average, the sjstem scores the current maifcet performance to establish the value of the trending component. Bullish and bearish values are calculated by loofcing at three situations independently: the simple moving average signals derived from both the long- and short-term frends, and the double moving average signal generated by combining the two frends. The two techniques of single and double moving averages are exactly as freated in Chapter 5. The mosl important of the three factors is the long-term frend; second is the short-term frend, and last the relative position of the fest to the slow moving average. The sfrongest upward-moving frend is generated when current prices are above both the festerand slower-moving averages and the faster average is below the slower. The opposite positions would result in the sfrongest downward-frending component. Trends are considered neufral if the most important element, the long-term movmg average, conflicts with the other two factors.

The rate of change in open interest is considered a secondarj reinforcement of the frend. Using a concepl different from the usual charting techniques, Commodex considers it a bullish sign if there is an increasing growth momentum in open interest combined with rising prices. The growth momentum is the difference between the rate of increase of the open interest and the 20-day moving average of the open interest. Continuing a bull move with rising prices and rising open luierest is a classic concept of charting. The bear move is confirmed with rising open interest and falling prices. Commodex also considers a drop in open interest along with falling prices to be a bullish factor. The movement of volume is freated in the same manner as open interest and can confirm a bull or bear frend. An increasing volume momentum with rising prices is siq3port of an upward move; other combinations indicate bear frends.

Added together, the signals are ranfced from a sfrong buy to a sfrong sell with lesser degrees in between. The sjstem must be given credit for the quantification and balancing of these elements, which are generally freated as highly interpretive charting techniques. The rules for appljing the daily signals to trading combine both the individual sfrength of the signal with the movement of the Commodex frend index. The frend index itself acts as an overbought-oversold indicator, encouraging profits at specified levels and considering a position reversal in more exfreme situations; stops are placed using the 20-day moving average with predetermined band penetrations Additional objectives are based on the

profits accrued from a trade, with a 50°o retum justifjing a protective stop on part of the positions, and a lOOSo profit requiring the liquidation of half the position. These money management concepts are an important aspect of any sjstem and tend to round out Commodex.

The Commodex sjstem has been used as an example of combining techniques that have been shidied individually in previous sections. It is important to see that each element is simple to understand and avoids duplication; it has single and double moving averages (with and without bands), momentum indicators derived from volume and open interest, an overbought-oversold indicator formed from a combination of an elements, frading stops, and liquidation based on money management and inverse pjTamiding.

Combining Trends and Trading Ranges

Trends and trading ranges account for most price movement; a small percentage cannot be identified. Most

traders must decide whether they intend to profit from frend moves or sidewajs martlets and then develop techniques that target that objective. A frend is concemed with the periodic tendency of prices to move in one direction without a limiting objective, and a frading range is a definition of price containment. The role of time is different for the two approaches. The frend is expected to advance with some regularity over time, while prices within a frading range can move in any direction with varjing speed, providing the upper and lower bounds are not violated.

Trading ranges and frends take tums dominating price movement, with the trading range most common. Long-term frends are easj to see, as for example, represented in the bujing power of the U.S. dollar and the Consumer Price Index. When there is no actual or anticipated change in fimdamental factors, prices fluctuate in a restricted range, bounded by siport and resistance levels. The top of a trading range is ofien wider than the bottom because of the inherently greater volatility at higher prices; but even with the broader formation, there is a clear point at which prices have gone higher and are no Icnger within the frading range.

It ofien happens that the two techniques of price frend and trading ranges are in conflict: The frend is up and a resistance level is encountered, or the downfrend is stopped by a siport level. Established siport and resistance levels cannot be ignored, neither can a major frend; it makes sense to combine both features. The simplestway is to use the trading range while prices are within that area and follow the frend when prices break out either above or beloW (Figure 22-10).

The exact rules depend on the size of the frading range, the speed of the frend, and whether the support and resistance levels have been established over time. Recent range formations offer less of an obstacle than those that have survived for some time. If the range is narrow, a moving average buy or sell signal invariably occurs at the resistance or support level and is met with an initial reversal. For a larger range, a medium-speed moving average signal may occur closer to the center of the range and allow some oppcrtunity for frend profit before resistance is encountered. For very wide ranges, there maj- be ample latitude for a moving average to signal entrj and closeout without the interference of siport and resistance levels. One method of combining the two techniques is to use whichever signal is generated, regartUess of the sjstem. The following combinations of short position entrj and exit are possible for medium to wide ranges:

1. Enter a new short position when the moving average tums down.

2. Enter a new short position when prices enter the sell zone around the resistance level

3. Enter a new short position when prices fall below the support area, then close out the short position when the moving average tums up.

FIGURE22-10 Combining frends and tradmg ranges.

] Folio* Ihe trend

.............

if hjghef prices

zone Follow the trend ifiowpfices

4. Close out a short position when prices enter the buy zone above the support level.

5. Close out a short position when prices break through a resistance level moving up.

The same rules apply in reverse for long positions. The advantage of this filtering method is that new short trades are not entered at siport levels, which causes immediate losses or prolonged trades of litUe result.

TRADING LIMITS-A DAMPENING EFFECT

The existence of trading limits in some of the U.S. markets has ahvajs been the concem of novice speculators who seem convinced that their first entry into the market win result in an adverse locked-limit move, shipping them of their future and happiness ovemight. There is no doubt that many people have suffered from being caught on the wrong side of a move, but trading limits would tend to protect rather tiran harm them. The limits established by the exchanges are intended to allow free frading within the normal range of price fluctuation. While the agricultural maikets still have traditional daily limits, the financial markets have gone to a sjstem of suspending frading for a short period, then reopening with laiger limits, if properly established, these limits should rarely be reached. As prices rise, the volatilitj of the maiket increases, and so the limits must be expanded to correspond to the new price range. At this point, trading risks become proportionately greater and maigins are raised.

The piwpose of limits is to prevent an immediate reaction to unexpected news from moving prices to unjustifiable levels. The first reaction to bad news is usually more exfreme than is realistic, and when proper assessment of the problem is completed, the results are rarely as bad as initially thought. Limits also serve to give the losing frader more time to capitalize the position, or meet a margin call, rather than be forced to liquidate

Frequency of Limit Moves

Sugar prices, which had a spectacular rise in 1974, can serve as an exfreme case of changing volatilitj Beginning in 1969, Table 22-4 shows the remarkable price move seen in the frequency of limit moves. The confracts shown have no duplicate dates and do not include the actual delivery month.

TABLE 22-4 Frequency of Limit Moves

The obvious conclusion to be drawn from the pattem of the sugar maiket is that limits did not expand quickly enough at higher price levels. In 1975, 22.6Soof all trading dajs closed atthe limit; at low price levels, the limits were of no consequence. The 1974 price of sugar increased by 15 times its 1969 price, while the limits only expanded from 50 to 200 points, a factor of four. Limits present a problem for a frading sjstem because they artificially alter the volatility of the market. The consolation is that a frading sjstem is useless in a maiket that is locked-limit for a week. You might find it enlightening to test your sjstem on a maiket and period where limit moves were evident to see their effect on signals and performance.

The only realistic way of comparing the effects of limit moves against a limitless maiket is to compare the U.S. and London exchanges. Using the May 77 Coffee confract for the month of February 1977, we can show three separate examples of limit moves in the U.S. confracts and the corresponding London action. Figure 22-11 illustrates the closing prices of the two maikets on the same day. Two problems must be considered: the conversion of sterling to U.S-dollars, which was fixed at $1.70 for the entire comparison, and the overlapping hours of the New York Coffee and Sugar Exchange and the London Sugar Exchange. The exchange hours may have some difference in the daily price bul should not appreciably affect the comparison.

The first U.S. locked-limit day was on Februarj 7, when the NewYork coffee prices for the May 77 confrad moved to 230.82 (up 300 points). On the same day, London prices for the May 77 confract moved from 218.80 to 230.56 (actually 2883 to 3038 sterling). A trader who closed out his position on the open in the London market would have taken a flu of 2850 and a loss of 12.67 cents per pound relative to the U.S. market. The traders locked into their New York position could have gotten out on the open on February 8 at 233.50, a loss of 568 points. Had both the U.S. and London traders waited for the closing prices as shown in Figure 22-9, the corresponding losses would have been reduced to 1176 points in London and 318 points in New York