Saturday, February 25, 2012

GGT Timer Systems

The blog is not on hiatus; it can be found (almost) daily at, in the "GGT" forum.  You'll need to register, but access to all of the blogs is free.


GGT Timer Rules

I've been working with my colleague Bob Wilson in quantifying various aspects of GGT.  As many of you are aware, GGT employs three timer systems:

  • Short-term:  uses a simple 4-day moving average to cross the Long-Cash Ratio (LCR) value, which is calculated daily.  Buy when the 4d crosses the LCR from below, and sell when the 4d crosses the LCR from above.  You cannot calculate the LCR but I can, and either Bob or I do it every day and post the results at my GGT forum @ Effective Volume.
  • Intermediate-term:  based on Dr. Alexander Elder's Force Index (FI) methods, this timer uses the slopes of the 13d and 34d EMA of the GGT prices, in addition to the sign (+ or -) on the 13d Force Index.  Buy when the FI(13) is turning positive AND the slopes of the 13 and 34d GGT prices are (turning) positive; 
  • Long-term:  a basic, keep-you-out-of-trouble timer based upon the 5d and 65d crossing of moving averages.  Buy when the 5d crosses above the 65d from below, and sell when the 5d crosses the 65d from above.  It's that simple.  Not optimal, but quite simple.
Like all timers, each of these has their benefits and detractors.  Benefits are getting you into the market, as well as out, as the markets dictate.  Detractors are simply that they lag, and they may whipsaw from time to time, and more importantly, they do not always work.

Here's the current status of the timer system:

As with all my images, right-click on the figure to open in a new tab or window.

Shown above is the last 2.5 months of signals.  The short-term timer is on the left, the intermediate-termed timer is in the middle, and the longer-termed timer is on the right.

You can clearly see that the short-term timer moved to cash on 2/9, forcing a sell at the open or early on 2/10.  The intermediate and longer -termed timers have remained long, so it's important to choose what your period of comfort is.  Typically, the longer your timer, the less worried about intra-signal drawdown you are.

Looking back over time, you can see the impact of the different time frames on the different timers.  You can also see the impact of the shorter time frames and the concepts of whipsaw -- look closely at 12/8.

There are benefits to having multiple timers.  The end result is that you often have equities in play, satisfying some deep urge to play the markets at times when it probably is more prudent to be on the sidelines.  As long as you entered at the correct time, leaving positions on, based on the time frame that you are focused upon, is most likely the smart approach.

Market Entry

For all my systems, the key entry is when the 5d slope of the LCR turns positive, and the 2d and 3d are either 1) already positive, or 2) are turning positive on the day that the 5d turns positive.

Where does this (seemingly) obscure entry rule come from?  Here's the results of some testing that I did concerning the best time to enter the markets:

It will take you a bit of time to understand how to apply this table.

Let's say that the markets have been in a tail spin, heading south.  Everything (LCR slopes) is negative.  Let's further say that the markets stabilize for a couple of days, turning the slope of the 2d EMA of the LCR  positive.  Since 2009 if you would have entered the market the next day near the open and sold at the close of that day (unrealistic), you would have turned $1 into $1.56.

If you wait until both the 2d and 3d slopes turn positive you would be at $1.67.

Finally, if you wait until the 2d, 3d, and 5d slopes turn positive you would have turned that $1 entry into $1.86.

So, where are we today?

Here's the LCR slope table:

There's other things on this table -- simply look at the columns that say "Slopes of LCR EMAs".

As you can see above, all of the slopes through the 55d slope are negative (red), and the 65d is super close to moving negative.

In the context of the table, which I'll annotate below, we have a situation where ONLY the 65d is positive:

IF, and this is a big if, the 2d moves long on Monday 2/27, then historically, with ONLY the 65d being long, we've only gained 12% on that entry (early Tuesday, 2/28) over the past 3+ years.  If the 2d and 3d both turn, but the 65d is the only other long, then we increase that value up to 17% edge.  You see where I'm going with this.  Playing late stage signals, while certainly profitable, is not nearly as profitable as being early in the game.

Given the present scenario, I have no expectation of having the same edge as the beginning of a cycle.  This goes to reason -- at the beginning of a cycle the markets are typically oversold, and at the end, when ONLY the 65d is still green (positive), the markets are overbought.  

So, how does TIMER entry work?

IN ALL CASES, we need to have the slopes of the 2d, 3d, and 5d EMAs all positive at the end of the day, whether simultaneously or sequentially (it doesn't matter):

  • If concentrating on  the short-term timer, then buy at or near the open of the market on the following day.
  • If concentrating on the intermediate-termed timer, then review the status of the intermediate-term timer.  If LONG or MIXED, then buy at or near the open of the market the following day.  If in CASH, you are "enabled", but you must WAIT until the intermediate-termed timer moves to a LONG or MIXED state before entry.
  • If concentrating on the long-term timer, then review the status of the long-term timer.  If LONG, then buy at or near the open of the market the following day.  If in CASH, you are "enabled", but you must WAIT until the long-termed timer moves to a LONG  state before entry.

Refer to the timer summary above, specifically on 12/21.

On 12/21 the short-term timer fired long.  You would have placed orders for positions early the morning of 12/22 under a "short-term" strategy.  You also could have placed orders for positions early the morning of 12/22 under the "intermediate-term", because of the "Mixed" status of the Elder timer.

On 12/27 the long-term timer moved long.  You would buy your anchor positions early the morning of 12/28 under this timer.

On 12/30 the intermediate-termed timer moved to CASH.  You would have sold early the morning of 1/3.

On 1/3 the intermediate-termed timer whipsawed back to LONG.  You would have BOUGHT early the morning of 1/4.

On 1/30 the short-term timer moved to CASH.  You would sell your short-term holdings early the morning of 1/31.  If investing in an index such as GGT (looks like the ETF "VTI"), you would have a gain of nearly +6% for the signal.  You would continue to hold your intermediate and long-term positions.

On 2/1 the short-term timer moved LONG -- another whipsaw.  You would enter early the morning of 2/2.

On 2/9 the short-term timer moved to CASH.  You would sell your short-term positions the morning of 2/10 for a gain of around 2.1%.  You would continue to hold your intermediate and long-term positions.

As of this writing the short-term timer is in CASH, effectively blocking entry to the markets.  I'm waiting for the 2d, 3d, and 5d EMA slopes to move positive.

Timer Performance

With the rules established, what is the performance of the system since 1/1/2009?  Here's an equity chart, applied to the GGT index, which again, looks like the ETF "VTI" from Vanguard:

It should be evident what you are looking at here.  The yellow highlighted area shows last year's period where trend following simply sucked -- there is no other way to describe it.  Using trend following methods the markets did not work last year, and in general, the 5d/65d timer suffered the most.

Overall though, you can see that these timer methods all produce very good gains and keep you in the markets when you need to be in the markets, which is important.  Here are the statistics for each timer:

As you can see, and as you saw from the equity graph, the timers all manage the down side adequately.  The 5d / 65d method has more losers than winners (6 compared to 3), but the winners are stellar compared to the amount lost, on average (30.23% vs. -4.16%).

From a cumulative performance perspective, the Elder FI timer has the best overall stats to date.  Next in line is the 4d LCR, followed by the 5d / 65d timer.


I've presented a methodology to use the GGT system timers to positively impact your portfolio.  Purchases were made into the GGT index as the example benchmark -- investing in stocks with a beta higher than the Vanguard Total Index (VTI) would obviously change your individual performance.  Over the past 3 years waiting until the 2d, 3d, and 5d EMA slopes of the LCR moved positive had a significant impact on entry edge.  Depending upon your risk tolerance, and depending upon your time frame, three timers are available (for free!) that work well overall in trending markets.