Support Desk |
ETFOptimize Support > Help Desk > Knowledgebase

Search help:

Why Don't You Have Any Short-term Signal Systems?




Q:  "I still would like to see a very short-term more active algorithm from you guys. Why don't you have any short-term signal systems?"

A: We are working on a new technical-only signal system that may provide something closer to the active system for which you are looking, but we may never do a pure daily or intraday system. Here's why:


We have done quantitative testing of investment systems since the late-1980s, and have spent a considerable amount of time testing short-term, intermediate-term, and long-term trading systems. A consistent problem with short-term (daily) signals is that they are too inaccurate, have too many whipsaws, and lead to too many erroneous trades that destroy returns. Simply put, there is too much random noise in the short-term data that skews signals and causes whipsaws.

Also, our models work diligently to identify robust trends, because trend-trading is the most profitable investment approach. Profits are the basis of Trends, and Trends are a function of Time.

A trend, that is — represented by a chart moving from the lower left to the upper right — is created by profits (consistently increasing share prices) — i.e., profits—which require time to accrue. By definition, Trend is a function of Time.

Symbolically, the Trend Investing equation is…


TR = (n)T x ∆Price⬆︎

where TR is Trend, T is Time, and ∆Price⬆︎ is the directional delta (change) of
prices. There can also be 
profitable  downward trends, represented by  

TR = (n)T x ∆Price⬇︎

Put another way, Trend is equivalent to 'n' amount of Time multiplied by the
upward(downward) change in Prices. Obviously, a larger amount (n) of Time
multiplied by increasing Prices will result in greater profits, so we want the
amount (n) of Time to be as large as possible.





Because of the random noise in day-to-day prices, daily signals are often far too inaccurate for use in quantitative systems that seek to identify trends. The longer the timeframe considered, the more accurate the indicators.

However, monthly or quarterly signals are too slow to provide optimum entry and exit points. Therefore, we usually use longer-term quantitative composite measures, constructed from multiple uncorrelated indicators, with weekly assessments. In other words, we may use indicators based on 200, 500, or even 1,000-day moving averages, but we assess those long-term signals every weekend. The average hold time for our models is about three months—or four trades per year, but each model assesses signals every week. 

For intermediate-term quantitative signals, we find that each Friday's closing price usually provides the most reliable and accurate signals for any given equity or ETF, without incurring too much lag. 

Daily closing prices have far too much noise, and even worse, investors often make discretionary decisions based on these noisy, mid-week prices and invariably get badly burned. Discretionary decisions exacerbate the problem of noisy prices by introducing speculation and human emotional biases.

The failure of investors to 'beat' this combination of irrational forces is the reason many new investors abandon investing altogether as an impossible undertaking. Even for those that stick with self-managed investing or turn their money over to a mutual fund manager, according to Dalbar, Inc., the average long-term annualized return attained by individual investors is only 2.6% per year (only 0.2% above the long-term rate of inflation).

All that said, if we ever find a short-term trading system that is consistently profitable and has a legitimate edge, with 70-90% winning trades, we will definitely offer it to subscribers. However, in the last 40 years, we have yet to find one that is comparable in profits to a weekly assessment system.




Was this article helpful? yes / no
Related articles How Often are Leveraged Positions Used?
Cancellation of Strategy Subscription
Should I Add a Stop Loss to My ETF Positions for Additional Protection?
Article details
Article ID: 14
Category: Knowledgebase
Rating (Votes): Article rated 3.2/5.0 (16)

« Go back