pradical way of evaluatmg a portfolio without requiimg special mathematical knowledge or expensive software is to use a spreaddieet program. Using a classic stock and bond mix. Table 23-2 shows the monthly rehims of the S&P 500 (S&P) and Ihe Ldiman Brothers Treasury Index (LBTI) combined into a portfolio of 6(Po stocks and 40"o bonds. Columns and show the monthly percentage rrtums of Ihese Iwo inputs and columns E and F have the correonding cumulative retums. At the bottom of columns and are some basic calculations that describe the performance. The S&P has an annualized rehim of 16.4Po with a monthly standard deviation of 3.32"o; bonds have a 7.650orehim with a2.6Postar-

BPeterL Bem.-tein. The P.table ME A in bve.-huent (jliu Wiley&Sons, 1995,p 25j.

of Stoclcs and bonds u«ing a Spreadsheet

Sep 94 2.«5 -3 09 71 2 1 - . * -1 94

Nov94 - * .- 1. 9 I 02 -8.41 -27S

C¹c94 1.48 2.42 I Mt. 2.50 -S.»9 -O.90

r,* Vi 3 90 2 344 899 -0.4 5.22

» »5 29 OA4 203 I 1 <)4 Oil 725

AprVS 294 ire 24 l-iae t 99 9 7-2

1 95 4. > 790 S5b 18 9.89 ISM

un9S 2,32 IJOt t82 21 20 IO.M 17 IO

1 95 . -l-M 1-34 24.52 9.31 18.44

02S 2.13 1 2477 1144 19.44

Sep 9b 4 22 1.73 3 22 2B.99 13-17 22.6*

Oct 9S -O it, 293 0,96 28.fi3 l« IO 23 *2

Nov9S 4 3» 2 57 33 02 18 7 37 2a

Dac9S 193 2.67 2 23 34.9S 2134 29.SI

Jan 96 3.26 -0.02 195 38.21 21 32 31.45

Feb 96 0.69 -182 151 38 90 l«.50 29 94

Mar 96 0.79 - I 8S 27 39.69 I4«S 29.68

96 IM -172 02 4I03 12 91 29.79

May96 2-28 -0- I 3S 43-32 1289 3115

Jun96

29.38

: &

20.4*

97 S97 2 34 4.52 63 95 20.3ft

97 609 1.14 4 11 70.O4 2150

Jun 97 I 94 3.46 74 52 23.44

Jol97 796 5-89 7.3 82.48 2933

/Vug 97 -5-60 -2 87 4.51 76.88 26.46

Sep 97 S.48 2.84 4.42 82.36 29-30

deviation. Tlie ratio of AROR to standard deviation shows that the stock maiket has been rduramg a much better reward to rid: ratio during these 4 years.

Using the standard 60°o stock and 4000 bond portfolio, column D (row 5) becomes + 5*. + C5*.4 and column G becomes the cumulative value of that portfolio. When the same statistics are calculated for the mix, the AROR is 13.13 0 with a standard deviation of . . As expected, the retum of B.Bois about 60° of the difference between the bond and S&P retums; however, the standard deviation has inaeased only (rather than 40°o) of the difference between the two initial standard deviations. The ratio of 4.84 is very close to the S&P ratio, reflecting the sharp drop in rid; due to ccmbining Ihese iwo assets. This easy method allows us to lockfcr maximum drawdown, plot ttie original assets and final portfolio, and even apply a sliat to ttie new equity stream.

INDIVIDUAL TRADE RISK

The first line of Me in controlling risk is the individual trade, although sequences of trades are also an important consideration. Trades can be viewed as one continuous event for the pmpose of assigning risk; alternately, a strategj may consider separating the trade into two or more parts, for example (I) from the time of entrj until the trade becomes profitable, and (2) from the time it is profitable until the end of the trade.

Rid; on Initial Positions

We associate the entry point of a trade with a period of greatest uncertainty. Depending on how quickly you anticipate the new trade, prices are either about to change direction or have just changed. Each tjpe of trading method has its own intrinsic rid; associated with initial entries. The trend philosophy will take small, frequent losses and reenter the trade as many times as necessary., the breakout technique will only limit losses by the size of the current trading range. It is not clear that the accumulated sequential losses of the trend method is less than a single loss in the breakout approach; and, in exchange for these two distinct approaches, the profit profile is also different.

Additional Risk Contiol

Most traders prefer specific rid; controls added to the one that occurs naturally when the trading method reverses position. These are generally considered stops, because they are converted to price levels and subsequently into orders that win force the exit for a current position. Three of the more general approaches to determining the level of individual trade risk are

1. An edimate based on initial margin, for example, 50010 to of initial margin. This is loosely related to long-term volatility and lags considerably An estimate of long-term volatility may be a more satisfactory altemative.

2. A percentage of the portfolio, for exanple, I.0°o to 2.5"o. This concept of equalizing rid; (and perhaps reward) across all markets is very popular; however, it is not sensitive to individual markets, and as with many stops, imposes artificial overrides, if the volatility becomes very high, as in the case of the S&P, then this risk level could be reached on every trade causing consident losses. It is, therefore, necessary to determine when a market with exceptional volatility should be removed fran the portfolio.

3. The ma.ximum adverse excursion determined by historic evaluation. " A stop is placed just beyond the maximum adverse excursion for each trade, or 2.5" o, whichever is smaller.

Stops

Stop orders provide the ability to control risk most of the time. The reason it is not-allthe time" is thai exchange-traded marfcets can gap open, leaving your stop far behind. When stops are placed very close to the market, they are hit with a fi-equency resembling a random dishibution; this is not particularly helpful to performance unless you only want to remain with a trade if it performs properly fran the beginning. When stops are placed much farther fi-om the current price levels, they can serve to protect you fi-om extreme losses, but not ahvajs. When a price shock occurs, stops are often filled far from their intended level and can actually be filled at the worst point of the move. A sjdem should not be profitable because of stops, but must work before stops are introduced.

if stops are to work, they must be based on values other than an arbitrary financial limit. Some worthwhile possibilities that change with the marfcet are:"

1. Advance the stop by a percentage of price, as in Wilders parabolic sjdem.

2. Use a swing high or low point, based on a percentage minimum swing.

3. Use the highest high or lowest low of the recent n periods.

4. Apply a method sudi as Kaufinans Adaptive Movmg Average (KAMA) as a stop.

5. Adjust the stop by the volatility, such as three times a 10-day average true range. Standard Deviation Stop

As a sound statistical measurement of ride the standard deviation can be used to determine stoploss levels. in a method called the Dev-stop, Cjnthia Kase uses the following steps to create stoploss levels for both long and short positions:

1. Calculate the true range (TR) of the past 2 trading dajs using the highest high and lowest low of the 2-day period.

2. Calculate the moving average AIR of TR (in step 1), using 30 periods for intraday charts and 20 periods for daily charts

3. Calculate the standard deviation of the true ranges in step 1 using the same period as in step 2.

4. The stop-loss values are DDEV=ATR + (f*SDE\), wheref = 1, 2.06 to 2.25, and 3.20 to 3.50, and where the larger values of the pairs correct for skew and the larger numbers allow for larger rid;.

5. The dev-stop for long positions is Trade high - DDEV; the dev-stop for short positions is Trade low + DDEV

This method acjuds for volatility using standard statistical measurement and is applied to the extreme profit of a trade to prevent unnecessarj loss of equity.

Kaufinan on Stops

As with many techniques, the use of stops has both good and bad features. The good aspects are summarized in the increased control of risk-in particular, the unexpected, extremely large risk. It is an added assurance that volatility will not cause losses that are out of proportion to sjdem performance and expected rehims.

ideally, a Stop order entered into the market is expected to automatically get you out of a position when prices move against you. It has the advantage of forcing the trader to decide, in advance, the size of the maximum loss, so thai risk is under control. It avoids

jnEisk," TecAaiicalAuBlysiE-f SocbaCiumoditieE (iMober 19931 605

lad-minute decisions and the tenptation to hold a losing position with the hope that prices will recover.

Stops are normally resting orders; that is, they are held by the fioor broker to be executed when the price is reached. For the piuposes of discussion, a stop will be any form of an order that is intended to limit losses to a fixed amount, whether the order is placed in advance or monitored real-time.

Rid; Proteaion or False Hope?

The use of stops, or the intent to limit losses to a predetermined, fixed amount may grve a false sense of security For , a Stop order that is reached during an illiquid or quiet market will result in laige slippage; during a fast market, or a price jump, it will often result in the worst fill. Large traders cannot enter stops because they will move the market and guarantee substantial slippage.

Maiket Noise and the Frequency of Stops

The use of small stops, or the trading practice of exiting a trade with a small loss, causes the total performance of a trend strategj to be worse. Small losses occur frequently due to maiket noise and are not an indication that the trading strategj, or the trend direction, is wrong.

Larger stop-losses offer some benefit for longer-term trading, when high volatility has caused the natural