Saturday, February 5, 2011

February 4th Weekend Update

There is a combined GGT and GGT, LLC meeting this coming Saturday, February 12th, at 10:00 a.m. to 1:00 p.m. at the Burke Centre Library, 5935 Freds Oak Road, Burke, VA  22015.  Please RSVP in one of the GGT Yahoo! forums so that I can print an appropriate number of handouts.  I anticipate that there will be a Web-Ex that remote folks can call into and listen to the dialog as well as view the slides in real-time.
  • GGT LCR System indicators are bullish 
  • GGT Price System indicators are bullish
  • A new indicator, the LCR Summation Index (described below in detail), is on a bullish run and we can't ignore the strength
  • A new indicator, the Pricing System Summation Change (PSAC) value (described below in detail), is indicating that entry into stocks on Monday is probably a riskier trade.
  • Elder FI(13), applied to the GGT system, is bullish
  • We can use GGT New Longs and Effective Volume ( to identify candidates that are experiencing institutional buying on the day of recommendation

The LCR System

The LCR is an abbreviation for Long-Cash Ratio, and it is the number of stocks in the GGT database that have a LONG rating divided by the number of stocks that have a CASH rating.  We'll deal with how a stock gets a LONG or CASH rating in another entry sometime in the near future.  While the absolute value of the ratio is important (e.g, is it 0.5, 1.23, 2.6 , etc.) the trend of the movement of this number is more telling.

Applying moving averages of various lengths to this ratio gives us insight as to what the short-term and intermediate-termed time frames are doing.  I apply 5d, 8d, 13d, 21d, 34d, 55d, and 65d averages to the LCR value, then I take the difference on a day-over-day basis to determine the slope of the moving average.  The result is the ability to produce the following table:

The left side of the table shows the slopes of various moving averages, while the right side shows the "slopes of the slopes".  As with all my images, right-click on the table to open in a new tab or window.

Slopes are an important indicator that goes largely ignored in the market, except the MACD (Moving Average Convergence-Divergence) indicator, which implicitly shows you the difference of two EMAs, effectively presenting the slope changes in histogram format.  I know, not exactly the same thing as what I'm doing, but close enough to get you in the same frame of mind.  Slopes are good indicators of direction of changes -- if the slope is positive in value then it simply means that the day-over-day change is positive.  If the  slope is negative in value then it simply means that the day-over-day change is negative. Moving averages smooth this effect, and this can be useful.  This is reflected above on the left side by the red and green blocks -- where they are green "bullish" we have those slopes reporting in as positive, and where they are red "bearish" we have negative slopes.

As you can see by the staircase green "bullish" pattern in the lower left of the table, we are becoming more bullish as far as the database is concerned, at least on a 5d through 21d time frame.   The implications of this are significant -- the database of stocks that GGT follows (all above $1, all above $700Kshares in dollar*volume, all on the three primary exchanges, or about 2600 in total) is expanding in terms of price and volume participation.  Money is explicitly flowing into the markets, and this is reflected in prices moving up on higher volume as demand increases, which are necessary requirements for a stock to get a LONG rating.  If you believe (as I do) that market behavior is at least 20% of a stock movement (if not more), then our chances improve when the database is expanding, e.g., we see more green "Bullish" labels appearing on the left side of this table.  Remember, this table is not of price, but of the number of stocks that have a LONG rating to those that have a CASH rating.  The distinction is important.

The right side of the table reflects "slope of the slope", or in essence, it answers the question "is the slope pointing upward or downward?".  This isn't the same thing as having a positive slope or negative slope (why?  make sure you understand this), because we can have a stock that has a positive slope (LCR System: left side green "Bullish") but day-over-day, is slowing in momentum (slope of slope is pointing downward, LCR System: right side red "Bearish").   Here's a detailed view of the 34d moving average as well as the 34d MA slope:

In the figure above I've shown two periods where the 34d EMA peaked in value, and the slope correspondingly is crossing through 0 at these point (daily change is virtually 0 when it peaks, right?).  Then, as the 34d EMA trends downward, the slope is correspondingly moving more and more negative (dropping deeper in the pink zone).  Pay specific attention to how the slope bottoms out long before the 34d EMA bottoms out ... slopes are a leading indicator of the moving averages they track.

Compare the figure above to the LCR System table.  Scan down the 34d Slope column, identifying where the table moves from red to green and back.  You will see the correlation in the figure above -- when the 34d slope transitions from the pink area (slope was negative and now newly positive), the LCR System table moves from red "Bearish" to green "Bullish".  The converse is true too -- when the LCR System table moves from green "Bullish" to red "Bearish" the 34d Slope line above is transitioning from the white area to the pink area.

By now you've probably noted that the 34d Slope column in the LCR System table is red "Bearish".  If you look at the figure above you see that the slope line is in the pink zone -- it is also correlating that the LCR System table value is negative (below 0).  Note though that the 34d Slope line is pointing upward, and has been for the last 5 data points.  Now, go back to the right side of the LCR System table and scan down the RIGHT side where it says 34d Slope -- count how many green "Bullish" blocks you see -- the quantity should be 5.  So, the "slope of the slope" is clearly bullish as it is pointing upward, but because the LCR moving average is dropping day-over-day, everything is in the pink zone.  The positive "slope of the slope" for the past 5 days is a necessary condition for the database turning around -- we need continued "green" on the right side of the table to see the number of stocks with a LONG rating move upward.

The real question is how to apply this information to our trading.  Here are some general thoughts:

  • When the slope value hits a maximum negative value (e.g., is deep in the pink zone or deeply negative), as it reverses it is showing the market's reluctance to drive prices further down -- day over day fewer stocks are moving to CASH status.  The values around November 18th represent this description.  This doesn't mean that prices (or the LCR) can't trend horizontally for some time -- they certainly can -- but it means that the acceleration to the downside has at least stabilized.
  • Ideally, as this value trends upward from this maximally negative slope value, we want to enter new positions IF the trend starts and continues moving upward.  Psychologically, this is incredibly hard to do, because we've been in a sell-off mode, and for the most part, with November 18th as a prime example, the markets were incredibly uncertain.
  • A confirmation signal of entering the market is when the slope line crosses from the pink zone into the white area, or for you mathematically inclined, a line tangential to the moving average curve has just transitioned from pointing down to one that is upward pointing.  Rather than show a whole bunch of charts on a daily basis, I summarize this using the LCR System table that I presented above.
  • Your eye can certainly see that there was more red "Bearish" periods on the right side of the LCR System table during the months of December and January than the latter time frames, where there is more green "Bullish" periods.  We simply desire to "jump on the train" when there are more green "Bullish" periods than red "Bearish", relative to the recent past.

With respect to the last bullet, it's possible to quantify this visual concept by running a simple accumulator.  Here's the results of what your eye sees:

As you can see, there are multiple "waves" to consider, so to simplify this simply look at the column to the right labeled with the sigma (summation) symbol.  This column is constructed by taking the various accumulators of the 5d through the 65d and adding them, left to right per day, based upon the premise that if we're in an uptrend on multiple time frames, the summation column should grow and confirm this.

As the summation column increases from some local minima upward towards or in positive values we have some confidence that we should be entering the market on the long side.   Hence, incredibly negative values indicate oversold areas, and once we start moving out of those oversold areas (e.g., 1/12/11 and 1/25/11) we have some confidence that the markets are in our favor.  Looking back at past data, "all in" signals have worked well when the summation indicator drops below -12 or so, although it obviously takes nerves of steel to do this because there are generally multiple tests of these lows when the market is bearish.  Note that the lowest summation value ever recorded in 2.5 years of data is -26 and the highest is +108.
[ Unfortunately, I was traveling all week, so I couldn't participate as much as I wanted to.  Our GGT, LLC fund is about 33% invested, and my personal Top25 ETF fund is 100% invested.  If the sigma column keeps growing I'll continue to add positions. ]

Going forward, we're solidly green across the board, and the accumulator has been increasing steadily like a horse out of the stable, so we cannot ignore what the LCR indicators are telling us.  I note with caution that the LCR Summation Value has increased 5 days consecutively running, and while there are instances of 6 and 7 days of consecutive increases, these are rare.  Note that again, in 2.5 years of data, if we include 0's as being "positive", then we've had  three instances of 10 consecutive days of increases (6/7/10 through 6/18/10, 2/28/10 through 2/22/10 and 10/29/09 through 11/11/09) and only one instance of 9 days of consecutive losses (1/7/09 through 1/20/09).  

To see where we are in this run compared to other runs of duration 9 days down to +10 days up, the following chart gives the number of days with the accumulator being consecutively driven into the ground (decreasing accumulator day after day) as a negative value, so -4 is 4 days of being hammered, and +5 days is 5 consecutive days of increasing momentum before a negative day intervened:

The higher incidence of stepping back for one day (-1 = 48 occurrences) is simply that after running upward, the markets take a day to lock in gains, then we resume.  We're at +5 which has occurred exactly 10 times in 2.5 years, and you can see, the likelihood of going to 6 without a reset is less.  Certainly, it can happen (news events, etc.), but I would expect a pullback to lock in gains here any day now.  Of particular note is that when we hit two or three days of successive hits in this indicator, chances are it's moving to the upside, at least for 1 day.  Although I've not tested it, I think that Larry Connor's strategies could work well with this indicator.


The Pricing System

The pricing system is organized much like the LCR System, in that we:

  1. take the average price of the database (equal-weighted), 
  2. apply a series of moving averages to the price series (5d - 65d), then
  3. take the difference of each day's moving average to determine slope.  
  4. We then take the day-to-day difference of the slope values to get the direction of the slope, or as I call it, the "slope of the slope".
The result is the same tabular presentation format as the LCR System, but now we have pricing information:

Lotsa green on the slope side (left).  Prices clearly have been in an uptrend for a long time.  In fact, look at the 65d slope column -- solid green "Bullish" since 11/16/10.  QE2, whatever, we're in an uptrend and there is no argument to this.  "Don't fight the tape", as the expression goes.

The left side of the Pricing System table shows whether the respective moving average is pointing upward (gaining price at a faster rate day-over-day) or pointing downward (losing price at a faster rate day-over-day).  We've been in a choppy market as of late, with prices seeming bouncing back and forth.  Here's a way to tackle the noise:

Just like in the LCR System, we can look at the slope of the slope area as "when is it a good time to buy", or better, "when is it NOT a good time?"   Take a look at the following graph:

The figure above plots the GGT price index (right axis, red line) along with the daily changes in the Price System accumulator.  Starting at the lower left of the figure, work your way up the GGT price index line, noting when the GGT price index line is jumping upward, as well as when it is falling.  Now correlate this with the changes in the blue Price System Accumulator Change line ...

What you should observe is the following:

  • When the pricing system accumulator change (PSAC) falls into the green zone, prices have pulled back and this presents a good area to purchase stocks.
  • When the PSAC fails to fall into the green zone but does fall, GGT price has been more-or-less stable/horizontal. 
  • When the PSAC peaks at a value over +5 or so, the value of the GGT price index has fallen within a short period of time, which would put pressure on any purchases made at this time.
It seems that there are good buying days and bad buying days, with the bad buying days being at the top of the PSAC cycle and the best buying days at the bottom of the cycle.

We have just come up off a Wed/Thur sequence of being at -14 and -10 respectively, and we finished Friday at a value of +2.  Monday is probably not the day to purchase stocks.

I'll start including this value in my summary when I blog, as I think it can give a significant edge to our efforts.  It seems to me that if we restrict our purchases to those days where we are in the green zone, we stand a better chance of the trades moving upward.

For those of you interested in backtesting, simply send me an email and I'll send you the dates in Excel format where purchases were "authorized", e.g., when the PSAC was below -5.


Contra ETF Watch

Contras continue to look ugly overall.  Both Bull and Bear Power (Elder creations) are negative, which tells us that the Bears for contras (or Bulls for long equities) are fully in control.  Swimming against the current is never advised, hence I'm not going to say much more until we have some strength appear in the contras that I watch (82 in all).

This being said, give a close look to the Bond Bears (SMB, TBF, TMV, TBT, PST, TYO, etc.).  They continue to outperform all other contra ETFs, and if interest rates move up due to inflation or otherwise, they will do well.  I'm thinking of pulling the bonds out of my internal index calculations because these simply march to the beat of a different drummer.


GGT New Long ETFs for Monday, February 7th

A question recently came in about how to select New Longs and the order placed in the sort on the GGT "By Recommendation" page, posted every evening by my colleague Joe A.  On this particular sheet ETFs are sorted first by recommendation (New Long, Affirmed Long, Long, Cash, Affirmed Cash, New Cash), then are sorted GPB, or "Gain Per Buy" descending.  This simply applies GGT to those equities, and then sorts them descending from the highest gain per signal to lowest.

Yes, discarding anything below 0% is a good first step.  Additionally, removing any GG vs. BH (Greek God vs. Buy-Hold) that is below 0% is a good idea too.  Note that these are only valid if you strictly adhere to GGT buy/sell rules, which many/most of you do not.

So, let's step back and see how the "By Recommendation" sheet could be useful to you, simply with the primary sort of "By Recommendation" (hence the name).

First of all, let's recap how a stock gets a "New Long" recommendation in the first place:

  1. Behind the scenes, a computer is cranking away 24/7, literally, looking at what combination of moving averages to price and volume, in combination with rates of change of price, maximize an equity curve,  But rather than pick the peak, the program "dithers" each of the values once it finds a solution, and backs off the "peak curve fit" in the direction of those values that produce the minimum change.

    Think of it this way:  the program basically finds the peak of the mountain, but then slices off the top of the mountain to create a flat plateau.  You can move the magic numbers a bit in any direction and the value of the equity curve will not move (much).

    The lookback period is typically 2 years, and it is applied as a sliding window.  This means that if I calculate the same equity two consecutive days the data will almost be equivalent, but will have a slight difference, due to a day being added and the oldest day falling off the record.
  2. Once the magic values are found, they are loaded into the spread sheets every Friday.  These typically appear in the "Control Panel" in columns R-W, header names EMA1, EMA2, Volume, ROC1, ROC2, ROC3.
  3. IF today's price and volume action are above this magic combination of numbers, the stock will most certainly be some form of LONG status.  If yesterday found it below these magic values and today above, THEN, and only then, is it assigned as a "New Long".

    Hence, "New Long" equities have price and volume appreciation, relative to past periods that gave good equity curves.  
Now, think about this.  What is Effective Volume? It's price changes and volume, and we like it when price changes and volume start increasing.  What this means is that it's quite possible that New Longs can feed a screen based on EV, and this concept is exactly what I've been evaluating for the past few months.

Here are GGT New Long ETFs, filtered through an EV commentary:

  • IWP, the iShares Russell Midcap Growth ETF, has seen some fairly significant buying, especially near the end of Friday.  A whole series of 25K blocks (about $1.4M per traunch) around 3:45 drove Large Effective Volume (LEV) upward by the end of the day, along with the price.
  • It's hard to ignore the price movement in SCO, which saw LEV increase throughout the day and saw some solid buying action just after 11:00 (200K shares or so).
  • VOT is another mid-cap growth ETF, and it clocked in a positive-trending LEV while the retail Small Effective Volume (SmEV) sold off in the morning and never returned.
  • FVD is a dividend ETF and it's been under solid accumulation, as evidenced by the LEV curve, all week.  Friday was no different, and the LEV kept increasing while SmEV held steady.  Note that FVD is thin in terms of $-Volume @ $1.5Msh; I like to consider ETFs above $10Msh.  Note that someone picked up 100K shares on 1/30 in this, so there is some institutional movement.
GGT New Long Stocks for Monday, February 7th

There are 70+ New Long stocks for Monday, so plowing through each of them isn't as highly desired as you may think.

The first thing I do is throw out any stock that has a $-Volume level of less than $10Msh.  This reduced the list to about 50 or so.

Next, I use a combination of filters to get rid of any stock that isn't meeting general Elder criteria:  FI(13) > 0, 13d EMA > 34d EMA.  This reduced the list to about 30.  Time to plow ...
  • UAL, United Airlines, saw some impressive buying on Friday. 385K shares were traded at a bit after 2 pm, then another 371K shares about 30 minutes later.  Both significantly contributed to LEV.  About 3:40 someone sold 225K shares, but from that point to the end of the day the buying continued, forcing LEV to move positive.
  • WTFC, Wintrust Financial, saw a number of transactions occur throughout the day that added to LEV until the last 10 minutes of the day. 
  • ALB, Albemarle Corp, is in LEV accumulation
  • BDX, Becton Dickinson, reports earnings in 2 days and is experiencing significant LEV accum
  • PHH had two blocks of 183K shares trade within 20 minutes of each other and the price didn't budge until the second block.  Solid SmEV and LEV accumulation.
  • BC, Brunswick Corp, saw huge accumulation THURSDAY at the close of business, and LEV didn't look back all day Friday
  • DBD, Dibold, saw a flurry of buying/selling with a net LEV increase in the last 10 minutes on Friday.  Interesting.  They report earnings in 9 days.
  • CNK, Cinemark Holdings, saw solid, steady LEV accum all day Friday, with two big lot grabs throughout the day.
  • BBY, Best Buy, has been under solid accumulation for a long time.  Friday was no different.  Some price pressure mid-day saw the retail boys running but the LEV kept bouncing upward on the dips.
  • AVP, Avon, which I've been ignoring overall, saw about 200K shares in 2 minutes just before closing, causing the price to jump up and LEV to skyrocket.  AVP reports on 2/8.
  • ARRS, Arris Group, beautiful LEV pattern on Friday.
  • NKE, Nike, solid accumulation for several days.  They report on March 17th
  • CLX, Clorox, reported before the market opened on Friday.  Amazing LEV all day Friday.
  • SAPE, Sapient Corp, saw large lot buying throughout the day, all contributing to a positive-moving LEV.  In fact, LEV never decreased except for a brief sell just before closing, then it skyrocketed to the day's high.
  • TRP, Transcanada Corp, has been under solid accumulation for days and Friday continued the pattern
  • CRI, Carter's Inc, the little kid's outfit store, has been seeing big lot purchases for several days.  It reports on 2/24.
  • JNPR, Juniper Networks, massively jumped in the last 5 minutes of trading in terms of LEV.  
  • MRX, Medis Pharma, has something going on with it.  A HUGE move in volume at the end of the day on Thursday pushed LEV and price up, and the same thing happened in the last 15 minutes on Friday. 
  • CIEN, CIena Corp, saw amazing buying on Friday, driving LEV up in major steps.  This one saw serious buying, with over 400K shares traded at 11:18 @ $25, or over $10M in one pop.  Volume clocked in at 127% over the 50d average ...

All of these are good candidates with significant institutional buying.  The ones at the top of the list are in my buy zone, the ones at the bottom of my by list are outside of my buy zone (CIEN is 10.3% above!!!).  I intend to wait for these to pull back, and where I see good value, I'll enter.

Obviously though, probably not Monday ....


Remember, you are responsible for your own trading decisions, and I am not.  Please do your diligence, and please take ownership for your actions.

Make it a great weekend.



Position disclaimer:  I own or influence positions in the following equities:  AAPL, BGU, BTI, CREE, DAG, DIG, ERX, FAS, HPQ, HSFT, HSY, IEO, IEZ, IM IYE, JJC, LULU, MVV, PALL, PSQ, RIMM, RJA, RSX, SCOK, SDS, SOXL, SSO, TGT, TWM, UPRO, UYG, UYM, VDE, VIT, WPI, XHB, XLE, XME, XOP.