What I've set out to achieve with this indicator and exploration combination is a simple set of tools that can simulate the more sophisticated External Relative Strength Analyzer (ERSA) component of SpyGlass. This is done by basing a scan of selected securities on the ERSA Simulation indicator and sorting the results into ascending or descending order. The sorted results will closely match the ascending or descending percentage results produced by ERSA. With the scan completed any threshold (i.e. number of securities to include/exclude) can be calculated using the following formula.

Securities in Results Report * Threshold Percentage / 100.

I'm not a collector of MetaStock add-ons but I do have SpyGlass. I honestly think the money spent on this add-on is one of the best trading-tool investments I've made. Having said that I should perhaps qualify my comments by admitting that I'm not totally convinced that ERSA is as well thought out or as flexible in operation as it could be. However, the purpose of this article is not to criticize ERSA (nor SpyGlass as for that matter), but rather to demonstrate how the ERSA ranking function can be closely approximated. If you already own SpyGlass (or FIRE as it's now called) you will have absolutely no use for these tools. Feel free, however, to check the accuracy of simulated results against those from the genuine article.

ERSA splits the rate-of-change for each security into four quarters and applies a user-selected weighting to the ROC (or Relative Strength if you prefer that term) of each quarter of the total look-back period. The weighting values allow a user to adjust each data segment's influence on the total Relative Strength rating. A rule enforced by ERSA is that the total weighting of all quadrants must add up to 100%. I've decided not to enforce that rule in my indicator but I do suggest that you abide by it all the same.

Typically a weighting of 40% is applied to the most recent data segment and 20% to each of the three previous segments. On running ERSA across a folder (or portfolio etc.) to calculate the External Relative Strength, all securities are assigned a percentage ranking for each bar available to the ranking process. These values are stored as a text file for each security and among other things they can be used to qualify entries for any trading system (exits too if you've a mind to do that). ERSA rankings are calculated for historical as well as current data so they can also be used for system back-testing.

Unfortunately my simulation indicator cannot be used for back-testing. However, it can be used to accurately assess where any preferred percentage threshold lies for the current bar of the securities under scrutiny. Let me expand a little on that by referring you to the ASX Weekly system from the September 2006 issue of MSTT. That system uses an ERSA threshold of greater than 80% to qualify all other entry conditions. All that means is that the system cannot initiate a trade on any bar where the ERSA ranking is less than or equal to 80 (i.e, less than 81).

ERSA reports percentages as whole numbers, and so when ranking the S&P500, for example, it is highly likely that each percentage value would be returned by five securities. The five highest securities by ERSA ranking would return 99%, the next five 98%, and so on down to 0%. For most purposes that's a fine enough distinction.

How should the ERSA Simulation indicator be used? I suggest that the default periods and weighting values be set as required (40, 20, 20, 20 are good weighting numbers to start with). You'll need to run an exploration and call the current-bar indicator value with Fml("ERSA Simulation"). With the exploration completed the "ERSA Sim" column should be sorted by left-clicking on the report column header. Right-clicking a second time will reverse the order from ascending to descending. Should your exploration scan 500 securities then it's easy to work out that an 80% threshold level will separate the 100 highest ranked securities from the remaining 400 unwanted securities. It's that simple. The simulation value is not a percentage but it doesn't need to be for you to work out which are the top 20 percent by ranking.

Contact the author for a copy of the complete article with all illustrations.


MetaStock® is a registered trademark of Equis International, a Thomson Reuters company.