I've been working hard to roll out new features at Alert HQ. I always try to do some backtesting to verify that these features are working as anticipated. Actually running the various screens and putting out the results at Alert HQ, though, always yields new insights.For example, I recently introduced Trend Busters. This was an attempt to implement a system that would identify the recent trend and determine whether that trend had been violated. This was initially rolled out using a simple test: if the trend was pointing upward and the most recent price was 5% below the trend then it was a SELL signal. If the trend was heading downward and the most recent price was 5% above the trend then it was a BUY signal. This seemed to be pretty workable but in practice it turns out that sideways action will trigger a false signal.
This post is to explain some of the changes that have been implemented to make the Trend Busters signals more accurate. Here are some of the improvements we have made to BUY signals:
- Verify that the closing price of the most recent day is higher than the closing price of the previous day
- Verify that the low price of the most recent day is higher than the low price of the previous day
- Verify that the closing price of the most recent day is at least 5% higher the closing price of the day when the most recent minimum price was established
- Verify the most recent day's low price is higher than the high of the day when the most recent minimum price was established
Our recent backtesting has validated that this new system rejects the most questionable signals. I am excited to roll out the new Trend Busters. The list will certainly be shorter than it has been in the past but the stocks and ETFs that do make the list will be much better candidates for the Trend Busters title. Be sure to check them out when the newest list is released this Saturday (tomorrow!).
For details on how the original process worked click (here) to read the introductory post.
Comments
Post a Comment