During the mid 80s, Jake Bernstein and I conducted a seminar together. One of the topics he presented was his Dual Exponential Moving Average (DEMA)/Stochastic combination method. Over the years, Jake taught me many things but this particular technique, altered in a specific way, remains among the most powerful in my trading arsenal. The way Jake taught this method was to use the Stochastic in the traditional manner and filter it with the

DEMA buy or sell. In other words, the Stochastic and the DEMA had to both be on a buy, or both be on a sell, before a confirmed Trend signal up or down was given. So what was a buy on the DEMA? Better yet what is the DEMA? The DEMA is a derivative of Gerald Appel's MACD 4 (Moving Average Convergence Divergence) originally developed by Mr. Appel for analyzing stock trends. The MACD is, as Mr. Appel says, quite simple. You take the difference of two Moving Averages of price and create a Moving Average of that difference. The difference of the original two Moving Averages and the Moving Average of that difference, can be plotted as two lines, one fast and the other slow. The equations are in Appendix E.

MACD

Notice, I've used the same two wavy lines on Chart 5-2 to show the MACD that I used to show the Stochastic in Chart 5-1. All I've done is change the scale, since the MACD (DEMA) oscillates about the zero line, while the Stochastic travels between zero and 100. In our work we will essentially ignore the scale for both indicators and simply observe the penetrations of the wavy lines.

Jake Bernstein took maximum advantage of the MACD by using specific exponential Moving Averages rather than whole number inputs as Gerald Appel did, hence the term DEMA 5 .

Gerald Appel, The Moving Average Convergence-Divergence Trading Method (NewYork: SignalertCorporation) .

5 Jacob Bernstein, Short Term Trading in Futures (Probus Publishing Company, 1987).

Like the Stochastic, when the fast line crosses the slow line from below, you get a buy. You get a sell signal when the fast line crosses the slow line from above. As much as I like to tweak things to achieve maximum advantage, I have never found any combination of inputs that matches, much less exceeds, those that Jake developed, i.e. .213, .108, .199. These exponential inputs can be simulated by "period" inputs of 8.3897, 17.5185, 9.0503, if the software you are using is programmed to take "period" inputs and simulate exponential Moving Average smoothing. Among those graphics programs I know of which accurately program this study are CQG, Inc., Aspen Graphics, TradeStation®, and our own CIS TRADING PACKAGE. I'm sure others do as well, but I have not confirmed that fact.

If you're a "shoot from the hip" type and think I'm getting just a little too particular about these details, that's your prerogative. I must emphasize what I believe is important. It's your option to choose to ignore what you please. I'm not saying you will lose money if you don't follow this precisely. I am saying that you should know what it is that you are doing and not make unwarranted assumptions. Besides, we need lots of "shoot from the hip" types. These traders often provide us with the other side of our trade.

Okay, let's assume we have properly calculated and displayed the Stochastics and MACD studies (I will refer to the DEMA as the MACD from this point forward for clarity). Here's how I use them.

For my purposes, the MACD is the more reliable Trend indicator. I have continued to use Jake's numbers, thereby leaving it strong. I have deliberately weakened the Stochastic by inputting 8, 3, 3, rather than the stronger 14, 3, 3, originally used by Jake and many others. Above are Charts 5-4 and 5-5 of the Stochastic and MACD for the daily March

U.S. bond contract. Notice that the Stochastic study has a more ragged appearance than the MACD. The flowing lines of the MACD give us a smooth presentation of Trend. This is what we want! The number of buy and sell signals given by the MACD are infrequent compared to the Stochastic. By displaying these two indicators, one on top of the other, in separate windows, as in Chart 5-6, I can evaluate when to fade (go against) a market and when not to.

If both the Stochastic and the MACD are on a buy, the Trend is up. If the weak Stochastic signals a sell, I can buy the dip associated with that sell, as long as the (strong) MACD buy remains intact. Notice the solid up trend shown on the MACD between mid- January and mid-February. We had an excellent buying opportunity against Fibonacci support, as the weak Stochastic went into a sell, then got back in gear to the upside. The same can be said on the sell side for the period from mid-November through mid- December. What is more subtle but nonetheless opportune, is the MACD up trend to the far left of the chart. While casual observation suggests that the Stochastic remained in an up trend, real life trading was just the opposite! Why? During the period from mid- October through the November high, there were many times when the Stochastic signaled unconfirmed sells during the day. We can't see those sells by looking at this chart, because the close of the day is the point at which the (confirmed) indicator is computed and this is what we see on Chart 5-6 when we look back. Traders acting on those unconfirmed intraday signals in real time, however, gave us an opportunity to go long when they sold. We could then take profit at pre-calculated Fibonacci Logical Profit Objectives when the Stochastic went back into the buy mode and their buy stops were hit!

By observing the combination signals on the most frequently used charts, i.e. 5, 30, and 60 minute, daily, weekly, and monthly, we can get a bird's eye view of where the weak players are (Stochastic), as well as the position of the strong players (MACD). My aim is to buy dips (Stochastic sells) at Fibonacci retracement points in an up trend (MACD buy), or to sell rallies (Stochastic buys) at Fibonacci retracement points in a downtrend (MACD sell). In this way, I combine Leading (Fibonacci) and Lagging (MACD/Stochastic) Indicators, in such a way as to "safely" interact with price action. You should also note that the traditional requirement for a Stochastic signal to be at the extremes of 25 or 75 is ignored. Just like the MACD, I require only a crossover of the fast line through the slow line for a signal.

Chart 5-7 shows a five minute S&P in a thrusting down move. It's difficult to see the extent of the thrust, since the bars are confined to only one third of the chart. This was done so I could show you the action of the MACD and Stochastic. Initially, both indicators are signaling a sell. An interim price low is reached and the Stochastic goes into the buy mode. It is bringing in the weak longs and getting rid of the weak shorts. The down trend, as defined by the MACD, remains intact. Observing this type of action will show you how trailing stops set in improper areas give knowledgeable players a perfect opportunity to take away positions from weak players, i.e. to buy a dip or sell a rally in the direction of the prevailing Trend. Chart 5-7 shows how the buy stops are hit, driving the market up to Fibnode resistance. After the up move, the Stochastic gets back in line with the MACD and the market returns to its previous direction, perhaps to new lows. This type of action is repeated again and again in a variety of Time Frame charts. Just be sure you are involved in thrusting markets in order to help avoid possible whipsaws.

In teaching the use of this MACD/Stochastic combination signal, I typically break it up into levels of complexity and teach up to the level possible, depending upon the presentation setting and the experience level of the students. The above explanation involves level 1 (waiting for both indicators to agree before defining a Trend), and level 2 (the concept of fading the weak Stochastic indicator while positioning yourself in the prevailing Trend). Later we will have examples of level 1 & 2 utilization of this technique. Level 3 (anticipating or acting on an unconfirmed signal) will be discussed somewhat, while level 4 involves sliding your Time Frame and is too complex a subject to be adequately covered outside a classroom setting. To give you some idea, however, as in the example above, the half hour Trend would typically be on a sell to further support our Stochastic fade on the five minute chart. There will be more examples after we cover Fibonacci analysis.

Now, let's drop back, and look at this from a different perspective. If you think about this approach and consider the mathematics of the Stochastic, you will see how a market can be made to turn. Consider a large local or, more likely, a group of locals who are short the market. If they can hold prices at a given high for several bars (keep prices from going higher), it will force the (weak) Stochastic to turn south. The weak longs start selling their positions and weak shorts initiate new positions on the sell side. Now the locals (and we) can buy those sell orders. The locals can take their several ticks profit while we can position ourselves for the expected new high, or a move up to a Fibonacci expansion Point. If we tried to buy stop the old highs instead of buying in on the dips, we would be

at a point of high market slippage. Upon being filled we would have to suffer through yet another pullback, while the locals who have fed us the sell orders, endeavor to make a profit. Ifwe buy the Stochastic sell and the MACD ends up breaking (giving a sell as well as the Stochastic), we know we're wrong and we take the next rally out. If we are operating on a short term Time Frame and have adequate experience employing this method, it is possible to break even, pick up a few ticks, or perhaps suffer only a few tick loss, even when we are wrong!

Let's take a look at a relatively simple example on daily crude oil, Chart 5-8.

We obviously have a thrusting, up trending market as defined by the 3X3. If you played this market primarily from the long side on the way up, you'd have done very well. You wouldnot have gotten into trouble by selling point Tl, or by buying points T2 or T3. You wouldn't have made any of these trades, even though Trend rules learned previously would have justified such a play. (We will discuss Directional Indicators which overrule Trend in the next chapter. They would have you buying at Tl, and selling at T2 and T3, which are all at near perfect Fibonacci retracement points. Pardon me for digressing, but a quick look ahead is sometimes useful.) Now, back to the point.

The entire downtrend after the high was contained by the MACD. The rally up to Fibonacci resistance at point T3, supported by the Stochastic, gave us a perfect opportunity to get short.

The same can be said for the up move preceding the high. The up move was almost totally contained by the MACD, while the Stochastic gave us ample opportunities to get long when the market retraced.

FREQUENTLY ASKED QUESTIONS:

Wouldn't it be better to have two strong Trend indicators, instead of one weak, and one strong ?

No. The weak Stochastic shows the hand of the weak players. It can also show the Strength of the market. If the Stochastic gives a sell and there's no discernible movement down in price, watch out for a big move up!

Conservative Carl: Do you wait for the bar (time period) to close before making a determination that the indicator has given a signal?

This question leads us into level 3 understanding. By definition, you want confirmation to be certain of the signal. But, if you wait for confirmation, a significant part of the move may be in place. By waiting, you are paying for insurance you may not need. Just as it is acceptable to anticipate the DMA crossing of price, before the period closed, it is also acceptable to anticipate these Trend signals, prior to the close of the period. Be sure you get confirmation of what you have anticipated by the close, or get out immediately!

/ 've always bought when I got a Stochastic buy. How can I now sell?

If you are going to be among the 5% to 15% winners, you will have to be open to applying methods and procedures that are different from those of the masses. If it were as easy to win as following a Stochastic crossover, where would all the losers come from to pay the floor, as well as the off floor winners?

Hyper Hank: So I buy when I get a sell on the Stochastic and sell when I get a buy on the Stochastic, right?

No, you fade the Stochastic in a thrusting market in the context of a Trend, supported by the MACD. You don't simply BUY or SELL. Then you employ methods of entry, covered in CHAPTERS 8, 9, 10, 11 and 13.

Why don't you use the 25/75 barriers before assessing a buy or sell signal on the Stochastic?

Any crossing of the fast line through the slow line is considered a signal because of the unique way I am using the indicator. It may also be useful for you to note that trading experience (not formal computerized research) indicates that stronger signals, on both the MACD and Stochastic, are typically given if there is a greater angle of attack at the point of crossing. See Charts 5-10 A & B. The stronger appearance generally is indicative of markets that are moving and turning, rather than consolidating.

Trend Analysis MACD/Stochastic Combination 69

Diligent Dan: "In the first example of daily bonds (Chart 5-5), it looked like the MACD barely gave a sell near the end of January and then got right back to the up side. Is there some way we could have avoided being whipped as we saw this action unfold? "

Since this was a daily-based signal, it's likely I would have been whipped as you suggest. It's unlikely I would have wanted to take home a position that was against the MACD. There are, however, ways of avoiding being whipped under similar circumstances. If, for example, the Fibonacci support levels have not been penetrated by price, you could have a trading plan that would accommodate a minor break of the MACD. This would give the MACD an opportunity to correct. I give myself this latitude if the MACD signal is of the weaker variety, as in Chart 5-10B A minor break of the MACD on an intraday chart is also easier to live with than a break on the daily, since you can find out quickly if the position will hold on the anticipated Fibonacci retracement. You may choose not to allow this much discretion in your trading plan, until you gain more experience working with the concept.

SUMMARY:

Let's summarize the high points of our second Trend analysis tool, the MACD/Stochastic combination.

* Both the MACD and Stochastic give Trend signals when the fast line penetrates the slow line. These signals remain intact until another penetration occurs. The signal is confirmed at the close of the period.

•  The MACD/Stochastic combination is applicable for all Time Frames we use.

•  The values I use for the MACD (DEMA) are .213, .108, .199.

•  The values I use for the deliberately weakened Preferred Stochastic are 8, 3, 3.

•  To get good results using these indicators, you need to investigate the formulas used to
create the studies involved as well as the method of programming the mathematical inputs
into these formulas. The graphics software used to display the charts should show
market-aligned bars, not time-aligned bars.

•  By utilizing a specific weak Stochastic and a specific strong MACD, we are able to
make informed judgments about what the weak and strong players are doing.
Consequently we can determine how to best interact with price to achieve our goals.