The Holy Grail Returns (12/14/04)

With the return of interest in stock market investing has come the inevitable... The black-box-holy-grail-of-market timing has magically returned. All you have to do is pay $300 a year, follow the buy/sells blindly and with a little leverage you will make 50% annually, so goes the pitch. And IT WORKS! There are now websites pulling in millions of dollars with 5,000 subscribers, each of which will almost certainly eventually become members of the "market timing doesn't work" club.

Considering the above, I really feel that I should comment based on a couple decades of experience with market timing. I must admit that much of what follows I learned by experience. However, I have also learned that one can succeed greatly with market timing if they are realistic about their expectations.

The first thing I want to say is that not all black-box timing websites are dishonest. I think it is extremely misleading to promote and recommend a high-risk system aiming for 50% annual returns without also recommending that only risk capital should be used and that the chance of failure is high.

I think there are systems that the operators think are so good that they want to keep secret although I've seen many of these systems and I've never been too impressed. In most cases, the reason for secrecy is to allow the operators to change parameters for marketing purposes.

I know that some operators of these services know full well that the chance of success for their clients is very low. Lastly,. I'm not accusing any one person or company of anything illegal. These are just my opinions.

The Realities of Market Timing

It is EXTREMELY easy to develop a strategy that shows hypothetical
100% annual returns trading the NDX long and short with leverage.

With the systems in ULTRA, I was just able to do it in less than five minutes. Here are the details for those interested:

@SYS
cot1.sig, 1 *ND
vol.sig, 1 *ND
lhd2s.sig,1 *ND
yroc.sig, 1 *ND
snbm.sig, 1 *ND
pent.sig, 1 *ND
seas2.sig,1
pres.sig, 1
@ALL
0,-100
1,0
2,0
3,100
@END

(All systems use default settings)

And the historical results trading the NDX (1986 to present):

Trades: 4 per year (71% winners)
Annual Return (long only): 36% per year
Annual Return Long/Short w/ 200% Leverage: 107% per year

Now all we need is a crack marketing staff and we too can have 5,000 subscribers and make a million dollars a year selling black-box buy/sell signals.

Seriously, the probabilty of attaining such an ambitous goal is extremely low. The best investors in the world struggle to return 30% annually using the skills they've obtained over decades of experience. Does it seem likely that a strategy developed in five minutes is likely outperform the best investors in the world?

If a black-box system fails, the operator is free to make whatever rule
change is necessary to be able keep his marketing machine alive.

These rule changes are often described as, "adapting to a changing market" or something similar.

In the early 90's I developed a black-box system called U31 that was my best attempt at reproducing the way an "expert" makes timing decisions. I made one rule change in 1994 and then decided I would never do so again because:

1. I wanted to create a study in real-time performance.
2. I wanted to be honest with my clients about the reality of market timing.
3. I don't believe that curve-fitting formula tweaks improve the probabilty of success.

In real-time from 01/02/95 to 07/25/00 (5.5 years), U31 recorded the following results trading the NDX long-only with no leverage:

Trades: 17 (3.1 per year)
Winners: 77%
Annual Return: +33.5%

This was as close to "holy-grail" performance as a system can do. It made me, and a lot of others, a whole lot of money. However, this performance was the exception, not the rule. Assuming that any mechanical strategy is likely to perform this well in real-time can be a big mistake.

Every system has a possible market scenario that will kill it.

U31 was largely breadth-based. In the bear market of 2000, breadth was very positive because of the large number of bond related funds and ETFs traded on the NYSE. Before 2000, positive breadth was a very accurate indicator of accumulation. During the bear market, positive breadth became a worthless indicator of money moving out of stocks and into bonds. This caused U31 to not be able to generate sell signals. The buy signals were good but since U31 couldn't sell, the results ended up terrible.

Luckily, because of the risk in the overvalued market and the risk inherent in U31 (invested 60% of the time) we abandoned U31 in early 2000 for lower risk strategies. We did our best to inform clients that ULTRA did not have a stop-loss of any kind, could be locked into a buy, and that no strategy should be trusted solely. Investors must know that failure is possible for any strategy and that diversity is key. Diversity is exactly why we created ULTRA in its current form with almost 100 individual systems.

In real-time from 07/25/00 to 12/31/02 (2.5 years), U31 recorded the following results trading the NDX long-only with no leverage:

Trades: 9 (3.7 per year)
Winners: 0%
Annual Return: -41%

Obviously, if I had been running a black-box advisory making me $1 million in annual profit, I would have tweaked U31 at its first hint of failure so that I would still have my "holy grail" to advertise. I could still today, very easily make a couple of minor changes in U31 that would make it again appear to be the "holy grail" of market timing.

The simple (and ridiculous) rule, "If year equals 2000 or 2001 or 2002 then don't buy" would again make U31 a "star". Don't laugh. If I were running a black-box advisory how would anyone know the silliness of my new rule.

Good Strategies can make a come-back.

The logic in U31 is still very good. That is, identify when the market is too low by some measure and buy upon some type of price trigger that indicates a likely change of direction. In fact, in the last two years U31 has made a significant come-back.

In real-time from 01/02/03 to 12/06/04 (2 years), U31 recorded the following results trading the NDX long-only with no leverage:

Trades: 7 (3.62 per year)
Winners: 71.4%
Annual Return: +19.2%

So, what is one to do? Should U31 be abandoned completely because of it's 2.5 years of failure? I don't think so. I think that all timing systems should be approached realistically. Because U31 is invested often (60% of the time), it is high risk and should be used in combination with other systems and only as a part of an overall investment strategy.

As your annual return goals increase, your probability of reaching those goals decrease.

There's nothing wrong with following a high risk strategy hoping for huge gains. Just know that the odds of making those huge gains are not high. Systems that seek huge gains are high risk by nature. There is no such thing as a system seeking 30% annual returns that is not high risk. High risk strategies can easily result in catastrophic failure.

You MUST diversify.

Knowing that any strategy can fail, investors must diversify their strategies so that if one fails they will not be decimated. To have the best chance of performing well with reasonable risk you must follow a set of strategies diversifed in:

• Risk- If all your strategies are high risk, you are at risk of total failure.
• Logic- You don't want your strategies to be correlated in logic or they could all fail.
• Assets traded- Standard asset diversity is still a great idea. The Russell 2000 (RUT) was barely effected by the 2000 bear market.

To succeed you must know the details of your strategy.

There is no free lunch in the investment game. Those with the best and most information are the ones that succeed. When your strategy is having a tough time, you MUST know exactly why so that you can make an intelligent decision on whether to ride out the storm or replace that strategy with another. Since you're using a set of strategies, the failure of one is not gong to be significant to your overall financial status.

If you are following a black-box you will simply give up at the first hint of failure and move on to the next holy-grail. The black box operator will reoptimize and be ready to convince a whole new set of "investors" how easy it is to obtain 30% annual returns outperforming Warren Buffet. It's a cycle that I've watched for 20 years and I'm sure will repeat over and over again.

To succeed you must be realistic in your expectations.

Most investors believe market timing is impossible. Some venders want you to believe that 50% annual returns are possible (or even likely). The truth lies somewhere in between.

A market timing money manager I respect greatly believes that one can equal the return of the indexes with lower drawdowns thereby outperforming on a risk adjusted basis. Considering that when you own a portfolio of stocks, that portfolio becomes highly correlated with the indexes and underperforms because of commissions and fees, equaling the indexes with less risk is good performance.

I believe that:

• With timing, you can have an excellent chance outperforming the indexes by five points with much lower drawdowns and risk than the indexes over the long-term with good timing strategies.
• As your annual return goals increase, your chances of obtaining those goals decrease.
• In years when the indexes show large gains, expect to underperform a bit.
• In years where the indexes collapse you should be able to outperform the indexes significantly and that is where you'll get most of your benefit from market timing.
• Lastly, if buy/hold were my only option, I'd avoid the stock market completely.

Obviously you get to form your own beliefs. I hope I've helped that process a bit with this document.

Steve Hunter, ULTRA Financial Systems Inc.

ULTRA Financial Systems LLC
P.O. Box 3938, Breckenridge CO 80424
Phone: 970-453-4956 Fax: 970-453-2467