Time the Markets: Using Technical Analysis to Interpret Economic Data, Revised Edition (Gal Zentner's Library) by Charles D. Kirkpatrick

Time the Markets: Using Technical Analysis to Interpret Economic Data, Revised Edition (Gal Zentner's Library) by Charles D. Kirkpatrick

Author:Charles D. Kirkpatrick
Language: eng
Format: epub
Publisher: Pearson Education
Published: 2012-06-14T16:00:00+00:00


Earnings Yield

Most academic studies of the factors affecting stock price returns show that earnings yield has the highest correlation. Of course, most of these studies, unlike this one, fail to measure an out-of-sample test to see whether the correlation holds up in a set of data unknown to the system, or if it is just over-optimizing the existing, in-sample data. As you have seen in the preceding chapter, a relationship between one set of data must be tested against another out-of-sample data before we can judge that it might have predictive value. Regardless of how successful a system is during the in-sample data test, it must be disregarded if it cannot show profitable, minimal-risk results in unknown, out-of-sample data.

As with most tests in this book, this optimization creates a moving average crossover system with a filter. Such a system has a set of moving averages and a percentage filter by which the first moving average must cross the second moving average before a signal is given. There are thus three variables: first moving average, second moving average, and percentage filter. The final moving average length and percent filter parameters are determined solely through the walk-forward optimization.

To give you insight into the entire process, for earnings yield in this chapter alone, I go through the entire optimization, testing, and evaluation process in detail. In the experiments remaining after this one, I report only the resulting figures and add some commentary on the each indicator’s character.



Download



Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.