PROFIT TAKING WITH LOUIS MENDELSOHN
Over the years, Lou Mendelsohn has been not only a persistent presence among trading system vendors, he’s also stood out as one of the few who advocated complete disclosure of the systems being sold. Not content to disclose his own system, ProfitTaker, and its refinement, ProfitTuner, he’s also written and spoken out publicly on the hazards of buying undisclosed trading systems. He’s also one of the most realistic voices on the fad of optimization.
With ProfitTaker now well established and his family comfortably ensconced in a Florida ranch, Lou’s turned his attentions to — horrors! — quantifying the impact of new fundamental information on tradeables. To see where all this might be leading and to check on any possible softening of his opinions, we corraled Lou on the phone for several hours a few months back. Here are the results.
Can we go over a little bit of your background, and how you got into this game? Have you always been a trader?
I had been trading equities for nearly 15 years before I got into commodities in the late 1970s. At the time I was a hospital administrator in Tampa, Florida for one of the large hospital management chains. With my M.B.A. background I was very quantitatively oriented. I bought and tried all sorts of trading software programs, many of which turned out to be a waste of time and money. Some didn’t even work. Initially I was using CompuTrac, but when gold and silver took off I quit trading stocks and options altogether, and started designing my own software specifically for commodities.
Was this analytical software, or was it a trading system?
Both. I consider analytical “tool box” software and “black box” or “gray box” system software seriously lacking from a trader’s standpoint. The analytical software has so much flexibility that it encourages traders to jump from one set of technical indicators to another. Then there’s “black box” software which doesn’t disclose any of its trading rules, and “gray box” software with some rules disclosed but others kept secret from the user. These undermine the discipline that is so crucial to successful trading.
Do you feel that’s been successful, that the use of computers has eliminated the human element here?
I never intended to eliminate the human element, just the emotional, undisciplined element. In my opinion, trading software should be viewed as a tool in the hands of a thinking person. If the software’s trading rules are logical and completely disclosed, and can be customized by each trader to his or her own trading style and risk propensity, then this will cut down on the indecisiveness and lack of discipline that so often results in unnecessary losses and missed profit opportunities.
I know a lot of traders who get their signal but then go against it, or fail to execute. I hear from so many people who just don’t have the discipline to consistently pull the trigger.
There are a lot of that. Back in October when the stock market crashed, the top performing model from my monthly ProfitTuner report gave a sell signal on the S&P 500 on the close of October 9th to go short on Monday’s open, October 12th. That’s a week before Black Monday. Many clients called to say that they took ProfitTaker’s short signal then. But some clients, who didn’t have the guts to take it, were kicking themselves afterwards for overriding the short signal and standing aside. At least they weren’t long, but look at the enormous profits that they would have made if they had pulled the trigger. Fortunately, a lot of others bought puts based on ProfitTaker’s short signal.
They were confident enough to buy puts on the short signal?
Yeah, but even then some didn’t make as much as they could have.
What markets do you prefer to trade?
I’m mostly interested in the currencies, bonds and stock indexes.
Seems to be a favorite area for computerized traders. For some reason those commodities or futures contracts seem to behave better than some others.
They have been known to have more of a trending characteristic.
Have you studied that characteristic?
Well, of course, ProfitTaker is to a great extent a trend-following approach. It’s designed to catch breakouts and latch onto trends. It’s constructed so that it can’t miss getting onboard a trend and riding it. Recently I’ve been working on a new software program called Trader, which is designed to correlate fundamental econometric information with the trading direction of various markets.
Can you give us some conceptual discussion of that?
Yes. What I’m basically doing is looking at government economic reports from Japan, West Germany, England and the United States. Also included are reports from some of the other trading partners like Hong Kong, South Korea, Taiwan and Singapore. In addition, the software considers a variety of private surveys and other economic and sentiment information that is reported on the news wires or published in the media. It then uses this fundamental information in a technical way to help the trader make quick trading decisions in various markets, based on the recent causal relationships that exist between the economic information and these markets.
Can you give us an example of that?
Sure. Let’s take the U.S. trade balance as an example. First of all, the market develops an expectation before the report is actually announced. This is usually based on surveys of prominent economists, institutional traders and other financial analysts. Their estimates take into consideration seasonal factors and economic information from other related reports. For instance, the Japanese and West German monthly trade reports give a clue to what the next U.S. trade figure will be. Then the market discounts some expected value, which is either higher, lower or unchanged from the previous report’s figure. This affects trading in various markets including the dollar, currencies, metals, stock indexes, bonds, etc. The software would give the directional impact that these three possible expectations would have on the markets so that the user immediately knows what direction to trade. All of this would be instantly available on the screen. Of course the user could edit, add or delete any of the report items or markets. He or she could alter the directional relationship between the items and markets, and designate some markets as “influence items” on other markets.
O.K., then what happens after the report actually comes out?
When the report is announced, the figure is either the same as expected, higher than expected, or lower than expected. This has immediate implications, particularly for day traders. They could take advantage of the situation if, for instance, the trade deficit was expected to widen, but did so to a lesser extent than expected.
How do you measure those impacts? I take it you’re looking at past experiences and trying to find the relationship from that.
Yes. I’m looking at the recent historical relationships, and simply organizing and presenting all this otherwise disparate information on one screen.
Would that be some kind of blind parameter or would the user know what the relationships have actually been defined as?
The software gives him the directional relationships that have existed. However, if he wants to develop additional relationship scenarios for the same market, let’s say under different stages in the economic cycle, he could add these to the software for his own customized use.
Are these relationships mathematical?
Yes, but for trading purposes they only need to be directional.
How do you see clients using your Trader software?
Well, I think it will be dynamite as a stand-alone tool, especially for a day trader and anyone involved in trading the financials. It would be very useful for a longer term investor to help him decide when to restructure his portfolio.
A futures trader could also use it as a filter to what he’s presently doing. For instance, as long as the trade figures are bearish on the dollar, the trader could use this information to avoid taking long dollar positions or short Yen or Mark positions. In fact, an aggressive trader might even sell short on sharp rallies in the dollar when he gets a buy signal using tight stops to protect himself.
To switch subjects, how many people subscribe to the ProfitTuner, if they’ve already taken ProfitTaker?
A lot. ProfitTuner is valuable because it gives the trader an up-to-date picture of over 20 different markets each month. In many instances, it calls attention to profitable opportunities in markets that the trader might not be following himself. This way he can start trading in a new market immediately with a profitable model. It also cuts down substantially on the amount of time a trader needs to spend researching and testing models.
So there’s a substantial number of traders who would prefer not to do a lot of historical testing or optimization.
Well, ProfitTuner is not designed to completely eliminate testing. It’s complementary. The most effective way to use ProfitTaker is to use the ProfitTuner models as a starting point. As the month elapses the trader should do some fine-tuning and testing on his own to double check the models and to see if they can be improved.
How much back data do you recommend they use?
ProfitTuner presently tests the most active contract months with rollovers for nine months. However, I’ve been researching this issue for some time. I’ve broken the issue up into two time periods, the historical testing period or “testing window” as I call it, and the real time trading period or “trading window.” I’m interested in finding the optimal testing period to use when developing a model, and the optimal trading period to use in real time before retesting and changing the model.
I would just guess from my personal experience that it would vary.
Yes. My research to date suggests that these window sizes are not fixed but vary according to the market.
Have you thought of defining windows by their bullish-bearish nature?
Sure. That’s another approach. Rather than going back nine months or one year or two years or whatever, testing windows can be defined for each commodity over different time periods. That way the trader can build up an inventory of unique models for every market under each type of trading condition, whether it’s bullish, bearish or sideways.
Do you do all this work yourself, or do you have a hoard of analysts and programmers?
I do all the research myself. Of course on the ProfitTuner, I do that on a mainframe.
How else do you spend your time?
Basically, I spend much of my time providing consultation to my clients, conducting technical research, and developing and testing out new ideas for incorporation into ProfitTaker and for inclusion in other investment software packages such as Trader. I have a very close rapport with many of my clients, and try to make myself as accessible to them as possible. I even have a special telephone number dedicated solely to this. I also do a fair amount of writing and have been invited to speak at various investment seminars and on Financial News Network. There are also a lot of joint venture proposals that come my way, some of which I am pursuing. One involves a book on technical analysis that I have been working on.
What’s on your front burner for this year?
Well, I’m introducing Trader. I’m also finishing up a revision to ProfitTaker. I’m also customizing ProfitTaker for a few institutional clients who have particular requirements, and am getting more involved in providing consultation to them.
Do you have a sense of who’s doing the most interesting new work in technical analysis these days?
Sure. I keep in touch with other traders and technicians around the country who are actively involved in researching new ideas for trading methods and systems.
John Bollinger, Larry Williams, Philip Gotthelf and Jake Bernstein come to mind. Last year I even considered setting up a technical advisory panel strictly for the purpose of kicking around research ideas and brainstorming. But the response was too much. Almost every one of my clients wanted to be a member of the panel. Don’t forget, I’m running a business. It would have taken some sort of university auspices to really get something that big off the ground and funded properly. The research alone could easily run into the tens of thousands of dollars to begin to do anything really extensive. So I had to put the idea of a formal research panel on the back burner for the time being.
So what are you doing instead?
I involve my clients in a more structured way, by sending them questionnaires before I start on a new revision. Then I can rank order what research projects and enhancements to ProfitTaker would be most beneficial to the greatest number of clients. Also, many clients stay in frequent touch with me on the phone. Some even try out different research ideas on their own, like filtering ProfitTaker’s signals with other technical indicators that they’re interested in, or testing day trading models on fifteen or thirty minute intervals. For instance, Terry Apple, a professional trader who has written quite a few review articles for various magazines on different trading systems, has done a lot of research himself on ProfitTaker over the past three years that he has been using it.
Has this been fruitful? What changes have you been able to make?
To be honest, I have been reluctant to tinker with ProfitTaker’s trading rules. At first many ideas look good, but through extensive research they have to demonstrate a substantial improvement in performance before I would add them to ProfitTaker. Terry called me up a few days ago, in fact, to say that he had just given up on testing a particular indicator which had initially looked promising at filtering out some losing trades. But the problem is that for the few bad trades avoided, other very profitable trades were also filtered out.
Are you saying that you’ve found the holy grail?
No, not at all. I’ve never said that. But ProfitTaker is relied upon daily by an awful lot of very sharp traders. Over the years it has withstood the test of time by making money. And that’s the bottom line. But it’s not a sure bet. It takes thought and effort, just like everything else in life.
So you intend to keep on with your research?
Absolutely. I expect to keep looking for ways to improve the performance of the software and encourage my clients to do the same. But as the old saying goes, “if it ain’t broke, don’t fix it.” I don’t fool with a good thing just for the sake of coming out with a new version. But I’m constantly on the lookout for ways to enhance ProfitTaker’s performance and share it with my clients.
Do you keep track of your competitors?
Who do you see as the major competitors to the ProfitTaker package?
I don’t really see ProfitTaker as having any serious competitors. I’ve been at this for so long already that I’ve seen many software programs come and go. It’s like a revolving door. What separates the wheat from the chaff is staying power. The only true test of trading software is how well it performs over the long haul, year-in and year-out in real time trading.
What about optimization? You’ve written quite a few articles on this subject. Are you still an advocate of optimization?
I’ve never been an outright advocate of blindly performing optimization. It’s very tricky business that most traders and even many software developers don’t fully understand. Too often software is designed with so many trading rules or indicators that the models get curve-fitted to the historical data. In other cases, traders ignore basic statistical principles and use a testing window that is too short to generate enough trades for a valid model. The end result is that optimization as a concept is blamed, when in fact it’s the way that the procedure is performed or faulty conceptual design in the software that contributes to the problem.
A lot of traders want the flexibility of using many indicators. Are you saying that this is bad?
Yes. For several reasons. First of all, it runs the risk of curve fitting. The object should be to find a model that captures the general tone of a particular market. A lot of the published research comparing the profitability of various technical trading systems found that over the long run the simple approaches perform much better than more complex ones. This makes a lot of sense to me, maybe because of my strong math background.
What do you mean?
It’s easy. The more trading rules or indicators used, the less likely the historical performance can be replicated in real time. Remember, testing isn’t the name of the game. It’s to make money in real time trading. Also, when there are too many optimizable models or trading rules available to the trader for analysis, he tends to get bogged down researching everything. This syndrome, often referred to as “paralysis of analysis” can become very addictive. But it’s a very poor use of a trader’s time and energy. That kind of spadework ought to be done by the software developer, so that the trader can devote his efforts to trading without spending an inordinate amount of time designing and testing trading rules.
Your system basically uses moving average oscillators and momentum indicators. Do you work with any other indicators? There’s 50 or 60 of them out there that I could think of.
Let me give you a little bit of background on ProfitTaker’s trading rules. I didn’t set out to recreate the wheel. I found from my own research and from other published research comparing various trading systems, that a moving average approach outperforms most other schemes over extended time periods. Then I worked on refining this to overcome a major limitation that I felt it had.
What was that?
I added additional capabilities to overcome the limitation of lagging behind the market, particularly on exit signals to protect profits in open positions.
How did you do that?
I added “protect long” and “protect short” rules, using two independent sensitivity channels or bands for both the long and short sides of the market. That way, ProfitTaker triggers a signal to get out of a position long before typical moving average approaches would normally give a liquidate long or cover short signal. I also added an execution timing parameter that determines the best time to put on and take off positions for each market. l published an article about that in 1983.
That’s the good thing about moving averages they can’t put the trader on the wrong side of the market. It’s just mathematically impossible. And that’s why I’ve always used them, but l hate to admit it sometimes because so many people are using exotic indicators.
I don’t have a problem with any indicator no matter how mundane some people think it is, as long as it works. I’m not impressed by what happens to be the “in” thing today. I’ve seen too many ideas get shot down. A technical indicator or trading system doesn’t have to be exotic to be profitable. On the contrary, if the rules are easy to understand and make logical sense to the trader, he’s more likely to maintain his discipline and pull the trigger consistently. Too often these exotic indicators are touted just to make a lot of sales.
What about your spare time. It doesn’t sound like you have any. What do you do when you’re not working?
Actually, I have very little free time, especially since I moved last year to a small ranch in the country just outside of Tampa. My wife is really into horses and we have an Arabian and a pleasure driving pony. Then there’s the little calf she gave me for our anniversary last Fall. Between that, our dogs, cats, geese, rabbits, and of course my two very active sons, my spare time is pretty well taken up just tending to the flock. It’s relaxing though. A lot of times I just take walks around the property to enjoy the outdoors. It’s mentally rejuvenating which allows me to be even more productive when I get back to work.
Louis Mendelsohn can be reached at Market Technologies 5807 Old Pasco Road, Wesley Chapel, FL 33544, (813) 973-0496.
Reprinted from Technical Analysis of
Stocks & Commodities magazine. (C) 1988 Technical Analysis, Inc.,
4757 California Avenue S.W., Seattle, WA 98116-4499, (800) 832-4642.