Download the last wee k of prices.
Estimate the inst rument risk.
Update the stop loss gap calculation for any positions that you might ope n next week.
Update any exchange rates you need to work out pos ition sizes.
Calculate the size of positions needed if you were to open a positio n next week.
Decide which dated position you’d own if you were to open a position and evaluating whether a current position needs t o be rolled.
That leaves the following tasks to complete on weekdays: Check to see if any trades require opening (which can be done using free internet price charts) or closing (which can be done using aler t services).
Do any opening trades that are required, using pre-calculated pos ition sizes.
Do any rolling trades that a re required.
Update stop loss levels (and any stop loss orders left with your broker) using your pre-calculated stop loss gaps, if the price
makes a new daily high (for long positions) or low (for short positions).
Summary
That is the end of part two of the book. Congratulations for getting this far! You know how to use a trading system to trade a single instrument, using a single opening rule. Once you have practiced trading the Starter System and are comfortable with it, you will probably want to start improving it. Part three of the book explains how to do this by diversifying the system: adding more instruments and more tr ading rules.
20 20
77 75 The SPY ETF tracks the S&P 500 US equity index and is highly liquid. It is comparable in costs and minimum capital to large cap stocks like Facebook, although Facebook currently has a slightly lower minimum capital level and cost level. I’ve chosen this ETF rather than any individual share as it’s possible to get much longer data history for the S&P 500, which makes it more suitable for back-testing.
20 20
77 76 Some instruments have both closing and settlement prices, some have just one or the other. There is a subtle difference between these two types of price. The closing price is usually the price at which the last trade of the day was done, whilst the settlement price is averaged over a short time period before the exchange closes. Where you have the choice, you should use the settlement price, as it is based on multiple trades. This makes it more reliable.
20 20
77 77 Appendix A includes links to sources of data. To name just two websites that are active and free at the time of writing: quandl.com and barchart.com
20 20
77 78 Technical note: Specifically, the effects of carry (contango and backwardation) will be removed from the price. If we take bond futures as an example, these tend to be in permanent backwardation with the spot price above the future. This means that going long the future and regularly rolling it will earn an additional return which translates into an upward drift in the stitched price. This in turn introduces a long bias into a trend following system which uses the stitched price rather than the spot price. Removing this bias by replacing the stitched price with the spot price will reduce back-tested performance and will also reduce future performance if bonds continue to attract a risk premium.
20 20
77 79 Apparently, the adjustments made to prices are evocative of ships rising up or down in the locks of the Panama Canal. There are other more complex methods for back adjusting prices, but I prefer the relative simplicity of the Panama method.
20 20
78 70 I’ll explain later in the chapter when you should switch from one dated product to the next.
20
78 ¹ At the time of writing quandl.com provide this data, but at a premium price.
20
78 ² There are roughly 251 trading days in a year, depending on where and what you are trading. But as standard deviation scales with the square root of time it’s easier to pretend there are exactly 256 days, since the square root of 256 is 16.
20
78 ³ Technical note: Without making assumptions about the ratio of intraday and overnight volatility it isn’t possible to calculate the relationship between standard deviation and ATR with an explicit formula. This figure was derived by taking an average of the ratios of each risk measure across 37 different futures contracts.
20 20
78 74 This rule is here to keep the Starter System simple. In theory we should adjust both stop loss gap and position size when instrument risk changes, but for the vast majority of traders this is unnecessary work which adds no significant value. I discuss more complicated position management rules in chapter ten.
20 20
78 75 To be clear: this means trading at most once a day. This is not the same as day trading, where you always close your positions before the end of each day! As I’ve already mentioned you should avoid day trading: you will trade too quickly, and your trading account will be emptied quickly.
20 20
78 76 This is the value of all the open positions in your account, plus any unused cash. Some brokers describe this as the net liquidation value.
20 20
78 77 I’ll explain how you can do this properly in chapter eleven.
20 20
78 78 Or settlement price. Where both closing and settlement prices are available, make sure you are consistent with your choice of which to use.
20 20
78 79 This depends on whether the futures are cash settled (if the position expires you get paid your accrued profits or losses) or physically settled (if long you receive actual bushels of corn or bars of gold; if short you have to deliver them), and whether the broker allows physical settlement (most brokers do not allow physical settlement except for certain FX futures, as most retail traders do not have the garage space to store the 5,000 bushels of corn represented by one corn future).
20 20
79 70 One advantage of futures is that you usually have more flexibility about when to roll and which contract to hold, compared to CFDs and spread bets. The reasons for this early roll will become apparent in chapter eight.
20
79 ¹ This is potentially confusing. If the opening rule has been short for weeks, why have we only just closed the trade? It’s
because the opening rule and closing rule are different: the Starter System uses a moving average to open trades, and a stop loss to determine when to close a trade or not. In part four of the book, I’ll explain how you can use the opening rule to close as well as open trades.
20
79 ² Remember the account we are using to trade gold with is a GBP
account, and we are betting on the price in £ per point (even though gold is priced in dollars). Hence no FX conversion is needed.
Part Three: Diversifying
Chapter Seven
Adding New Markets
I opened this book with a quote from legendary investor, Warren Buffett. Here is another of his gems:
“Diversification is protection against ignorance. It makes little sense if you know what you are doing.”
20
Many traders concur with Mr Buffett 79³ – diversification is a dirty word. In their opinion you should become an AUDUSD FX
trader, or a crude oil trader, and stick to that. Trying to trade AUDUSD and crude oil is a waste of time. This attitude stems from days of old, when professionals traded exclusively ‘on the floor’
in physical, rather than electronic, exchanges. Exchange floors were split into pits , each trading a different instrument, and traders were fiercely loyal to th eir own pit.
Today’s trading gurus still claim that you should find an instrument that ‘suits your trading style’, build up ‘expertise’
in that particular market, and then trade it exclusively. I don’t agree with these self-proclaimed experts. For traders like myself, who use a systematic approach, diversification makes a huge amount of sense. In this chapter, I will explain why, how, and when you should diversify your trading across different instruments.
Advantages and di
sadvantages of diversification In the Starter System we picked a single instrument to trade, but it was impossible to predict which instrument would be the most profitable. From figure 2 , there is no robust historical evidence to say that any one instrument performed better t han another.
But choosing one instrument is risky. The Starter System looks for trends in markets and it’s unlikely that all the worlds markets will simultaneously have profitable trends. If we are unlucky, we could pick an instrument that turns out to be a poor performer, like the S&P 500 margin trade in the previous chapter.
By trading a variety of instruments simultaneously you will be in a better position to catch any trends wherever, and whenever, they appear.
For this reason, large systematic hedge funds are constantly looking for ways to diversify the set of instruments they trade.
When I was working at AHL, I recruited Ben – a smart young graduate. Soon after Ben joined, I assigned him his own project: adding a large batch of new instruments to one of our flagship funds. Ben wasn’t too happy and probably thought this boring and mechanical task was a waste of his undoubted academi c abilities.
I said: “You probably don’t appreciate this right now, but this project will add more value to this company than pretty much anything else”. I was right. Over the next few years the fund went on to win several awards for outstanding performance, much of it thanks to those ex tra markets.
Why diversification works: the maths
Let’s suppose you’re trading the Euro Stoxx 50 equity index, which is the instrument I recommended for CFD traders in chapter six. Another CFD that’s available with a reasonable minimum capital and isn’t too expensive are US 10 -year bonds.
Why would it make sense to trade both of these instruments, rather than just one? If there is no evidence that one instrument is better for trading than another, then how can trading two instruments improve your expected returns? In fact, you should not expect diversification to improve returns. Diversification is good because it r educes risk .
Risk is reduced because the returns from trading US 10-year bonds will be different from those earned in the Euro Stoxx. Some days your 10-year bond trading will be profitable and the Euro Stoxx will lose money, whilst on other days, Euro Stoxx will beat bonds. Trading both instruments will reduce the variability of returns in your trading account. To confirm this, I back-tested the returns of the Starter System trading both instruments. I found that the risk of the entire trading account came out significantly lower than the 12% risk target: 9.2% a year.
Lower risk is nice to have, but the risk of the Starter System is already fairly modest. Most traders would prefer to stick to the original risk target and swap the diversification for higher returns. This is done by applying more leverage and running the trading strategy for each instrument at a higher risk target.
To do this you should increase the risk target on each individual instrument by a multiplying factor, which in this case is 12% ÷
9.2% = 1.304. This factor is the instrument diversification multiplier (IDM) . The IDM exactly compensates for the reduction in expected risk which comes from trading both instruments.
Now we have:
Euro Stoxx trading strategy, with a tar get risk of: IDM × risk target = 1.304 × 12% = 15.6%
US 10-year bond trading strategy, also with a target r isk of 15.6%
Using formula 19 I can work out the expected pre-cost return of each instrument:
r = (SR × s) + b
Where SR is the Sharpe ratio, r is the average return, b is the rate we can borrow at and s is the standard deviation. From chapter five the expected Sharpe ratio without costs for one 20 20
instrument 79 74 SR =0.24, the risk target including the diversification multiplier s = 15.6%, and b the rate we can borrow at is currently around 2.0%. The pre-cost expected return on each in strument is:
r = (SR × s) + b = (0.24 × 15.6%) + 2% = 5.7%
If we expect both instruments to provide the same expected return, then we will also make 5.7% on the trading account as a whole. The individual instruments have risk of 15.6%, but what is the expected risk for the whole account? Thanks to the power of diversification the risk will be lower by a factor of (9.2% ÷
12%). The risk for the entire account is: 15.6% × (9.2% ÷ 12%) = 12%
So now we have:
Euro Stoxx CFD trading, with a target risk of 15.6% and an expected ret urn of 5.7%.
US 10-year bond CFD trading, also with a target risk of 15.6% and an expected ret urn of 5.7%.
Entire trading account, achieving a target risk of 12% and an expected ret urn of 5.7%.
Back in part two, on page 100 , I calculated that the Starter System is expected to make 4.9% before costs. Adding the extra instrument has increased the expected return from 4.9% to 5.7%
without changing the account level target risk, which remains at 12%. The extra leverage added by the IDM has converted the risk 20 20
reduction into improved expect ed profits. 79 75
Diversify or trade a cheaper instrument?
The calculations above are done on a pre-cost basis. But adding more instruments will probably increase your costs. Rather than choosing a cheap instrument with higher minimum capital, you will have to opt for two instruments with lower minimum capital, that are likely to be more expensive. If you have enough capital is it better to diversify, or to cut costs by trading a single cheaper instrument?
Let’s find out with an example. Before reading this chapter a US
CFD trader with $25,000 would have looked at table 5 and decided they should trade German 10-year Bonds (known as ‘Bunds’ by professional traders). Bunds have a lower cost than both the Euro Stoxx and the US 10-year bond CFD products. But they need $10,500
in capital; and ideally twice that: $21,000. Alternatively, with $25,000 the trader could trade Euro Stoxx (requires $6,500
minimum, and ideally $13,000) and US 10-year bonds (ideally nee ds $12,000).
Which is the better option? We start with formula 19 : r = (SR × s) + b
Where SR is the Sharpe ratio (0.24 for the Starter System), s is the standard deviation (12% for one instrument), r is the average return, b is the rate we can borrow at (I assume 2%). Subtracting costs, I get the formula for post-c ost returns: Formula 26: Expected post- cost returns r = (SR × s) + b − (c × s)
Where c is the expected cost in risk-adjusted returns (equivalent to Sharpe ratio units). To translate risk adjusted costs into annual returns I multiply by the standar d deviation.
Trading only Bund CFDs wit h c = 0.038: r = (0.24 × 12%) + 2% − (0.038 × 12%) = 4.42%
Now for the diversified CFD account with two instruments: Euro Stoxx and US 10-year. For Euro Stoxx ( c = 0.04) with s = 15.6%
(because of the diversification multiplier): r = (0.24 × 15.6%) + 2% – (0.04 × 15 .6%) = 5.12%
For US 10-year bonds ( c = 0.074) also wit h s = 15.6%: r = (0.24 × 15.6%) + 2% – (0.074× 15 .6%) = 4.59%
To calculate the return of the whole account we take a weighted average. The weights are the proportion of account value in each instrument:
r = 5.12% × (13,000 ÷ 25,000) + 4.59% × (12,000 ÷ 25, 000) =
4.87%
This is significantly higher than the 4.42% return obtained by trading Bund CFDs, even though Bunds are cheaper than both Euro Stoxx and 10-year bonds. In general, diversification is nearly always better than using a larger account to trade a cheaper instrument .
Why diversification works: the evidence Some of you may be unconvinced by the fancy maths above. Good!
Cynicism is a very desirable quality in any trader. Let’s look at some evidence which, as usual, is drawn from a back-test. It will tell us what happened in the past, but not what will occur in the future.
When looking at back-tests it’s important to use as many different instruments as possible; this is so we know that the results aren’t just a fluke, and also that the author hasn’t just cherry-picked their favorite. It�
�s also important to check the statistical significance of the results, just like I did back in chapter five when deciding which instrument and opening rule to use.
The results are in the box and whiskers plot shown in figure 19.
On the extreme left-hand side, you can see that the Sharpe ratio (SR) of the Starter System for a single instrument, with returns averaged across all the instruments in my data set, comes out at 0.24. As in previous figures, (i) the line in the middle of the box shows the average, (ii) there is a two-thirds chance the true SR is inside the box between 0.17 and 0.31, and (iii) the whisker lines outside the box show there is a 95% chance the SR is between 0 .1 and 0.39.
So, if I chose one instrument at random, and trade the Starter System with it, then on average I expect to have a Sharpe ratio of 0.24, with a 95% chance the SR was between 0. 10 and 0.38.
Figure 19: Adding instruments significantly improves returns –
Average Sharpe ratio with error bars as instruments are added to a trading strategy
What performance can I expect from trading two instruments?
Moving rightwards to the next box, if I add another random instrument the average SR leaps to around 0.3. Adding further instruments continues to increase the expected back-tested 20 20
performance of the trading system, 79 76 and once I have a dozen instruments the SR is significantly better than the SR for a single instrument.
This is very clear evidence that diversifying across different instruments will improve the Sharpe ratio, and thus our returns.
If you are able to trade 15 or more instruments you can expect to double your SR. There is no other change you can make to the Starter System which will deliver such an improvement, as you’ll realise once you have read the rest of the book.
In fact, this evidence actually undersells diversification.
Rather than picking instruments at random, as I did here, we can deliberately choose instruments that bring the maximum amount of diversification into our portfolio. This will produce even higher improvements in Sharpe ratio. I’ll discuss how this is done later in the chapter.
Leveraged Trading: A professional approach to trading FX, stocks on margin, CFDs, spread bets and futures for all traders Page 15
