Portfolio Design

Advanced concepts in stock investment portfolio design.  Fundamentals, technical analysis and many other related topics are discussed.

Use the Screener to Evaluate Stock Ranking Factors

The Portfolio123 screener is typically used to find stocks or ETFs with specific characteristics, or as decribed in the last post, to get a handle on current/future trends. But the screener is also a very convenient tool for evaluating stock factors for use in ranking systems.

Stock market analysis fueled by Portfolio123.

Today, I am going to examine some analyst estimate factors, in the hope of determining the best factor to use for a ranking system.

Screen Setup

Start this exercise by opening up a new screen. The stock universe is the S&P 500 Index. Quick Rank will be used for ranking. Quick Rank bypasses the use of a pre-defined ranking system, allowing the user to experiment with ranking system factors in a fast, convenient way.

The Method is set to Long/Short with maximum number of stocks set to 100 (applies to both the long and short sides). This approach will test a large number of stocks for both the high end and low end of the stock factor, a much more exhaustive test than simply looking at the top ten or twenty stocks.

Portfolio123 screener setup

% Earnings Surprise Latest Quarter (Surprise%Q1)

Now browse the reference under ESTIMATES until you find Surprise%Q1 (Earnings Surprise (Estimated vs. Actual), 1 Quarter Ago).

Screenshot of Quick Rank formula selection: Surprise%Q1

Screen Backtest

Now go to the BACKTEST tab and set up the backtest as follows.  Slippage and Carry Cost should be set to zero. Rebalance Frequency should be set to Weekly for this demonstration.

Note: Weekly rebalance should be used if you are developing a very high turn-over ranking system. 4 Weeks, or Quarterly are often better choices for most lower turn-over systems

Set the time period to the maximum that your account will permit. Click on RUN BACKTEST to process the historical statistics.

Screenshot of Portfolio123 screen backtest setup

Backtest Output

The output for the factor Surprise%Q1 is shown below. As you can see, the results are quite disappointing.

Screenshot of Surprise%Q1 backtest results

Below are test results for several other analyst estimate factors.

% Sales Surprise Last Quarter (SalesSurp%Q1)

Screenshot of Quick Rank formula selection: SalesSurp%Q1

Screenshot of SalesSurp%Q1 backtest results

Long Term EPS % Growth (LTGrthMean)

Screenshot of Quick Rank formula selection: LTGrthMean

Screenshot of LTGrthMean backtest results

Average Recommendation (AvgRec)

This factor is tested in the contrarian state (higher is better).

Screenshot of Quick Rank formula selection: AvgRec

Screenshot of AvgRec backtest results

13 Week Change in Average Recommendation

Screenshot of Quick Rank formula selection: AvgRec13WkAgo-AvgRec

Screenshot of backtest results: AvgRec13WkAgo-AvgRec

Total Sum of EPS Revisions Last Week (TotRevisionsLastW)

Run this last factor just to see that I am not totally wasting your time :-)

Screenshot of Quick Rank formula selection: TotRevisionsLastW

Screenshot of TotRevisionsLastW backtest results

Finally, a an analyst estimate factor with a good, reliable output, giving positive results while holding the top 100 stocks long and the bottom 100 stocks short.