Seeking Alpha
, QVM Group (90 clicks)
Profile| Send Message|
( followers)  

Visually inspecting charts is essential to “get the picture”, but using a computer to quantitatively screen for ideas that can be reviewed in depth, one-by-one is very helpful. It allows you to work with a large universe of possibilities that the computer can narrow down to a manageable list.

Just as financial fundamentals and valuation can be quantitatively screened using computers, so too can charts. While ordinary computers cannot pick up the near infinite variety of information that the human eye can see, they are much better than humans at testing large numbers of charts for a few specific quantitative features. Computers are millions of times faster, are totally objective, and don’t make mistakes.

As long as you can specify a quantitative attribute, the computer can tell you if that attribute is there or not. From the narrowed list of charts that meet your few quantitative tests, you can further reduce the list with visual inspection to ordinate and triage them. Price action alone is insufficient. You also need to review the fundamental strengths and weaknesses, the valuation, the business story and any operating or market risks you may perceive.

You can generate your short list of potential investment ideas by finding attractive chart patterns, then culling those that pass using fundamental and valuation criteria. Or, you can generate your short list by first doing fundamental or valuation screens followed by narrowing based on price action.

Either way you should attempt a fusion between fundamental and price action factors.

In a recent article, we explored the utility of the 52-week and 26-week averages to identify multi-year cycles in the S&P 500 — where the 26-week average crossing the 52-week average serves as signal for trend changes from up to down, or down to up, cycles. This is the illustrative chart:

click image to enlarge (1640 pix wide)

52-week/26-week SMA crossings


Others have done studies suggesting that the 200-day average and the most recent closing price provide a pretty good indication of up and down cycles. We think that gives too many signals to be useful to investors, versus traders who are more actively in and out of positions.

You need some other criteria (such as crossing the 200-day average by some percentage, or to stay on the new side for several periods) to filter out bad signals, all of which illustrates the advantage of using the relative position of a long and a short moving average as a signal for “quiet” money.

40-week (200-day) SMA/price crossings


Using the 40-week/20-week combination creates more (earlier) signals than the 52-week/26-week combination, but avoids most of the frequent and false signals of the 40-week/price approach.

40-week/20-week crossings


If you become strenuous for long position entry, and require that the longer average be flat or sloping up, as well as that the shorter average is above the longer average, and that the price is at or above the moving averages, you get very few signals, which is good when seeking multi-year trends. And if, you use the reverse of the criteria for exits, you get few exits (at least in the past 20 years).

40-week/20-week | upward slopes | P > averages


The green shows entry signals and the red shows exit signals (5 in and 5 out in 20 years ). There would have been times during the up trend when you were long were you would probably have been stopped out, but you would then have waited for the price to resume its above moving average relationship and re-entered. Presumably you would have been in cash or bonds during the down legs (we’re not considering shorting in this discussion) so that your assets would grow while stocks you would not own were shrinking.

By this more strenuous method, today’s S&P 500 index is not yet in an uptrend.

For those with more of a trading mentality, or that seek to enter closer to the turn from down to up to capture more of a new trend move (with greater chance of false signals and whipsaw loses), shorter moving averages might be appropriate. For our own purposes, we prefer the longer-term averages and the long price runs they signal.

Of course, there is no certainty that other indexes on bonds and foreign assets, or individual stocks will behave as nicely with this kind of trend signal as the S&P 500 did, but the broader the index, the more likely it behave similarly, we would think.

In all cases using moving averages, there is a lag, resulting in you joining an up move a bit late and exiting a down move late too, but moving averages are trend followers, which makes long-term averages good friends when seeking multi-year rides.

Our last article reviewed several key asset category funds, country funds and bond funds in terms of the relationship of their 40-week average (200-days), 20-week average, and the most recent closing price, to see which might be indicating intermediate-term upward or downward trending.

We did that by eye, one at a time, since there were only a couple of dozen. However, there are over 800 funds and over 8,000 stocks which makes timely visual inspection of them all impractical. So, we programmed a simple computer screen to review all ETFs and all US stocks using the same criteria.

We asked the computer to find ETFs with minimum assets and trading volume, where the longer and shorter moving average had recent positive (upward slopes), where the shorter average was above the longer average, and where the most recent closing price was at least 95% of each of those moving averages.

We ran the filter using three different sets of averages: 52 and 26 weeks; 40 and 20 weeks; and 26 and 13 weeks. The test found 270 non-short ETFs with at least $1 million of daily trading volume and at least $100 million in market-cap as our search universe. Of those 270, we had these price action results:

Search Results:

For the 52-week/26-week screen, we found 21 of 270 ETFs passed. All were bonds funds, except for a bio-tech fund and gold funds.

52-26_2009-04-28

For the 40-week/20-week screen, we found 17 of 27o ETFs passed. All were bond funds.

40-20_2009-04-28

For the 26-week/13-week screen, we found 97 of 270 ETFs passed. Of those, 16 were bond funds and 81 were non-bond funds (stock funds, currency funds, and commodity funds).

Conservative investors with high focus on capital preservation and loss avoidance, may find the 52/26 or 40/20 screen most helpful. Aggressive investors (really traders) willing to enter and exit more frequently, may find the 26/13 screen provides useful information.

Whatever method you use to find opportunities; price action, fundamentals or valuation; persistent trailing stop losses (using percentages) are an important way to cut losses while letting profits run — particularly in these times of high volatility and government policy twists and turns.

Source: Screening for Trends