Friday, April 27, 2012

Update for Friday, April 27th -- Elder Force Index Timer Confirms LONG

With the close of markets on Thursday, April 26, my Elder Force Index Timer has confirmed a move to the Long side:

The hold out here was that I use two methods -- a slower simple moving average (SMA or MA) and a faster exponential moving average (EMA) to calculate the Force Index (price change * volume). Both are positive on a 13d average, so we have confirmation. 

Shorting stocks (or Contra ETFs) at this point in the market is a low-probability proposition at best. This is an intermediate-termed timer and with the long-term timer (13d / 65d crossing), the direction is up from here.

As you can see, all of my primary timer systems are long.


The combined timer historically has kept us on the right side of the market:

If you have a long(er) term view, this could be a viable methodology to know whether you should be in the markets or out. I'm targeting "in" right now with a core position in IWM.


If you read my entry yesterday on my GGT blog, you know that I was going long in IWM.

I didn't get into IWM for a couple of reasons, all mechanical. To wit:

1) I set a limit order at the previous day's close (80.94)
2) I set a timed delay (active @ 9:35 a.m.) to allow the overnight book to clear. The markets opened below the close, moved down, and by 9:35, were above my limit. They never looked back.
3) I was away from my PC so couldn't monitor that this was occurring, or I would have attempted to get in by 10:30.
4) Lunchtime is typically weak, but not yesterday. Hence, my VWAP entry never executed.

I'm still largely in cash, with a GTC limit order at 80.94 still pending. Futures are up as I write so I need to get on the bus or risk being locked out of this entry. Shoot. I hate chasing.


Shameless plug:

I will be away from my PC most of today and all weekend. I am riding over 100+ miles this weekend with disabled veterans from the Pentagon to Gettysburg, PA. Read about it here: 

Please read my personal fund-raising page here:, and if you are so inclined, please contribute to this cause. We, in America (and all around the globe -- I am a verteran and was stationed overseas), benefit from these sacrifices and this is a good cause.



Wednesday, April 25, 2012

Update for Thursday, April 26th -- Combo Timer Signals New Long

Effective with the close of markets on Wednesday, April 25th, the GGT combo timer, which is comprised of both the Elder Force Index timer and the GGT 13d / 65d cross timer, has signaled a New Long.  Entering positions early in the morning the day following a transition of this timer to a long state has been profitable more often than not since November 2008.

Let's look at the timer performance coupled with the GGT index (which generally looks like the ETFs IWM or VTI):

(Right-click on the image to open in a new tab or window)

As many of you know, I have developed a number of timers -- their performance since 1/1/09 is shown above.  The best timer -- the 13d/65d timer, has considerable drawdown and most people would abandon it once they gave back half their gains.

Surprisingly, when we combine the best timer and the "worse" timer (someone has to be last), we get a complementary pair that works well together.  The "worse" timer is the Elder timer.  Combining with the best timer gives us the purple curve, and as you can see, it behaves well overall.  Not perfect, and I want you to look closely at the areas I've circled -- these are areas where all the gains have been given back and then some, which is certainly a possibility at this stage in the market.

You must be aware that no timer is perfect, all timers whipsaw, and all timers are subject to their worse drawdown IN THE FUTURE.  Internalize this statement, and make sure you understand it.

Now, with that caveat out of the way, let's look at the performance statistics.  There is more than equity to quantify the behavior of a timer -- and you should at least be somewhat familiar with how I do this.

The timer has been in CASH since 4/4/12.  At that point, since 11/25/08, the timer was responsible for turning an initial $1 into $2.03, a gain of 103.7%, and a compounded rate of return of 23.15%.  The timer is only in the market 49.5% of the time, so consequently, it has a very good efficiency -- a bit under 2.1.  Values above 2 are considered quite good.

The worse trade in this entire time, not counting commissions and slippage, is only -2.41%.  This is accomplished by executing early the following morning as close to the previous close price as possible (or better).

The best trade is +19.59%, and it goes without saying, when the system works, it works.

Dropping down, the system has only had 19 trades since 11/25/08, so this is a low-cycle timer.  I'd like this number to be higher, but my time machine has failed and I can't advance to the future to check the performance, so we must do this the old fashion way -- day by day and trade by trade.  The system values will start converging around 24 trades, and more is better.  All in time.

Dropping down a bit further, Van Tharp's t-test/SQN metric is a solid 2.65.  Not a great system in his book but not bad either.  We want values above 1.7 to assert that our gains are not due to chance.  More trades improves the quality of this value.  I'm looking to see it go up with increasing trades -- we'll see.

Dropping down to the yellow line, mathematical expectation (ME) is 2.62.  This is an amazing number, relative to systems like Connors' and others.  This shows that there is a high degree of probability that the system will produce gains going forward -- we have a significant edge on the long side.

Finally, the PRR value is the Pessimistic Return Ratio, and it improves as the number of trades increases.  A value above 1.0 shows that we'll most likely make money, and values above 2.5 are considered very good.  We're there.

Again, I cannot stress enough -- 19 trades is a low number of trades for metrics on any system.  It is quite possible that the 20th trade will be a significant loser -- AAPL *IS* the market right now, and if AAPL goes down, the market will likely follow.  Unrest in Europe with the ECB, Spain, Italy yields, "Sell in May and Go Away", I could go on and on.  Markets climb a wall of worry and we're there.


A number of conditions triggered this timer long today.  The Elder Intermediate timer turned to "mixed" mode, which is essentially "long" but not confirmed by a slower-moving average method:

Essentially, I use both a simple moving average (SMA) and a exponential moving average (EMA) method to calculate Elder's Force Index.  The SMA is still below 0 (by very little), and the EMA is above 0.  Hence we get a "mixed" signal there.

Of some importance is that the 13d Force Index slope is pointing upwards -- it is gaining day over day.

Next, the 13d and 34d EMAs on the GGT index are positive slope - they are pointing upward.   Prices are moving upward.

Hence, by a pure EMA method, we have Elder LONG going into tomorrow.  If we incorporate a slower moving average to require confirmation, we're not there yet, hence we're Mixed.

I tend to be more aggressive so I'm long at this point.


The 13d / 65d timer was already long going into today:

As a result, we have the necessary and sufficient conditions for the combo timer, and we have a "Enter Position" signal.

It's difficult at this point to know if this signal will work; your crystal ball is as good as mine.

I had lunch a few days ago with a fellow HGSI user and I made a statement "start with ETFs ... if you can't make an ETF work, then stock picking will be more difficult over the long haul.  I believe that here too.

At a bare minimum, I plan to enter IWM tomorrow at the open (or even before).  I base this on answering the question "What does GGT look like in the most recent time frame?"   The answer lies in the Pearson correlation with the GGT index and various other indexes:

The closer the value is to 1.000 the more the GGT index is acting like the various ETFs shown over the time periods shown.  You could argue that the VTI and IWM are close -- they are -- and I simply am picking IWM because I picked VTI last time  :)

How big will my position size be?

Let's say you have a $100,000 account.

Let's further suppose that you only want to risk 0.5% on this "newfangled" idea from Paul.

Here's how I calculate the position size.

Account Equity:  100,000
Risk:  0.5%
Amount to Risk:  100,000 * 0.5% = 500.00
Worse Trade in System:  -2.41%
Position Size = 500.00 / 2.41% = 20,746.88
Closing Price IWM:  80.94
# shares = (round down) 20,746.88 / 80.94 = 256 shares
Stop Loss:  80.94 * (1-0.0241) = 78.99

With this stop set, your maximum exposure on a 20K position is $500.

REMEMBER:  the worse trade has NOT occurred, and the worse drawdown has NOT occurred.  This could be a bumpy ride.

For those who regularly read my entries at the Effective Volume site (, here are the obligatory price and LCR model tables and figures, without commentary.

I intend to set my stops as indicated, and let the markets do what they do.

Enjoy your Thursday.



Tuesday, April 10, 2012

Update for Tuesday, April 10th -- Cash is King

I blog (almost) daily at; please register there to view my updates.


Aside from positions in UVXY and AMLP, both of which are longer-termed holdings, I'm in cash.


The Price Slope Model is losing the bullish battleground, with the 34d slope now transitioning negative. I've used this in the past as a "canary in the coal mine" and if we hold the negative value today the overall direction of the markets is most likely continued down.

Note that the GGT price index fell -1.28% on volume that was -18% below the 50d MA for volume. This was not a "nail the coffin shut" sell off. I do note too that the database strength index, which combines price, volume, and rate of change of price information, did drop to under 0.3, which historically has been a ripe area for reversals to the up side. Time will tell.

Certainly, the last 4 days have seen increasing bearishness, so we need some green on the right side of the table or we'll continue to transition negative on the left side of the table. If all the slopes turn red, this bull leg is completely dead.


The LCR slope model has been bearish for some time and continues so:

With more stocks in the database with a "cash" recommendation than with a "long", there is plenty of fuel to go higher if Mr. Market decides to reverse.

Monday saw a -30% drop in the LCR value, which is the 52nd largest drop out of about 450 that I've registered. Notable.

There's nothing bullish about this table. It has been telling us that cash is king for some time.


The short-term timer, which is based on a simple 4d MA and the Long-Cash Ratio (LCR) value, has been in cash for several days. Graphically, this is what it looks like:

You can see that we are heavily extended below the 4d MA and we'll need a huge rally to move the stocks from the bearish side to the bullish side. I'm not overly optimistic of any transitions today.


The Elder Intermediate-termed timer, which uses the 13d Force Index, the slope of the 13d Force Index, a 13d Pricing Slope (GGT), the 34d Pricing Slope (GGT), and a big AND statement, is completely in CASH, and has been signalling cash since 4/4.


The Longer-termed 13d/65d crossing timer is still long, and it is telling us that the 13d is still 6% or so above the 65d:

Of interest here that your eye can see the lower highs and lower lows (LHLL), and if you look back at the inception of this timer, the results are pretty clear.


Correspondingly, the three timer systems that I employ are short-term cash, intermediate-term cash, and long-term still invested, BUT, I combine the long-term and the Elder intermediate-term timers to give a better result. Hence, I'm in cash.

The performance of the various timers is quite good, and they are my roadmap forward:

You can see that the long-term 13/65d is still long and is the best performer, but it also has you fully exposed with higher drawdowns than any other timer.

Right-click on the following image to open in a new browser or window:

The combination timer is at the far right and is the one that is the best performer. You can see the various statistics on the other timers. Although the 13/65d [fixed typo] cross timer is the best performer, it only has 3 total trades, so the statistics are suspect. Nevertheless, we put performance in the bank, and this one does not lie. Note that the exposure of 63% is high relative to the others.


Overall, I'm on the sidelines. Nothing is telling me to enter the markets, and nothing is telling me that I should be preparing to enter either. I'm focusing on other things today.



Thursday, April 5, 2012

Update for Thursday, April 5th - Elder Intermediate Timer Signals Cash

I blog daily at; please register and follow me there if you want daily commentary.  I generally make time to blog here only if I get a transition in one of my measurement/timer systems.


With the close of markets on Wednesday, April 4th, the Elder Intermediate-termed timer has transitioned to cash. This is a trigger event to sell long-term holdings, and I will do so at the open.

Shorter-termed holdings should be looked at very carefully for exiting. Longer-termed holdings, such at those purchased around January 5th and which have a solid gain, are probably safe but my models give no further selling guidance beyond what I'm posting today and history has shown that waiting for the 13d/65d model to transition to cash on it's own is generally a "give back solid gains" decision. I'm moving to cash and waiting for the next entry signal.

The latest entry the morning of 3/13 resulted in a net +0.63% gain overall to the combined 13d/65d model:

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

Note that ME has dropped a bit due to the smaller gain but PRR has increased, giving me further confidence at this approach.


The price slope model is mildly bearish/bullish, with a few of the shorter termed slopes below 0 and the remainder above 0. This is mixed and the bulls/bears are fighting the battles on above-average volume:


The LCR slope model is completely bearish and is telling us not to enter stocks on the long side.


Of interest to me is that there is some life left in the long-term timer before it crosses into bear land:

Note the distance between the end of yesterday's signal and the 0 line. Barring a major collapse of the markets, the long-term timer should remain long, indicating that long-term positions can continued to be held.


When I calculated net Effective Volume divergence this morning I was surprised at the apparent amount of net inflow into equities. There are far more stocks that are attracting money on decreasing prices than the converse, and this leads to the conclusion that this dip is being bought:

Time will tell whether this is the correct interpretation.


I'll be away from my desk most of the trading day today.



Wednesday, April 4, 2012

Update for Wednesday, April 4th - Another Short-Term Whipsaw

I do not post here daily; I post daily at  I post here as time allows.


With the close of markets on Tuesday, April 3rd, the short-term LCR timer has transitioned to CASH. 

Under the rules of the re-entry timer, the index position entered into yesterday will be closed at the open. The Elder intermediate-termed timer and the 13d / 65d long-term timer are still long, so positions that were entered into early in January or the middle of March, which have gains, should continue to be held.

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

While any system that whipsaws is frustrating, you can see that the secondary re-entry system manages the trades quite well (as long as you execute flawlessly), with the worse trade ever recorded at -5.46%, the average loss at -1.59%, and with a significant positive edge (ME = 0.84, want > 0, and PRR = 1.5, want greater than 1.0 at a minimum).

I'm not as anxious to sell, and backtesting that I have conducted shows that a 1% trailing stop loss (TSL) placed on the trade 5 minutes after the market opens typically works in your favor (e.g., you get a higher price on exit than the opening). I intend to place the TSL and let it ride from there.


The Price Slope Model is wavering on turning bearish, but simply has 1 day of bearishness developing on a good bullish base:

Although the 2d and 3d slopes turning negative is undesired, the weakness we see is characteristic of being within a range.


The LCR Slope Model has resumed the bearish stance it has had for a while and tells us that more stocks are moving below their historical, optimized performance levels and that we should not be buying stocks today:


I'll be away from my desk all day today until near the close.



Sunday, April 1, 2012

Update for Monday, April 2nd - Update on Timer Systems


Overall, the long term trend is up but the intermediate trend is under pressure to move to cash, and the short term trend is in cash. The Long-Cash Ratio continues to contract, meaning that day-over-day, more stocks are moving to a recommendation of CASH than in the other direction, so moving long at this time would be swimming up stream. My personal view is that positions entered into in early January and around the middle of March should continued to be held if they have profits, but I will prune any underperformers and move to cash for those that are underwater. To accomplish this I have set a 1% trailing stop loss on the underwater positions and if they move up, great -- if they drop, I'm out.

Here are the details:

The price slope model (PSM) has been increasing in bearishness over the past few days:

As with all my images on this site, right-click on the picture to open in a new tab or window at the original size.

Price moved up a slight amount (+0.13%) on volume that was -3% below the 50d MA on Friday, which is unremarkable.

The slope of the 2d price EMA moved positive on Friday, but overall, you can see that behavior for the week was net down. While there is a positive expectation for re-entry on Monday on the long side under these price conditions (2d positive, 3d - 8d negative, 13d - 65d positive), the chances of these trades working out in a historical sense are only between 36% and 44%. The following chart shows the historical results compared to entry when the entire price slope matrix is negative, and the 2d transitions positive (the highest possible chance of the trade working):

It's clear that relative to coming out of a complete oversold condition that our present conditions are not any where near ideal. From the table you can see that you should be buying when:
  • All of the price slopes are negative and the 2d transitions positive (reference point)
  • The 2d is already positive and the 3d transitions positive (88% success relative to 2d transition)
  • The 2d is already positive and the 5d transitions positive (55% success relative to 2d transition)

Furthermore, the table suggests that re-entry on pullbacks can be incredibly powerful if conducted early in the recovery. Hence,
  • If 3d already has a positive slope, and 2d transitions from negative to positive (due to a short-term pullback), and the rest of the EMA slopes are negative, then there is a 91% chance of success.
  • If the 5d already has a positive slope, and 2d transitions from negative to positive, and the rest of the EMA slopes are still negative, then there is a 61% chance of success of the entry.

The table confirms what most of us already know -- participation early in a new move is key if we want the highest probability of gains. As the move matures the chances for gain decrease.


The Long-Cash Ratio slope model continues to contract, with the LCR dropping another 1% on Friday despite the weak move up in price:

All of the LCR slopes are negative, although we did see some strength in the LCR acceleration on Friday. Bottom line, there is no compelling reason to move long when the database is showing us that stocks are falling below their historical optimized performance thresholds.


I've refined the longer-term timer and have moved the 5d / 65d EMA crossings to a 13d / 65d set of crossings. Additionally, when I do this and combine secondary entries to the Elder Intermediate Timer Exits, the historical performance is outstanding and represents the best model that I've been able to develop to date.

Here's the performance table:

Across the top are the different timer conditions that I've chosen to display. Along the left are the descriptions on the various metrics I use. Let me quickly explain these.

Cumulative Performance is since 11/25/08, which is when the GGT database was expanded to over 3000 stocks. The data in the database is "stable", and the methods used to calculate the GGT optimized levels have not changed since this time (interpretation has, but methods of calculation have not).
Compounded Rate of Return is exactly what you think it is.
Market Exposure may be new to some of you. Timer systems have the ability to sit in cash, so the lower the value, the less you are exposed to the market. In general, we want this number to be lower, not higher. If we were in the market 100% of the time (buy and hold), we would see a 100% here.
Efficiency is a related to market exposure. It is the amount of Cumulative Performance divided by the Market Exposure. Large Efficiency values indicate higher gains with lower market exposure, which means lower overall risk.
Worse Trade, Best Trade, Average Trade, Median Trade, Stdev, etc. are all self explanatory.
Mathematical Expectation (ME) is the "edge" of the system. We want this to be positive. The more positive, the better. Note that ME has a "flaw" of sorts -- it does not look at consistency of trades.
Pessimistic Return Ratio (PRR) counters the "flaw" in ME and incorporates the consistency of trades. The more trades, the better PRR will indicate your "edge". We want PRR to be above 1.5, and values above 2.0 are considered very good.

I've circled two columns. The right column is the system which my focus will be upon going forward, and the left circled column will also be the focus of my efforts in the event I want to "jump into" the market mid stream (in the event I miss the initial entry for whatever reason).

The timer rules for initial entry are explicitly this:

Enter the markets on a new signal:
  1. if the 13d price EMA crosses the 65d price EMA from below AND
  2. the Elder Intermediate termed timer is NOT in CASH.

Re-enter the markets if stopped out or on a failsafe condition:
  1. if the LCR has been above it's 4d SMA for two days (this is my short-term timer system) AND
  2. the Elder Intermediate termed timer is NOT in CASH AND
  3. the long-term timer is already long.

Re-enter the markets if stopped out or on a failsafe condition:
  1. if the 5d LCR transitions from negative to positive slope AND
  2. the Elder Intermediate termed timer is NOT in CASH AND
  3. the long-term timer is already long.

Exit rules are simple: Exit all positions if the Elder intermediate-termed timer transitions to CASH.

The entry/exit dates for the system are as follows:

Additionally, there are times when there is an edge to adding positions to the system. I'm keeping track of those dates, and the results were in the timer table above in the "Secondary Entry + Entry Performance" column. This too has a significant positive expectation and is a very efficient system for managing secondary entries. Here are the dates and status of the system for secondary entries:


Overall, my plan for Monday is unchanged. I see no compelling reason to enter positions at this time, and as far as selling is concerned, positions entered recently that are weak will be sold with a 1% trailing stop loss in place.  I do have hedging positions on in all accounts and plan to add to those positions if a reversal in the long markets is confirmed.