Issue 4

Bollinger Bands as a Filter - Dr. Michael Miller

In our ideal world there would be a method for eliminating trades that had a high probability of generating a loss.

From June thru July, real time trading using Swing Catcher Daily files generated profits of \$14,362 on 52 trades and \$10,862 on 61 trades in German Mark and Swiss Franc, respectively.

One might consider it a foolish endeavor to try to improve a system that has generated profits in excess of 800 per year, but the use of Bollinger Bands clearly eliminate some trades that would have produced loses.

Bollinger Bands are upper and lower envelopes around the 20-day moving average. The bands are based upon 2 standard deviation units of the commodity.

Prices tend to remain confined with the upper and the lower bands. Any move outside the band indicates a short-term overbought (sold) condition. Consequently, it would not be recommended to initiate a new position at this time - set and forget computerized stock trading.

Charts in Print Copy

The above chart represents March German Mark for the month of January. On January 25, Swing Catcher generated a long position. The following trade if taken would have generated a loss of \$1,387. Clearly, the price is above the upper band and one would assume that the DMark is overbought.

In order to quantify the overbought/oversold situation, the following formulae are used:
long positions--(upper band-price)/price
short positions--(price-lower band)/price

In the above example, the upper band is 62.50 and the closing price is 63.00. Consequently, (62.5-63.00)/63.00=-.79

This number quantifies how far above (or below) the price is from the band. Further research needs to be done in order to determine the optimal filter level.

The following represents all trades that occurred when the price was outside the Bollinger Band for both Swiss Franc and German Mark.

In summary, 16% or \$4087, in losing trades would have been saved if one had avoided making a trade while the price was outside the Bollinger Band.

Charts & Table in Print Copy

Review of Buran's Grand Combo - Matthew Chiang

I recently responded to Bob Buran's promotion and was impressed by his "record book" of trade statements. I also reviewed the video tape (preview but no methodology disclosed), and I talked to Bob himself before I decided to buy the system.

I believe Bob is an honest guy, and he was helpful in answering my questions. I was very concerned with drawdowns and consecutive losing trades, so I specifically asked Bob on this aspect, and his answer was "once in a blue moon."

I then bought the System. However, when I ran the software and studied the logic, I found that the system was not what I had expected.

Suffice it to say that Bob's Grand Combo system works similar to a volatility breakout system (VBS). Like any VBS, Grand Combo can get a lot of whipsaws if the market is non-trending, but overall, consecutive losses are mild and manageable (there are exceptions, of course).

For each trade, Grand Combo has a volatility-type and statistically sound stop-loss, but the program is not specifically engineered to reduce risk in this aspect, except an advise to diversify.

I have used Essex's ACE Currencies (a VBS system) and it performed poorly in this aspect (7-9 consecutive losses, could you believe this!). Plus, Futures Chart's Mini-Trend Setter also has significant consecutive losses (3-5 while trying to catch the major change in trend). So Grand Combo is better than others.

Overall, I should say Grand Combo is average to above-the-average on loss management - typically 2-3 losses in a row, sometimes as much as 4 losses, but they are not that frequent. Then you have 2-3 or even 4 wins to patch up.

So what's the catch? Grand Combo has a weakness I cannot accept. Unlike other VBS or trend-following systems, Grand Combo has very tight profit targets.

I was rather surprised that for some markets, the profit target is typically some 30-50 ticks, and you will be lucky to get over 100 ticks (depending on market, of course, the BP has more ticks, for example).

I ran some historical tests, and found that most markets made money, as are claimed. But I was not that impressed because most of the trades made very small profits, and you have to trade a lot to make enough money.

The average profit per trade (again, depending on markets) is \$300-\$500, some are over \$600 and a few are over \$1,000. The losses are similar.

I also dislike the idea that Grand Combo pays no attention to underlying market trend. It simply trades breakouts on either side. You could end up buying near market tops or selling near market bottoms.

I always struggled whether to obey the counter-trend signals or not while in a defined trend, or seeing a top/bottom is near. Even if the signals are in trend, you usually end up buying/selling near swing highs/lows.

This is common to VBS. But unlike other VBS, Grand Combo does not attempt to ride the trend. It is content to take small and quick profits.

Most trades are typically short term, from 2 to a few days. That's why the average profit is small. Even though the tests have made paper profits, I wonder if I could trade the system with such attention and frequency.

Bob and the brochures do say that you don't need to watch the market, it is true. But with such tight profit targets and trade frequency, I don't think I can afford a full service broker to watch the trades for me.

A plus about Grand Combo is, while wins are small, losses are also small, only occasionally do I have losses over \$1,000 per trade (Grand Combo does not trade S&P or its like).

After paper trading the system for 5-6 weeks, I decided to give it up. I pay much attention to drawdowns and consecutive losses, because I am a novice trader with limited experience and trade capital. I don't yet have the nerve nor the pocket to take 2-3 losing trades in a row while scalping small profits on others.

It would be different if I have some big winners to cushion the losers, but I don't see this a frequent event in Grand Combo. I prefer trading for short to medium term (a few days to weeks), and perhaps have some that can ride the trend.

Also, you need around \$25,000 to trade Grand Combo effectively because it's diversified among 30 markets. Somebody suggested that Bob's good results were not achieved 800 mechanically, and judgment calls were involved. There has been some controversy over this and other issues (on system drawdowns) in Club 3000 recently, and Bob has spoken a few times to defend his case, I believe one allegation is false.

I do not want to comment on this, as I respect Bob as an honest and good trader, but I would hesitate to put down real money trading his system, unless I cannot find another better one.

The principle behind Grand Combo is sound, and I think it works, but not for every style. It would be very suitable for very short-term traders who require a lot of action, and perhaps more suitable for overnight traders with 50 to 60 tick targets.

Observations & Stop-Loss Technique - Brian Radke

The following statements are my opinion only:

In thin, wild markets, such as "Bellies," do not enter "on the open." It is not unusual to see 40 to 70 point opening ranges in this market! Try to time your entry to catch as much favor in your direction as you can.

A little over a week ago, I took a "Swing Catcher" signal which later proved to be right, and entered about the time of the day the lows were usually "in" by. The market slapped me with a \$600 loss in about an hour (I was trading smaller than the recommended stop), but I believed in the signal and re-entered.

This is a vital key to any new trader. You must not become gun shy in markets such as these. If you have done your homework (or even if you haven't and you own "Swing Catcher"), you should begin to have a feel for the market you are attempting to trade, and re-enter after being stopped out. That re-entry was a winner!

It was simply a matter of believing and not being scared. Futures are not for you if you are searching for 800 winners. But again, don't feel you have to enter on the open.

The following is a little stop placement yardstick I use in some markets. It has just experienced some good results.

To avoid redundancy, this idea is based more on the open and closing prices to aid in stop placement. Also, I try not to use daily or true range as this again would be redundant, as most traders have thrust or breakout systems already in place although you could use an average daily range as a multiplier.

This is not a system for entry timing! Please refer to Feb. Bellies from 07/16-07/26. This only works during a trend and is not designed to determine the direction of the market, rather simply a guide to good stop placement. Look also at soybeans, gold, cotton and whatever else you wish.

I have this programmed on a spreadsheet for ease of use. The following is for a "bull" trend. Begin using after a confirmed swing low and once you have entered long. It is extremely simple.

In this idea positive or negative is not the point, treat all figures as positive.

1. Subtract yesterday's open from today's open=
2. Subtract yesterday's close from today's close=
3. Add the resulting numbers from #1 & #2 together and divide by 1.8
4. Subtract the answer from #3 from today's close. Place your open order stop at this point.

Reverse the process for short positions, i.e. add to today's close.

When today's stop is much below yesterday's (above in a "Bear"), leave your stop where it is to protect profits. This often signals the end of a swing! Don't be afraid to re-enter if the trend continues however. Markets such as bellies are known to bend your eyebrows often. Hang tough.

Variations:

2. Use a balance price (0+h+l+c)/4, minus a percentage of today's range
3. Use a 22 day (or any length you wish) average range in place of the 1.8 multiplier.????

I too, like many seem to have a weak spot for "ringed high or low" systems, and that is what Trident is. I believe the mesmerizing effect is due to the ease in which anyone can look at a chart and see without question, P1, P2 and P3. I am not about to slight Charles Lindsay, as it is very apparent that he is very articulate and intelligent.

I've literally, without exaggeration, spent well over 100 hours on my computer trying to make this type of system work in many markets consistently.

"Trident" is based on the "zero sum formula." This algorithm has been used for decades by statisticians and mathematicians as a measuring tool, and a reader should not be confused into thinking that it is or was designed for commodities as a market system.

It is foremost a yardstick, and one should not get overly excited about its predicative capabilities. It is very easy to see where all of the rest of the system, i.e. CP, EP, SD, etc. come from. They are all derived from looking at the personality of the markets and applying fail-safe in the form of equations.

I have no problem with this approach, but it is very apparent that the system becomes very cyclic in nature.

Leading cycle researchers always relate the same dilemma, cycles seem to disappear as quickly as they appear, and no one can seem to predict a market's swing high or low with any consistency in many markets. This is why there are exceptions (P2=I3 etc.) in the rules of these types of systems.

Read the first few chapters of the book and then the last few. You can see that the author has almost totally abandoned the ringed high, ringed low, approach, replacing it with "factor analysis" which is more similar to where our many oscillators of today come from.

Trading "Trident" must encompass all of its tools. One must use B2 & B6 and most importantly the variance rules. Even with that done, you will miss many moves and constantly be second guessing your next move, because of the subordinate nature of the minor swings.

Look at Jan-March Cotton. "Trident" made fantastic three day trades then came unglued at the seams in April. Even having said that I can say I enjoyed the book and it is very good reading.

FINDINGS: Use DP and TRP as an oscillator and not to form a price window. When trading minor swings, any change in measurements are helpful.

Since a ringed high or ringed low is not confirmed until the next day it is too late, so one must try to tilt the odds looking at today's prices. I had a system in place long before reading "Trident" that is surprisingly like DP & TRP.

In this system, I do not have to wait for confirmation of a P3. I have it programmed on a spreadsheet and my computer does all the daily crunching. It is not by any means perfect, but I'll share the portion of it that is similar to "Tridents" DP-TRP.

TRY THIS: Take nine days of prices, with today being day #1. Now construct two opposing Tridents using max and min. prices over a specific time frame, (9 days in this example). The multiplier used in "Trident" (0.625) can be changed! I have various figures "tweaked" for various markets.

You'll need to create 4 columns;

column # 1=call this DP #1 (or whatever you wish).=This is the lowest low of days 3, 4, 5 & 6, (remember today is day #1), minus the highest high of days 6, 7, 8 & 9. Multiply this by 0.685. Then add the highest high of days 1, 2, 3 & 4.

column #2=TRP #1=Lowest low of days 3, 4, 5 & 6, minus the highest high of days 1, 2, 3 & 4. Multiply this by 0.795 then add the highest high of days 1, 2, 3 & 4.

column #3=DP #2=highest high of days 3, 4, 5 & 6, minus the lowest low of days 6, 7, 8 & 9. Multiply this by 0.675 then add the lowest low of days 1, 2, 3 & 4.

column #4=TRP #2=highest high of days 3, 4, 5 & 6, minus the lowest low of days 1, 2, 3 & 4. Multiply this by 0.795 then add the lowest low of days 1, 2, 3 & 4.

What to look for: In a normal market column #1 should be below column #2 and thus #3 below #4. Instead of looking at the actual number each column offers, I choose to forget about the "window" and look exclusively at the relationship of the columns.

On the day that #1 goes above #2 AND #3 goes above #4, a minor swing top is occurring. Note that #3 may be above #4 for a few days before #1 goes above #2 and vice versa. On the days that #1 falls below #2 and #3 below #4 a minor swing bottom should be occurring. Use this as a reversal system very short term (even day-trading).

Couple these crossovers with 4 day highs and lows and you have a fair to average system, i.e. #1 goes above #2 and #3 above #4 and today's high is a NEW FOUR DAY HIGH, sell the next open. The opposite is true of course for long trades.

DRAWBACKS: This type of system trades blindly at times and gives no help during a strong trend. It does not work well in all markets with one set of multipliers.

It should not be traded without looking at daily support/resistance and momentum. I have many more facets built into my system but am offering this much to hopefully inspire someone to crunch the numbers for themselves, and maybe get some new ideas to a very old problem.

I would not feel at ease trading this system alone, although it can be brilliant at times and I do look at this when I get a computer generated signal I feel uncomfortable with.

I have two other systems to handle the facets that "Trident" type analysis overlooks, and the three together are very helpful.

I strongly encourage anyone to spend the time and effort to research on their own. System writers are not always smarter than the average trader, rather they usually have spent countless hours staring at a computer screen to come up with new and workable systems and hoping to find that "right combination."

What do you suppose might happen looking at Swing Catcher Harmonic files????

When I decided to turn my hobby into a full-time career, I checked out several quote services. My location near Winnipeg, Latitude 50 degrees North, ruled out telephone and hand-held quote services. My trading did not require "live" quotes, nor did the size of my account justify the added expense. The choice easily narrowed to only two vendors. I chose Futurelink because it had connections I was familiar with - Futures Magazine, Futures Update, FutureSource - and because the service supplied my basic requirements.

FutureLink provides futures and futures option quotes for all U.S. Commodity Futures Exchanges and also the Winnipeg Commodity Exchange. The New York markets are on a half-hour delay basis, updated every 10 minutes. The rest are "snapshots" every 10 minutes.

FWN supplies current news updates from around the world relevant to all futures, as it breaks, to assist in making decisions. FWN also has regular reports from each exchange separately for most commodities, so traders can be aware of what is happening.

Several important technical indicators are calculated and transmitted after the markets close each day. Daily, weekly and monthly bar charts for each commodity and contract month are updated daily. An overlay indicator assists in checking back on prices, volume and open interest. Recommendations and market analysis are provided by selected analysts. Supplementary commentaries and current weather maps can be added at extra cost.

A custom screen can be set up to keep track of positions along with profit and loss factors. Another custom page can be set up to follow specific commodities through the day without having to go to each page. The entire program is simple and easy to use, has most information required for regular trading, and is well recommended.

There are times when one might like quotes more often, a wider range of technical indicators updated through the day as needed, move overlays on the charts, faster reporting of news at critical times, and access to stock quotes, but of course these would be more costly and available through other services available through Oster Communications. Some computer programs are available through Futurelink to enable end of day quotes to be used in various spreadsheet and charting programs.

Support is provided through a toll-free number. There seems to be sufficient staff, all knowledgeable, very helpful and courteous. In most situations, I have found this to be a very good service.

The problems I had were in receiving the signal. Subscribers must supply their own satellite dish and computer. The receiver itself is purchased from FutureLink and imported with appropriate duties and taxes paid. Once purchased, it is yours. The recommended receiving dish, a four foot offset, is also available through FutureLink and may be cheaper than purchased locally even with duties and taxes. Subscribers should compare before committing.

This dish is equipped with a simple feed horn which has a clear plastic bubble on the receiving end. This bubble fogs up at any opportunity - cloudy days, rainy days, misty days and overnight. The signal cannot penetrate this fog. Either you wipe it clean several times a day in poor weather conditions, find some way to cover it, or do without a signal.

I installed the dish myself. It seems local installers have these in their catalogs but have had little experience with them. On questioning them about the foggy bubble, I was advised occasionally to remove it and not to be concerned about dust, dirt or bugs making webs.

I have my doubts considering the vagaries of the signal. This dish is recommended because it picks up the KU signal very well. FutureLink is not available on the C band, only KU band from Satellite K2 and difficult to get from this latitude and region.

Apparently it is also available on the ANIK satellite, but I got very poor reception when I tried it, not to mention the expense of importing and returning the trial receiver. While C band signals are forgiving and allow for some degree of misalignment, the KU band signal is not.

At this latitude your dish must be right on the signal - exactly! If there is any deviation - at all - there is loss of signal and you might get the quotes but not the news stories or charts.

For better or worse, my dish was struck by lightning, frying the dish, the receiver, and the I/O card in my computer. On questioning, installers scoff at protecting your dish from lightning. You be the judge. I was out of business for several weeks getting straightened around.

I went to an Alpha 10 dish with an all-weather protected feed horn. If there was any room for deviation with the other dish, there isn't with this one. It must be exact.

Be prepared to adjust your dish after a severe storm, large rapid temperature changes, and in deep sub-zero weather. Some days the signal does not come clear until after the sun is up for awhile. Forget about news stories during a heavy snowfall or rainfall.

The signal receiver is virtually trouble free. Your computer can be a basic XT or AT with a hard drive, but increasingly there will be more sophisticated transmissions such as color weather maps so you will need a more advanced unit.

Although you can use the computer for other duties you must run the resident program 24-hours a day to avoid missing any information pertaining to the commodities you follow.

Occasionally the program stops dead with no data units getting through. It wasn't a problem until an updated program was sent out. It can happen anytime during the day or night. I am assured it is not the program's fault. Luckily it is relatively easy to remedy.

At first, I used to call FutureLink about the poor reception on certain days as I didn't understand that clouds in the sky or frost in the LNB creates unsatisfactory reception. It was not very reassuring to hear that everyone else in Canada was getting satisfactory signals.

There seemed to be no explanation as to why my equipment was functioning unsatisfactorily. By chance, I learned that some people north of me were using the ANIK satellite and were getting very good signals. Well, I got worse when trying it.

Releasing the names of subscribers close to me to contact for discussion about the FutureLink reception seems to be a breach of confidentiality. I have to conclude there aren't any, no one is using it to trade, no one cares about the reception, or they are making the best of a poor situation.

I really like the FutureLink program and until I am ready to advance to more sophisticated programs, I would like to keep it. However, I am not happy with the reception.

If anyone is considering subscribing to the service, or already has it, and would like to discuss it with me for western Canada, I would welcome their calls or correspondence. Perhaps someone has some useful suggestions for me to use. I would appreciate any input. Now really, can I be the only one with these problems?

Using Ensign/Bonneville/CSI Data Formats, Data Conversion Frank Seitzinger

I can use my Trading System as early as 3:34 PM, after the markets close, by converting from my Bonneville Quote Machine, from their data format to Swing Catcher's CSI data format.

I use Swing Catcher's Utility/Edit menu selection 7 from the main menu. Then press F2 key and then press the F key to edit a QMaster file.

Next, select the file number to edit. Then simply change the file name to conform with the Ensign/Bonneville data format, for CSI/Ensign conversion.

In the Ensign System Setup Program, be sure to change the CSI to your new file name. Plus, add lower case c to the historical update time, i.e.: (1545c), for Ensign to convert data to CSI format with your CSI data file of choice.

I subscribe to Trend Index Trading Co. monthly parameter-on-disk Service. When I get the disk I install both parameters and included data files for the easy way to be sure all my data is correct and I have the current parameters.

P&L Report on Bob Buran's Grand Combo System - Matthew Chiang

I choose only 17 markets because I did not have time to follow more. Out of the 17, I only choose around 4-5 for actual trading.

Of the remaining 10 markets that Grand Combo trades, there are 3 markets that I don't have data on: Platinum, NYFE, Unleaded Gas. Cocoa's data is so bad, I discarded the results.

Since Genesis data is not 'clean', and that I do not follow the 10+ markets, I did not attempt to correct any error in the data files. Even for the 17 or so markets that I follow, I am uncertain if the data is clean enough.

Sometimes the error is very obvious and I can correct them daily. Other times the errors are not that discernible, and I don't even know where they occur. Therefore, the test results may not be a true reflection of the system's performance, but I think they are close.

From \$60.00 to \$80.00 round-turn slippage, and \$30 round turn commission is what I am experiencing now, because I don't trade that often to get good attention.

I ran some more tests today and have come up with the following list (appearing at the top of next page): Chart in Print Copy

(NOTE: These figures make NO ALLOWANCE for slippage & commission, which has a major impact!)

The excellent research done by Mike Miller on Bollinger Bands reveals how a filter can effectively be used as a way of making another methodology work even better. In this case, Mike is using Bollinger Bands as a way to bypass some losing trades, by avoiding trades if price is outside the Bollinger Band.

Mike's methodology seems to have excellent merit as a way to make his system work better. The techniques are also likely applicable to many other trading systems.

Matthew Chiang's comprehensive contribution and review about Bob Buran's Grand Combo goes into great detail. He must have spent lots of time working on it.

From what I understand and have been told by very reliable sources, Bob Buran is likely the number one trading system vendor (based on both sales volume and dollar amount of sales, during the time period he has been selling systems) of all time.

System Writer/Trade Station may have sold more copies but they are not Trading Systems, per se. They come under the classification of 'Toolbox' type of Programs, where the user can design his own system using the programming shell. Grand Combo/Trabos is a stand alone system that does not require the user to design the methodology or make algorithm modifications.

I believe Buran's System was available to the public for the first time less than 2-years ago. From what I gather by talking to numerous clients of Bob's and from very reliable sources, an amazing number have been sold.

It is interesting to note that very few opinions on Buran's methodology have been written. A large number of commodity traders have purchased Bob's Grand Combo/Trabos, yet very little has been written about it.

The readers of CTCN would be greatly interested in reading more reviews of Buran's methods.

Trading Observations and Stop-Loss Techniques submitted by Brian Radke are very interesting. Brian says that re-entering a trade in the same direction, even after being stopped-out at a loss, can be very profitable. That seems to be true, and in fact I have noticed it seems to occur often. However, one problem with re-entering after a big loss, is the fact many traders lack the discipline or nerve to actually do it!

Brian's stop-loss methodology could be quite valuable. It deserves to have some extensive research done on it to determine its validity. Does anyone know why Brian is using a division factor of 1.8?

Brian Radke's in depth research based on Trident Strategy, reveals he has put a tremendous amount of time and effort into researching Trident. He has shared with us some very detailed technical methods that may get the Trident Strategy to work better.

It looks like James Weber spent lots of time and effort trying to get FutureLink to work properly (apparently unsuccessfully). It seems very surprising that Jim says he 'likes the FutureLink Program' (second paragraph from end of his review), in spite of the many problems and great aggravations he seems to have experienced.

However, Jim also states he is not happy with the reception, which of course is the most important item!

Note: I experienced similar problems involving DTN. The most serious problem was losing data due to weather conditions.

Frank Seitzinger's ability to convert to CSI Data Format from Bonneville/Ensign should be applicable to many other software programs that use CSI Format.

Matthew Chiang's spent a lot of time putting together his detailed Grand Combo P&L reports. They do NOT include any allowance for slippage & commission, which unfortunately is a major factor in commodity trading! Matthew say's he averages \$60-\$80 slippage and \$30 round-turn commission per trade. Upon updating Matthew's P&L data by taking into consideration his (low-end) slippage and commission deduction of \$90.00 per trade, the figures change as follows:

Total trades 766 X \$90.00 for slippage & commission per trade=-\$68,940.00 deduction. Thus, instead of Net Profits of +\$31,857.00, the adjusted Net Profits becomes negative -\$37,083.00. Quite a significant difference!

Some trading systems only deduct \$50.00 (rather than Matthew's recommended \$90.00/\$110.00) per trade, for slippage & commission. However, even if we only deduct \$50.00 per trade for slippage & commission (766 X \$50.00=-\$38,300.00), from the +\$31,857.00 Profits, we still arrive at negative Net Profits of -\$6,443.00.

The accuracy of the Grand Combo P&L reports submitted by Matthew, is not and can not be assured.

I am sending a copy of this Newsletter to Bob Buran and asking him to check the P&L reports for accuracy and make comments, if he desires. If Bob responds, his reply will be published in the next issue of CTCN.