as a separate position and trade it in your main account. (This is one on the very few times I advocate trading 'both-ways').

As your account grows or shrinks adjust your hedge to represent your latest NAV.

BTW if people in this position don't do this, they are speculating, whether they like it or not, on the currency in which their account is denominated.

Inspiration

I was just recently having a private conversation with another trader, and I thought that his description of how he started; the obstacles he faced and how he stayed the course truly inspirational; and too good not to be shared. With his permission therefore, but with the promise of not revealing the source, here's an excerpt:

Leverage

It is a popular misconception that the market maker with whom we deal lends us money when we use leverage. This isn't correct. The following thread discusses what really happens, but also goes off to discuss a number of the finer points of trading fx (threads are like that!). For those interested in how things really work, it is a good read.

Margin Calls

The following explains in practical terms how margin rules work at OANDA. It was given in response to a newcomer considering using Margin Calls as a risk management tool. Not something I'd recommend!

If you are 'fully loaded' a margin call will take half of your account. If not, it will take even more !
Please understand what you getting into (ie the margin math's) before you risk real money with this strategy.
.......................................................................

Say you have a $1,000 account.

If your margin is set at 50:1 you can take out a $50,000 trade. OANDA requires $500 maintenance margin for this. Therefore if the position loses more than $500 (ie your NAV drops below $500), OANDA will close the position. The loss is therefore approximately 50% of your account.

However, if you had only taken out a $30,000 trade, OANDA only requires $300 maintenance margin. Therefore the position can lose $700 before OANDA will close the position. The loss on margin call is therefore approximately 70% of your account.

The same principals apply to any margin setting.

~chaffcombe

Profit Calculation

For those who are unsure of some of the calculations used to determine returns, the two threads below discuss the subject at some length. In particular the latter part of the second thread is worth reading, thanks to the input from Tom Schneider (tsch)

Profit Potential

Well, it's just about unlimited, isn't it? Here's something a wrote in a thread that deteriorated after a while. Never mind. BTW GrimFD also said somewhere else that the next Soros' and Buffet's are amongst us! I agree!

I'd just add to the Grims comments, that the opportunity for individuals to trade at interbank prices, at 50:1 leverage, multiple times an hour, 24/7 (if automated), is a new phenomenon. Who knows where or what the limits are? As has been said before; argue for your limitations, and they are yours – so why on earth would you do it?

However, you do need net profitable systems (those that make more pips than they lose) to go with your 'attitude', but I believe this comes down to what I've said earlier this week about having an unshakeable belief in yourself that you will find a way.

~chaffcombe

Psychology

To do well at this business you need an extra-ordinary psychology, but I think it all boils down to one crucial element, and that is ... (insert drum roll) ...

I'd say the essential core component is an absolute belief in oneself that you will find a way. This has to be unshakeable, or else you are in the wrong business. You cannot dabble at this and expect to get miraculous results.

~chaffcombe

Responsibility

GrimFD is a trader over on the OANDA forums. He is also a deep sea diver by trade, diving off the rigs. This must be just about the most dangerous civil profession there is, which makes him especially qualified to talk on the subject of personal responsibility.

1. You have to train and train and train till handling any event is second nature. It means you have to experience the events in a controlled environment. In fact you have to train to the point where you subconsciously react before you even consciously realise you have a problem. Spikes, slow markets, trends, re-quotes…with training/practice you can bring them on, no problem…trade those demo’s….experience the problems.

2. Know your craft. Equipment, procedures, back-up procedures. Which gasses, which mixtures for which depths, what tables to use? Your team. Who is responsible for what and how much you can rely on the individuals to help you prepare? The team here is the forums. People working towards a common goal. Everybody brings something to the table. Ask yourself what do you bring and how are you sharing? Learn things like what to do if you lose connectivity, if bad data causes spikes that kick you out of trades. Who to phone, how to get back online in a hurry etc…what happens in if you’re off-line 24 hours…

3. Take whatever you think is possible and time’s it by 10. Even then you’re not even close to what is really possible with the right motivation. If you think 10 pips a day is possible, aim for 100 and you’ll fail at 20 or 30. If you aim for 10 you’ll get 5.

4. Don’t pass the buck or blink under pressure. You’re the one swinging the hammer. If the bung implodes in your face and rips of your head it’s your problem. With a diving problem you probably have less than 45 seconds to sort it out (or as long as you can hold an exhaled breath, you only realise an equipment failure when you take the next breath). Stuff happens. You either stay cool and handle it, or you make the papers. If stuff happens in the market same thing. Might be somebody else’s fault, but it’s your problem. Stay cool.

5. Procedures. Work them out before the dive. Comms fail. Fine, 3 pulls on the umbilical or flashing light on the helmet. Primary fails, go to back up, back up fails…call for a Pneuma, Pneuma fails, signal like hell and hope your buddy was paying attention in CPR class, your last chance just became his retention span and how much attention he’s paying to the umbilical. …have those procedures in place for market events. Know your back ups as well as you know the trading…

6. Safety, safety, safety…on rigs you have to report near misses. They know for every 600 or so they have, one fatality occurs. If a rig owner sees the near miss reports climbing he’s on the radio, the rig’s getting sloppy and a fatality is one incident closer. Watch your near misses. If you’re making mistakes something’s slipping and your margin call is that much closer. Tighten up your ship…

Our job is to win with the hand we've been dealt...we can ask for a new card or two, but you can only draw so many new cards before you have to play the hand.

Scalping

Before proceeding, I'm going to wish wdbaker well, and assure you that I'm not trying to pour cold water on what you are suggesting. In fact, what you suggest is the way I would prefer to trade, if only it would work!

Following below is something I wrote in a previous post about the extreme adverse risk/reward scenario you get when you scalp with a market maker like OANDA, as opposed to dealing with a broker, and attempt to earn the spread yourself. Believe me, I wish the following wasn't true, but I don't see how it can be. Of course, nothing I've said below will stop you giving it a serious go, but you should be aware of the math’s!

'.....it seems to me that scalping for pips when dealing with a market maker is doomed to failure. Let me try to explain why.

Say we are trading a 2 pip market like Euros, and let's just say we want to make 5 pips, and are willing to risk 5 pips.

Ok. We buy at 60, t/p 65, s/l 55

Now here's the problem: To buy at 60 the market must be 58-60. To reach our profit target it must move to 65-67 (a 7 pip move). To reach our s/l the market must move to 55-57 (a 3 pip move).

Just look at that 7:3 risk/reward ratio. The market has to move over twice the distance for us to profit the same amount as we have risked! Those are terrible odds - we would stand more chance at roulette!

By comparison, if we are looking for 100 pip moves, say Buy at .9860, t/p .9960, s/l .9760, the market has to move +102 for us to win, and -98 for us to loss. It's still pretty horrible, but it's a hell of a lot better.'

Stop Losses

Michael; there's a lot of rubbish written about stop losses, and the relationship between t/p's and s/l's, so let me say that there are very few absolutes in trading; we all have to find out what works for us, and go from there.

For example in my personal experience I have found that stop losses of any point size are detrimental to my trading performance, so I only use them to protect my account against catastrophic ruin; not as a trading tool.

By this you can safely assume that I generally take profits at smaller intervals than when a stop loss is hit. This may go against popular wisdom, but it works for me - and that is all that is important. Somehow you have to get to the same place - ie find out what works for you.

My previous post was no more than an example that said the pure mathematical odds of success improve as you increase your targets. It was not advocating a 1:1 relationship between t/p's and s/l's. Try some different relationships for yourself - 7:1 5:1 3:1 1:1 1:3 1:5 1:7 etc., and see what works best for you. However, I can almost guarantee that if you use tight stops you will lose, because the spread (as explained above, and no matter how small) will get you.

Synthetic Pairs

Synthetic pairs are pairs that you calculate yourself from the underlying data of their components. For example NZD/JPY can be calculated from NZD/USD and USD/JPY. Normally you would be interested in calculating synthetic prices to arrive a price for a pair not offered by your market maker, but on occasions it can be useful to calculate the price of a pair yourself, where you also have underlying information. For example OANDA quotes EUR/AUD, but it is nearly always cheaper to trade the underlying (AUD/USD, EUR/USD - often by some 3-4 pips). You need to first do the calculations to see if it is worthwhile.

The following is a small extract of Java code, that shows the concept for calculating synthetic pairs

// --------- Calculate derived (synthetic)------------------

// default calculation of derived - (all non USD based pairs)
// i.e. AUD/NZD, EUR/AUD, EUR/GBP

crossDerivedBid = (pair1Bid/pair2Ask);
crossDerivedAsk = (pair1Ask/pair2Bid);

// If 1st pair non USD based, and 2nd pair USD based
if( (crossPair.equals("GBP/CHF"))
|| (crossPair.equals("EUR/CHF"))
|| (crossPair.equals("AUD/JPY"))
|| (crossPair.equals("EUR/JPY"))
|| (crossPair.equals("GBP/JPY"))

) {
crossDerivedBid = (pair1Bid/(1/pair2Bid));
crossDerivedAsk = (pair1Ask/(1/pair2Ask));
}

// BOTH USD BASED
if( (crossPair.equals("CHF/JPY")) )

{
crossDerivedBid = (pair2Bid/pair1Ask);
crossDerivedAsk = (pair2Ask/pair1Bid);
}

bidDiff = crossPairBid - crossDerivedBid;
askDiff = crossDerivedAsk - crossPairAsk;
totDiff = bidDiff + askDiff;

// --------- End derived calculation -----------------------------

The parameters passed to this extract are

1) The cross: ie AUD/NZD
2) The first USD pair: ie AUD/USD
3) The second USD pair: ie NZD/USD

The code then works out what the derived calculation is, so it can either be compared to the quoted cross rate, or simply it supplies the synthetic price of a pair not offered - say NZD/JPY.

Once you understand the basic methodology, you can use the same technique to calculate any synthetic price, by combining the two underlying pairs.

What it all basically boils down to is put the underlying pairs in the same terms, and then cross divide.

Tick Filtering

Every now and then one see posts over concerns about a market maker presenting prices differently that the way in which we perceive the 'real' market. Frequently this take the form of outrage against stop running, although this wasn't the case in the thread I've taken this extract. Anyway, my point below is that any filtering can basically favour us, providing we are flexible in out approach.

I've never really understood the concern over tick filtering.

Let's say for a moment that there is a mythical 'real' market, and various market makers base their prices around that. Some might dampen price movements a bit, while others elect for a more volatile approach. None of this can be particularly extreme, as any outside the norm will be arbitraged – no matter what it says in the small print.

Now providing spreads are common throughout neither dampened nor volatile is particularly disadvantageous to us, because each one favors a different trading approach. A volatile price feed favors scalping and support and resistance type trading, but is murder on stops, breakout trading and trend following, while the reverse is true if trading a dampened feed.

If one concludes that there is a bias in the price feed, simply pick a style that suits, or change market maker. If not, we should be concentrating on beating the market in general.

~chaffcombe

Trade Size

This comment only scratches the surfaces of what is one of the most important topics of all, but it does establish the basic criteria in the right sequence.

I don't think you can set trade size and then decide on your s/l. It should be the other way around.

First you decide the maximum % of your account you are prepared to lose if the trade fails in a worst case scenario.

Then you decide how wide your stop needs to be to accommodate things like normal market movements and the nature of your strategy, etc.

Finally with these two bits of information you can work out your trade size by simple maths.

Now if you want to know how I set my trade size... this is a description of the method I use, which I normally refer to as Variable Fractional Percent (VFP)

If you want another money management idea that is very responsive to performance, yet doesn't make you make too hard decisions up front try this:

Firstly you need to know your daily Net Asset Value (NAV) gain or loss in percent.

Start trading at a conservative 5% FFP (or whatever suits you). But instead of using a Fixed Fractional percent, you use a range say 2%-25%. Move up and down the scale by adding or subtracting half of your last daily NAV percentage.

For example start at 5% FFP

Next day profit on trades = 1.5%
Therefore your next FFP = 5 + (1.5 * 0.5 ) = 5.75%

Say you then lose 3%
Next FFP = 5.75 + (-3 * 0.5) = 4.25%

Obviously use ranges and a daily factor (here 50%) that suits you, but you'll find this method really rewards good methods, and lightens up very quickly on bad.

I use it myself, and find it very sound.

~chaffcombe

There was also a very good discussion on trade sizing methods (FFP, VFP and Fixed Ratio) here:

Turnover

I've been flamed a few times on these forums for saying that we should endeavor to traded as efficiently as possible because the spread costs of our transactions can become a significant issue.

Anyway, as many of you know I now trade with the api, and as a consequence I've just gone through my volume figures for the past month. Now I think I'm trading conservatively at the moment - low(ish) leverage, and modest trade frequency. Even so, I am turning over my account over 6000 times per annum! (That's USD trade volume for the month x 12/USD Balance). I even had one sub-account that turned over 2000 times, in the last month alone!

I know that is nothing compared to dgcfx's comment sometime ago that it's quite possible to do $80m turnover pm off a $10,000 account, but it still goes to show that turnover can be somewhat high - and possibly much greater than some might expect, which is good news if you are interested in the api, but bad news if you like to offset positions in different accounts.

But anyway; I don't want to get in an argument about that. All I'm saying is that some may like to do your own calculations - you may be quite surprised at the result.

OANDA's got a great business, haven't they?

~chaffcombe