01/9/2006 FastBreak - Year 2005 Review

Before reviewing FastBreak trading system performance we want to make a few announcements:

We are discontinuing the Discussion page on our website.  Automated “Spamming” programs continue trashing the page so we are moving to www.ft-talk.com, a subscriber site, to stay in touch with our users.

For information on all our products see FastTrack Commentaries 8160, 8161, 8162, 8163, 8164 and visit www.edge-ware.com

2005 Year in Review

FastBreak trading strategies we trade or follow had returns ranging from a 67% gain to a 10% loss for all of 2005.  This commentary will review several different types of trading systems.

The first version of FastBreak was released in early 1996. Potential users of any investment software product should always ask the question "What is actual performance?" Soon after the initial release we developed trading systems that could be monitored for "real time" performance purposes. We don't modify trading system parameters during the performance reporting period and report on the systems on a yearly basis.  The 2004 review is available in FastTrack Commentary 8164. If you would like to review all reports since 1997 visit our web site www.edge-ware.com and go to the Strategies page.

With the release of Standard FastBreak and FastBreak Pro version 4.0 in late 2001, we have the ability to run the same strategies in the Standard and Pro versions. All of the systems we report on here were developed with the Pro version but can be run in the Standard version.

We place sample FastBreak Pro developed trading systems on the Strategy page of our web site. These systems provide users and potential users a starting point on the many different types of trading systems that can be built with FastBreak. It also allows us to provide an honest benchmark for real time performance. Users of both Standard and Pro versions can download these systems. The Standard FastBreak 30 day demo that can be downloaded from the web site can also run these strategies but will not display the last 60 market days of trades.  Many of these systems are a number of years old.  We can now build better trading systems, but we keep these systems on our web site to show that many trading systems are effective for long periods of time.

We don’t often post or change the strategies on our web site because we believe that investors’ needs are very unique and users of our software should feel confident developing their own trading systems. 

This year we are excited about reporting on stock trading systems that have at least a full year of performance.  More and more mutual funds are imposing unreasonable short term trading fees, and trading stocks may be a reasonable option for both individual and professional investors.

The strategies we report on are "out-of-sample" for the entire year and in some cases the strategies have been unchanged for several years.

The 2005 US domestic stock market was not necessarily a good year for investors, with the exception of a few narrow sectors e.g., energy and gold.  Many international markets did have a very good year.  The US markets spent a lot of time wandering in no particular direction. As pointed out in the December issue of Almanac Investor Newsletter, the Dow Industrials index had the narrowest trading range in 109 years.  This is to say, the Dow traded within a range of less than 9% of where it ended 2004 and finished slightly negative for the full year.  The S&P 500 and the NASDAQ had similar action.  This kind of trendless, sideways market may be the most difficult to trade.

The bond market, as measured by STBI- (Treasuries) and DJ-2B (Commercial), couldn’t make up its mind which way it wanted to go.  The one surprise was the strong rally in the US dollar as measured by DXY-Z. 

FastBreak trading systems had very mixed results.  All systems preserved capital and several had exceptional results.  None of the systems had a “blowup” which is very important for mechanical trading systems.  The first rule of making money is to not loose money.   

We need a yardstick with which to compare our performance. Here is how common benchmarks performed in 2005:

 

Return %

MDD

UPI

S&P 500

3.0

-7.2

0.3

DJ-30

-0.6

-8.5

-.7

OTC

1.4

-12.5

-.1

STBI- (Treas. Bond Index)

6.7

-5.8

1.5

DJ-B2 (DJ Bond Index)

1.3

-6.3

-.4

DXY-Z (Dollar Index)

12.8

-4.6

6.1

Rapid reversals in many markets during the year made it difficult for many professional money managers. We also include two other measures of investment returns:

 

Return%

Managed Futures Funds

~ 3 to 4%

Hedge funds

~ 4 to 7%


The above values are approximate because of the difficulty obtaining reliable information on these programs.  Hedge funds have had considerable investor interest during the past few years as an alternative investment, and there has been a tremendous flow of money into these funds. Managed futures and hedge funds typically charge 20-25% of gains for their investment services. These funds are usually only open to qualified investors (those with a net worth of $1M or more) and minimum investments of $500,000 or more is not uncommon.  This is the second year in a row that was difficult for most hedge and futures funds, with many closing their doors. 

How did the FastBreak strategies posted on our web site perform?

 

System Return%

MDD

 UPI

Bonds posted 2/18/02 

8.9

-5.2

3.6

Growth w/RUTTR signal 

7.1

-5.0

1.9

International 

20.7

-11.1

3.2

Bonds LT Gov and HY 

1.4

-6.2

-0.2

Bonds Conservative zero

2.1

-2.6

0.1

Bonds Aggressive zero

0.5

-7.1

-.3

Juno only (shorts bonds) 

-1.5

-4.6

-2.4

Arktos only (shorts OTC) 

-1.5

-6.7

-0.8

Ursa only (shorts S&P)

0.5

-2.3

-1.3

Low beta Growth(w/RUTVOL signal)

3.9

-3.3

0.9

High beta Growth(w/RUTVOL signal)

12.1

-5.5

0.0

Income Strategy

-6.6

-9.7

-1.4

Weak Dollar

-6.1

-13.7

-0.8

The bond systems were satisfactory. After several great years, the bond market was looking very tired in 2005. The bond market was very choppy (same story as the last two years – war, fear of deflation, economic recovery, etc.) The FED raised rates a number of times during the year and that took the wind out of the bond market sails.  The original bond system that was posted back in 2002 continues to perform well.  The more aggressive bond systems that trade zero coupon bond funds underperformed.  These systems are built to take maximum advantage of a strong treasury bond market.  There were a couple sell offs in the bond market, but the downside didn’t continue to the extent that our short bond system (Rydex Juno) was able to turn a profit.

The international system continued to have stellar performance. This same system was up 29% in 2004, 63% in 2003 and 30% in 2002. This has been a fantastic system; however, since the strategy was built in 2002 many of the funds have closed or incorporated short term redemption fees. The results do not reflect those fees. We have maintained reporting on this strategy to demonstrate long term strategy performance. We have developed a new international strategy that uses closed-end country funds, and we will cover later in this report.

The Arktos and Ursa systems short the market (OTC and S&P 500). Both systems stayed in money market for the vast majority of the year. The systems went short during the two sell offs, but quickly went back to money market when the market recovered. Having a flat return may not look like good performance, but several investors who tried to short in 2005 had very significant losses. A trading system needs to be robust in all market conditions. We keep these systems in reserve when we have another 2001-2002 bear market.  These two bear market trading systems are somewhat dated.  We built newer, more robust bear systems with FastBreak Pro Version 5 and they had returns of 1 to 2% for the year and maximum drawdowns of less than 6%.

The growth system that uses the RUTTR market signal did well, but those systems using RUTVOL timing signal were somewhat disappointing. These market timing signals have been very popular with the FastTrack investor community. We admire the excellent work that went into building these signals, and we built trading systems to utilize the signals. Both RUTTR and RUTVOL had good trades, but there just wasn’t enough market movement to turn a significant profit. The signals did help avoid some of the worst sell offs and they were in cash for a substantial time.  The RUTVOL strategies captured “market” returns with much lower MDD.

The Income strategy struggled this year after three years of fantastic performance.   This strategy trades a variety of Fidelity bond and real estate funds.  With most bonds going nowhere and real estate funds being whipsawed, this strategy just couldn’t get traction.  Note: When we built this system in 2003 we built a similar program that used a different momentum ranking method.  We posted what we thought was the slightly better system on the website, but in retrospect, we should have posted the other system because it has significantly outperformed in the past two years (10.2% return, 9% MDD in 2005).  We will replace the existing system on our website with the one that has shown consistent performance over the last two years.  However, with both bonds and real estate funds looking tired the best days of this strategy may be in the past.

Our biggest disappointment was the Weak Dollar trading system.  We shouldn’t feel too bad about this system because the dollar index rallied 12%, and this system built for a falling dollar environment, lost approximately half that amount. We included two closed end world bond funds in the trading family.  These funds trade like stocks and it is an easy way to take a short position against the dollar.  The system entered the year holding GIM which has been a star performer.  However, in mid March when the dollar changed direction, this fund/stock plummeted over 13% in less than 8 days.  This is unparalleled in the long history of this fund.  The volume was massive and we can only guess that an institution(s) holding large positions dumped the fund.  Although the strategy performed well for the remainder of the year, it just couldn’t dig out of this hole.  Rydex and ProFunds have since introduced “strong” and “weak” dollar funds.  We find the Rydex version most interesting because they use 2x leverage.  We have built a Rydex trading system that incorporates these funds and we are pleased with the results.  Note:  Because these are new funds it is necessary to build synthetic test data.  This is easily done with a tool in FastGraph.  

Given market conditions the systems all performed reasonably well.  We all would like to see great results each year, but preserving capital until sustainable trends develop, either up or down trends, is a goal of investing.

Other Mutual Fund Systems of Interest

Potential FastBreak users often ask about other systems such as Rydex and Fidelity Sectors. Here are some other systems we traded or monitored during the full year to provide additional information. These trading files for these systems are not posted on our web site.

System

Return%

MDD%

UPI

Annuity (4 funds no signal)

4.8

-7.3

0.9

Rydex w/RUTVOL signal

5.0

-6.3

0.62

Rydex no signal

24.4

-9.8

5.4

Select w/RUTVOL

6.0

-5.1

1.4

Select no signal

16.6

-13.0

2.3

ProFunds (FastBreak signal)

21.7

-12.5

3.2

The annuity system holds four positions in a trading family of 45 funds. The trading family includes a very diverse group of funds including international and bond funds. This is the same system that returned 17% in 2004, 60% in 2003.  Several users contacted us about this program and we supplied them this trading system.

Rydex and Fidelity sector system performance was mixed.  These are the same two systems we reported on last year. The systems that didn’t use a timing signal did very well for one reason – energy funds.  The systems using RUTVOL didn’t perform because the signal was out of the market during much of the energy fund rally. We haven’t been trading the Select funds and the Select systems are very out of date. We report them here to show that even old systems don’t “blow up” like many trading systems.  The Sector system that doesn’t use market timing had very respectable performance.    

The ProFunds system has been trading real time since June 2004 and has averaged nearly 30% annually in real time.  The system uses the market timing signal described in FastTrack Commentary 8161.  This signal was built using the signal building option in Version 5 of FastBreak.  The bulk of the gains in 2005 were from energy, Japan, and biotech funds.  

We are beginning to feel comfortable with the ProFunds and Rydex funds.  For years we have been hampered by lack of long term data.  The availability of more data and new products from these two companies is very welcome. 

Stock Trading Systems

In Commentary 8161 we described three stock trading systems built using Version 5 of FastBreak Pro.  We had one strikeout, a triple and a homerun with these systems.  Please review Commentary 8161 for background.

The absolute star was the energy stock trading system.  It returned 66.8%.  Compare this to Fidelity Select Energy which returned 52% and Select Natural Gas returning 45%.  The energy trading signal performed well with one exception.  The energy signal gave a sell a few days before hurricane Katrina.  The signal got back into energy stocks very quickly but those few days reduced the system return.  If the energy system had stayed in the market during those few days the return would have been 75%!!  The only change we made to the system since writing Commentary 8161 was removing two low price (< $1 price) stocks from the trading family.   

Our triple was the closed-end country fund system.  Our best performing system over the years has been the open-ended international mutual fund trading system, but as noted earlier, restrictions have been placed on many of the mutual funds.  The closed-end fund system described in Commentary 8161 returned 19.7% which is very close to the open end system. 

Finally, our strikeout (well, maybe a walk).  The Large Cap Stock system that trades a family of S&P 500 and NDX-X stocks lost 9.8%  The problem was that the market had two sell offs that were just strong enough to trigger our market timing signal.  Our S&P and OTC bear strategies went active at about the same time and started to enjoy good profits.  However, the market sell off didn’t continue and reversed to the upside, but with weak rallies.  These two fairly small whipsaws were just too much to overcome in a year will little market upside.  The stock trading system that didn’t use market timing returned 15%.  We have studies that would indicate a small position in the system without timing may be a reasonable venture.

In Commentary 8161 we described combining the three low correlation stock systems (energy, international and large cap stocks) along with our short systems.  In the Commentary, we simply averaged the equity curves.  A better way is to trade into the bear market systems if the large cap stock system is in money market.  The bear systems had little effect either way.  If we combine the system how did we do?  The combined equity curve is called SEI (Stocks, Energy, and International) and is shown in the following figure.  Note: The combined equity curve may look volatile compared to the S&P but recall the discussion regarding extremely low market volatility.

The overall statistics would be:

·        23.2% return

·        -13.3% MDD

·        2.9 UPI

We are very pleased with these results.  It was a difficult trading year and this was our first attempt with the new market timing signals and stock trading systems.

What about 2006?

Some of our favorite trading systems over the past 2-3 years have featured bonds (treasury and high yield) and real estate funds.  These market segments started to look tired in 2005.  We are excited about the new ProFunds “inverse” real estate fund and inverse high yield bond funds.  Rydex and ProFunds have funds that short the treasury bond market and these funds may come into play if the bond bull market starts a substantial downturn.  As mentioned earlier, systems using the new strong/weak dollar funds look promising.  We think there is definitely a place in our trading systems for these funds.

Another favorite area has been international funds.  Both the open end and closed end strategies did well in 2005.  During the year we designed a new closed-end system that uses a closed-end country fund timing signal of our own design.  We built this signal doing the following:

·        Used FastTrack to build the “average” equity curve for a family of closed-end funds

·        Exported the family average as an FNU file

·        Used FastBreak Pro to build a timing signal based on this average FNU

·        Used FastBreak Pro to build a trading system using the close-end country fund family and the timing signal

This system returned nearly 26% but not all the data was out of sample for the full year.  Long term historical performance is very similar (~23%/annual) for both the signal and non-signal trading systems.  We like the signal version because it is out of the market during major down turns.

Energy stocks/funds have been very good to us the past two years, but all large sector moves finally end and we like the new inverse oil sector fund just introduced by ProFunds.  We hope more of these innovative products become available.  We have had success creating long term synthetic fund data to build and test our trading systems when these new trading products have been introduced.  FastGraph has a tool to build synthetic data if an underlying index is available.   

We want to continue with stock trading systems and new signals.  We want to find new systems with low correlation to existing systems, for example, we have been experimenting with gold stock trading systems.  Gold stocks often have a low or even inverse correlation with many other trading systems. We believe we have a workable gold signal; however, even the largest gold stocks are thinly capitalized and subject to wild swings.

Finally, one of the problems with sector fund investing is finding a suitable market timing signal.  There are instances when it is better to be out of all sectors, but the majority of market timing signals don’t work well.  We saw this in 2005 with the RUTTR and RUTVOL signals.  They provided very good signals for the general market, but the Rydex and Fidelity sector systems that used these signals underperformed because the energy funds were very strong during the time periods the signals were on a sell.  To try and solve this problem we have combined two of our signals.  The general stock market timing signal, described in Commentary 8161, is combined with the energy timing signal also described in the same commentary.  The signals combined (there is a tool in FastGraph that easily builds composite signals combined in the described manner) so that the combined signal stays in the market if EITHER underlying signal is on a Buy, and only sells if both signals are out of the market.  The intent is to stay in the market if either the general market or the energy sector is positive.  Early results from a Fidelity Sector system look positive. 

What is FastBreak and Who needs FastBreak

Unlike other trading methods, the FastBreak "mechanical" trading strategies gives specific buy and sell recommendations. Mechanical trading systems leave no room for historic "revision". These types of trading systems may not be for everyone, but we believe they should be considered if you recognize yourself in the following:

  • Do you have trouble "pulling the trigger" on trades, e.g., selling your losers, taking profits on your winners, waiting for a bounce to get out of a bad trade ...?
  • Do you have a limited amount of time to devote to your investments? If you are a full time investor, or enjoy sitting in front of CNBC and your computer all day, you may be able to do better with other investment methods. However, if you have limited time (or better things to do with your time!) 10 minutes an evening is all you need to download your FastTrack data and run your FastBreak strategies.
  • We have stopped reporting on some of the older FastBreak systems for a variety of reasons: Many of the mutual funds used in these old trading system families have since been discontinued, the market timing signals (unrelated to FastBreak) used did not stand the test of time, and we can build much more effective strategies with the functionality added to FastBreak in recent years. We felt that those systems don't accurately reflect the performance that FastBreak can produce.

Edge Ware, Inc.

Standard Disclaimer

As the saying goes, "Past performance is not a guarantee of future results."


l>