Friday, August 25, 2017

Unmasking the VooDoo: Sentiment


Usually, when I discuss the major parts of technical analysis I'll use an analogy of a table. Price is the top of the table and the most important part. The legs are volume, time and sentiment. They support price analysis but do not supplant it.

Everyone gets price from patterns to trends to momentum. Everyone get volume. Most people get the concept of time, too, which includes such things as cycles, analogs and some esoteric stuff such as Gann.

But sentiment is often the red-headed stepchild of the analysis. Sentiment is hard to gauge. It is hard to quantify. And it is usually fairly hard to gather. Just look at your charting software list of studies. You'll see RSI and cumulative volume. You might even see an Elliott Wave "recognizer" but you are not likely to see something called sentiment.

These things do exist but they are not derived from trading as all the other parts of the table. Sentiment is the true oddball of technical analysis because it is the only analysis that seeks to figure out what traders are thinking. It's not like everyone inputs their feelings into an app on their screens and sends it off for tabulation at a vendor.

Yes, I know there are surveys that do just that but for every collector of sentiment there is a different method and interpretation.

And there are many, many different types of surveys and proprietary indices. There are surveys of traders, of investors, of newsletter writers. Then there are surveys of the actions of traders, which are different from surveys of what traders think. There are surveys of money managers and of Wall Street strategist equity allocations.

There are a few oddball things such as the trend of carpenter tool prices and railroad car usage. Most people may have heard of the trend in oil rig counts.

Have you heard of the hemline indicator? They get shorter when people feel feisty and good.

How about the magazine cover indicator? This is a real thing and it is actually pretty good.

The ratio of football revenue to baseball revenue? Football is a darker sport. Baseball seems to do well when people are happy. Just ask George Carlin what he thinks about these two.

How about horror movies vs. romantic comedies? Same logic.

Giving credit where it is due, those last two come from socionomics, which looks at social mood as an indicator of the stock market. It actually works in reverse but they are closely enough correlated for this discussion.

And if that were not enough, there are a few things that are generated from trading including the famous VIX index, which is concocted from options prices. There is the ratio of puts to calls. Of put volume to call volume. Of equity only puts and calls. Of index puts and calls.

I supposed if we go there it is not far to include the number of stocks above their 200-day averages and at new 52-week highs or lows.

Have I convinced you that sentiment is wild and whacky? And hard to quantify? And hard to even obtain the data?

Using Sentiment

Let's assume you have access to some sentiment data or perhaps just have eyes and ears to follow the news.  What do you do with it?

First of all you have to understand that sentiment is a contrary indicator. And most of the time it tells you nothing useful. If the market is going up then sentiment should be happy.

It is similar to the trend is your friend until the end. Most of the time, the crowd is actually right and you follow along. Only at the extremes is it wrong and you need to go against it. You know, buy when there is blood in the streets. Sell when E*Trade runs ads with cab drivers buying islands. Or when the local pizza shop owner brags about how many houses he flipped.

Remember way back to 2008 when the market was going to hell in a hand basket? It was a brutal bear market and the government was panicking to do something - anything - to look like they were able to stem the tide. My father-in-law, rest his soul, called me to sell everything in his portfolio. I sold a little so he would not be mad at me but I kept the rest.

When the public gets so scared that they throw in the towel you know sentiment is at an extreme for bearishness. In fact, the American Association of Individual Investors survey was at a record high for bearishness. Something like 70% vs. its historical 28% (I am just going from memory on those stats).

Take a look at what was on the cover of Time Magazine on March 9, 2009. It said the economy was holding on for dear life. it talked about foreclosures, about nationalizing banks, toxic mortgages and your disappearing savings.

Guess what day saw the absolute lowest close on the Standard & Poor's 500 index? That's right March, 9 2009. It nailed the bottom.

Usually, there is a lag in these things so don't obsess over this perfection.

Why does this work? When editors or general interest media take the pulse of their readers and finally believe that readers want to read this stuff then we know sentiment is very one sided.  You may rag on Barron's covers but Barron's is not general interest. It's covers are driven by professional sentiment, not the general public.

The same thing works at tops. Think tulip bulbs. Bitcoin. And right now, some think Amazon.com. None of these markets, save for Amazon, had any connection with the real world. They were the greater fool, burning match theory in motion. They went up because of FOMO. You didn't think I could schtup in a millennial reference, did you - fear of missing out. Go buy some etherium. YOLO.

When an actual survey goes extreme, you can bet that whatever trend was in place is nearing its end. But remember, sentiment is not a trading trigger. It sets the environment for that trigger to be really meaningful.

If you can find a stock with a reversal pattern or some other technical positive in place and sentiment is awful - like SNAP was in mid-August - it might be ripe for at least a trade if not investment. Likewise, when all the Amazon haters finally acquiesce (i.e. shut up) then we can consider the stock is too high.

Friday, August 11, 2017

Unmasking the VooDoo: Moving Averages


As with all technical indicators, moving averages are just tools chartists use to figure out what they need to do in the markets. Notice I said they are figuring out their own actions, not divining where the market is going. Of course, the two are related but that is not the point.

What is the point is that pundits and journalists like to report when the market breaks one of the more widely followed moving averages. In this edition of Unmasking I am going to take a look at what a move through an average - or a "crossover" - really means.

Let's start with just what a moving average is. If you want the full definition feel free to fire up the google, the bing or even (shudder) the yahoo. And if you know what a moving average is, feel free to skip the next two paragraphs.


A moving average, or average, is just an mean price over a certain period of time. A 50-day average is the sum of all prices - usually the close prices - divided by 50.  Easy peasy.  Since the calculation is redone fresh each period, the value window of data "moves" over time. Get it? Moving average!


Yes, we can change the span, period, input type and even the weighting of the calculation using any number of oddball formulas but again, if you want to learn about that, it's back to the google for you.

So what is the big deal with these things? Do they create trade triggers? Which ones do we use?

To answer the last two questions: maybe and whatever floats your boat. Yeah, not much help but these are answers you need to develop on your own. I'll just tell you what I do with them.

And to the first question, the big deal is that they help us see the forest for the trees. They smooth out jumpy price action to let you see the major and minor trends. You change the parameters depending on what you are trying to do.

You do realize that the government actually uses a moving average on some of its big monthly economic reports, right? Why? They say it smooths things out. Score one for the chartists.

You can use averages in pairs or even triplets, garnering insights as the averages dance around each other and with the price plot. But let's KISS (keep it simple, schmegegge).

Let's talk about the 50-day moving average. Yesterday, August 10, 2017, the S&P 500 and Nasdaq both closed below their 50-day averages for the first time since July (June for the Naz). The financial headlines were quick to point that out and with the news of the day - a potential nuke crisis in North Korea - boy did they make a big deal. Gotta build the click through count.

Chicken Little was ready to sell everything and head for the hills. Now, can we take a peak at the headline from July 6, 2017 when the S&P 500 last closed below this red line in the sand?
July 6 - Nasdaq challenges major support amid geopolitical tensions, S&P 500 nails 50-day average
In the words of  that famous portfolio manager Homer Simpson, "Doh!"

Six days later, the Spoos reached another new all-time high. Talk about a crappy sell signal!

Let's look at the charts. As you can see, price action violated the average on four separate occasions before the current event since the November pre-election low. The 50-day average did not deliver a successful sell signal one time. None. Zilch.



That does not mean that a drop below the average is always meaningless but I'd rather look for something more reliable, such as a support break, to tell me to sell. Rather, I'd use breaks of the average to set up or confirm other signals.

So again, what is the point? The point is that averages help us figure out the trend. Just look at their name - average. Let's expand the chart all the way back to the start of the major rally in early 2016.


I added the trend channel with a little technician's license to make the upper line more meaningful. That means it touches price action more times.

Look at the 50-day average. It more or less runs through the middle of the channel. Price goes above and below it yet it still was a bull market. The average itself spent most of the time rising.

The next question about that time when it was not rising naturally will be, "How do you know it will go back up?" Well, you don't but there are other tools to use, such as the popular 200-day average and any number of indicators.

The point is that in a bull market being above the average is good but being below it is not necessarily bad. In a bear market, being below is good (for the bears) but being above is not necessarily bad (for the bears).

So, slap on your averages, short-, medium- or long-term, simple, exponential or VIDYA, and find out what the current trend is. However, selling just because price dipped below your average is not always a good idea.