Thursday, September 14, 2017

Unmasking the VooDoo: Guru Calls

I originally wrote this in August 1998 when I covered technical analysis for BridgeNews to explain what a big stock bull saw in the charts to suddenly turn him bearish. The names have NOT been changed because the call turned out to be pretty darn good.

What did that mean? 

Why a big bull turned into a big bear

This week, mega-stock bull Ralph Acampora, chief technician for Prudential Securities, did something that most people on the Street thought would never happen. He issued a sell signal. Shudders went though the masses. If this guy, who was famous for predicting Dow 7000 when the Dow was 1000s of points away and who clung to his Dow 10,000 forecast for 1998 before last week's fireworks, can turn sour on stocks, what will the itchy trigger fingers of the masses do?

Let's examine what really happened from a charting standpoint. First, Acampora follows trends and patterns in stock market prices. This makes him different from that other mega-stock bull Abby Joseph Cohen, chief market strategist for Goldman Sachs, who leans more heavily on fundamental analysis. We're not going to say here which is better nor who is right. As a column that attempts to debunk the mystery associated with the Acampora side, technical analysis, we'll just focus on what Ralph saw.

There are basically two keys to his new bearish stance -- market breadth and market trends. The former measures how many stocks are going up and how many are going down. It measures how much of each day's volume is attributable to each of these two groups. In other words, is the rising tide carrying most of the boats higher? If it's not, too many of the boats have sprung leaks.

The latter concept, trend, is subject to how the stock market is sliced and diced. There are blue chip stock indices, such as the Dow Jones Industrial average, that show the "market" fairly flat over the past few months. There are indices of big stocks, such as the S&P 100, that have been rising nicely. There are indices of small stocks, such as the Russell 2000, that peaked back in April and are down 30% already. Which one of these really is the market?

The answer is that none of them on their own fit the bill. Perhaps an unweighted index, such as the Value Line Index, which gives its approximately 1800 components an equal voice, is the answer. If it is, we've been in a bear market since April.

Acampora, as well as a whole host of technicians, saw that market breadth had deteriorated badly. More stocks were going down than up. More volume was attributable to declining issues that to rising issues. More stocks were setting new 52-week lows than 52-week highs. More S&P industry groups were lagging the S&P 500 than were leading it. This is curious in that we would expect the number above to equal the number below. The industry groups together are the S&P 500.

Again, Acampora was not alone in noticing that the small stocks were going down while the big stocks were going up. Stocks with Asian exposure, such as the technology sector, were sliding since the Asian stock meltdown last autumn. True, the tech heavy NASDAQ Composite was climbing but when we examine it more closely, we see that only four stocks account for nearly half of the index's movement. That's not the market, for sure.

The 299 point sell-off we had last week was underway well before Ralph said a word. No, comments from somebody famous did not help matters but the bottom line is that the stock market was already on shaky ground.

Let's now look into Acampora's new downside target -- 7400. This is 1200 points away from where we are today. First, the jargon. If a correction is defined as a 10% drop, we've already had one. A bear market takes us to a 20% drop so that's where Ralph gets his wording.

Second, how did he get his 7400 number? You can bet he didn't just pull it out of his hat. The Dow 7500 level was the arguable bottom of the July 1997 - January 1998 trading range. It served as the starting point for the rally that broke free from the range and lies at one of the key percentages to which markets often fall back during declines. For those of you so inclined, it is the 38.2% Fibonacci retracement of the rally that began in late 1994.

We cannot know today whether we will see 7400. More conservative technicians will wait to see if shorter-term technical levels are violated before moving into the bearish camp. However, the moral here is that we need to question our indicators, including our definition of the market or what our advisors are saying. Market breadth and small stocks were flashing warnings a long time before the Dow got smacked.

Acampora still looks for the magic Dow 10,000 in a few months. For long-term investors, he does not sound like such as bear after all.

Friday, September 1, 2017

Unmasking the VooDoo: Fibonacci

If you ever want to sound foolish during your next television interview on the markets, tell them that the market is approaching a Fibonacci level and will do X when it gets there.  It is sort of like doing a DIY network interview and saying the house will have four bedrooms because you are using a 32 ounce hammer.

Both the hammer and Fibonacci are tools used in their respective fields. A hammer is used in carpentry to, among other things, frame a house. A Fibonacci - usually a Fibonacci retracement - is used in investment analysis to frame the market's recent move.

Obligatory definitions

Here is the section where we have to define what Fibonacci is and we'll focus on Fibonacci retracements. If you have any interest in who Mr. Fibonacci was, feel free to fire up the Google. It is irrelevant to what we are doing here.

Mr. F's claim to fame was discovering a mathematical sequence that seems to describe a lot of geometry in the real world from how petals spiral on a flower to how stars spiral in a galaxy. And from that sequence we get the "golden ratio" or "divine proportion" which somehow describes pleasing and useful relationships in art, music, architecture and design in general.

The ratio of .618 to 1 comes from a sequence of numbers where any number is the sum of the two that preceded it. Here is it in all its glory:

0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, 4181, 6765, 10946, 17711, 28657, 46368, 75025, 121393, 196418, 317811, ...

As you can see, the numbers start to get huge. Fortunately, we rarely encounter any above 144.

And even more fortunate, the ratio of any two consecutive numbers settles on .618 fairly quickly. As follows:

0, 1, 0.5, 0.666667, 0.6, 0.625, 0.615385, 0.619048, 0.617647, 0.618182, 0.617978, 0.618056, 0.618026, 0.618037, 0.618033

First, this is the ratio calculated out to the number 610 in the sequence.

Second, and more important, you can see how the ratio oscillates higher and lower as it hones in on the .618 value. Any ratio beyond what is shown above is pretty much just rounding error away from that value.

Voila! the golden ratio!

Fun for math nerds
The ratio has all sorts of neet-o properties. For example, 1 / .618 = 1.618.

And 1 - .618 = .382.  Guess what the square root of .382 is? That's right, .618.

.382 / .618 =  .618

You have probably seen these two percentages on the charts an they are roughly equal to the simple percentages traders have used for years, namely 1/3 and 2/3 (or .333 and .667).

Guess what is not a Fibonacci number. 50% or .5.  Check out the sequence. It is nowhere to be found. And you cannot concoct it from math-ing 0, 1, and .618, either.

Still, traders do attach meaning to a 50% retracement or pullback and it has wormed its way into analysis using Fibonacci. Don't fight it, exploit it.

Using this stuff

Now is when we get to dispense with the formalities and use the name experience chart watchers use - fibos. Don't confuse this with fibbie, which is the FBI. Or fubar, which was what you may be feeling right now trying to digest all this fibo stuff.

Fire up any charting software package and under the drawing tools section you will see Fibonacci retracements and possibly Fibonacci extensions, time and arcs. They are all just different ways to look at two points on a chart - usually a high and a low - and then draw stuff off of them at specific percentages.

Take a look at this chart of the S&P 500 from 2007 to 2011. Using the basic set of fibo retracements of 38.2%, 50% and 61.8% look how nicely we could project where support and resistance might be.

Here is where you can get into trouble. These levels are likely places to find support and resistance levels but not always. And sometimes you might be tempted to add more fibo lines until you find one that fits.

My advice is to never add more percentages except for one... and it is a bizarre one. We'll get to that later.

Remember, we are only trying to frame the market, not nail all of its turns and accelerations. We'll use other tools to attack that part.

Anyway, we can see that the rally peaked at the 61.8% retracement and the ensuing correction stopped at the 38.2% retracement. Whoopee! It worked once. We can even see the market stumble in the vicinity of the 50% level.

Let's repeat this. Fibos are likely place for turns, not guarantees.  But we can beef them up a bit by measuring fibos from other points and see if they overlap, or cluster. If they do, then we have a better support or resistance target.

Check out how the rally from the 2010 pullback low to the highs of 2011 is 161.8% of the pullback itself. Something usable must be going on here.

Even better, what if moving averages hit those fibos? Or normal horizontal support or resistance? Or trendlines?

Getting the drift? As with any tool, the more things that point to a price level the more likely that price level will be important.

Now for the bizarre.  Some people use a 78.6% retracement level. Note how this is kind of close to 75% but that is not what makes it bizarre. It .786 is the square root of .618.

In my book, if a market can retrace this much of a previous trend, it is rather likely to retrace the whole thing.

The other fibo stuff

We touched on extensions, time and arcs so let's put in the quick definitions. A fibo extension is a number bigger than one and it projects where a market can go. Look for such numbers as 1.382, 1.618, 2.618, 3.618, etc.....

Draw an arc centered on the origin for each ratio (usually the original ones less than 21) to find trend targets.

And instead of measuring price, do it on the time axis. Think about looking for a 61.8% price retracement that took 61.8% of the time of the previous trend to happen. Make a fibo box out of it.

Fibos are also used in Gann and Elliott Wave analyses. For the latter, we like to see Elliott Wave 3 to be 1.618 or 2.618 the length of Wave 1.  That is a whole nother can of worms so we'll end this right here.

Use fibos on your charts but please use other things along with them. They can help you frame the market but like a hammer, you will have to decide where to put the door and windows. Fibos do not dictate where you buy and sell.