Real World Screener Performance

October 2, 2022 Printer Friendly Printer Friendly

2022 Market Performance Thru Q3

2022 has not been kind to investors, unless you are overall short the market. Since the beginning of the year, the S&P 500 has fallen by 24.77%, the Dow 30 has done slightly better, falling 20.95% and the Nasdaq has done significantly worse, falling 32.40%, losing about a third of its value since the start of the year.

If we go back a year and measure the numbers from the beginning of 2021, the numbers still are weak, but don’t look quite as panic-inducing. The S&P 500 has fallen 4.54% since the beginning of 2021. Still a negative return, but certainly of far less magnitude. The Dow 30 has done slightly worse, falling 6.15% since the beginning of 2021. The Nasdaq has done far worse still, falling 17.94% over the last 21 months.

So looking at longer time periods does help mitigate the poor returns of 2022 somewhat. But with that said, there is no getting over the fact that 2022 has been a disaster so far.

Screeners Performance in a Down Market

So these are the rough waters in which we now swim, volatile and bearish. With that as a backdrop, I thought it would be useful to know which screening strategies are holding up best vs. which strategies have performed poorly in this current negative market environment.

On July 15th 2022, I ran 13 of the most popular Stock Rover screeners, each pursuing a different investment strategy. After running each screener I put the passing stocks of each screener, into a corresponding equal weighted portfolio of the same name.

Note that from the screening date of July 15th through September 30th, the market has performed poorly. The S&P 500 has dropped 7.18%, the Dow 30 has dropped 8.19% and the Nasdaq has dropped 7.66%. We can use these numbers to benchmark the performance of our screeners.

The Screeners

The methodology I used was to run each of the 13 separate screeners on July 15th, 2022. Each screener generated its own unique list of passing tickers. The passing tickers were in turn used to generate 13 separate portfolios, each based on a specific screener.

Each portfolio had a differing number of stocks, based on the number of stocks that passed the screener. However, all portfolios were constructed so each passing stock was equally weighted within the portfolio, with each stock getting an allocation of $10,000 of virtual capital within the portfolio.

The 13 screeners that were tested were as follows:

  • Buffettology Inspired
  • CAN SLIM
  • Capital Efficiency
  • Dividend Growth
  • Fair Value
  • Growth at a Reasonable Price
  • Large Cap Value
  • Piotroski High F-Score
  • Safe Performers
  • S&P 500 Outperformance
  • Stock, Industry and Sector Momentum
  • Strong Buys
  • Top Stocks

All of these screeners are provided by default when you first create your Stock Rover account. If you have been a Stock Rover member for a long time, the screeners in your account may not match this list, but all of the above screeners are available from the Library and any of them can be imported into your account.

By using the Stock Rover Screener Manager, The detailed descriptions for the screeners can be found in the description within the screener itself. And the screener criteria can be viewed from within the Screener Manager when the screener is selected.

An example with the Strong Buys screener selected is shown below.

Strong Buys Screener

The Portfolios

Each of the 13 portfolios created from the screeners is named in a consistent fashion, as follows:

“BT – screener name – 7-15-22” where “screener name” is the name of the associated screener. BT is an abbreviation for Back Test.

For example the portfolio associated with the “Fair Value” screener is named “BT – Fair Value – 7-15-22”.

All of the portfolios created for the backtest exercise are available for import from the Stock Rover Library.

So that is the setup. What were the results?

Let’s begin with the number of stocks that passed each of the 13 screeners when they were run, because there were widely differing amounts of passing stocks across all the screeners. The 13 portfolios and the number of positions in each portfolio are listed in the table below.

Portfolio Number of Positions
BT – Buffettology – 7-15-22 50
BT – CAN SLIM – 7-15-22 3
BT – Capital Efficiency – 7-15-22 68
BT – Dividend Growth – 7-15-22 127
BT – Fair Value – 7-15-22 50
BT – Growth at a Reasonable Price – 7-15-22 90
BT – Large Cap Value – 7-15-22 50
BT – Piotroski High F-Score – 7-15-22 54
BT – Safe Performers – 7-15-22 39
BT – S&P 500 Outperformance – 7-15-22 41
BT – Stock, Industry and Sector Momentum 7-15-22 7
BT – Strong Buys – 7-15-22 50
BT – Top Stocks – 7-15-22 21

You will notice that the number of stocks in a given portfolio vary from 3 (CAN SLIM) to 127 (Dividend Growth). This dispersion can dramatically impact both the performance and the risk reward profile of a screener. Generally the more stocks in a portfolio, the smaller the performance variation and the lower the volatility. Or in other words, portfolios with fewer stocks will inherently have a higher risk/reward profile.

The Results

So how did all of the screeners do?

To answer that question, I made use of the Stock Rover Portfolio Analytics Facility to assess both the return and the amount of risk taken by the portfolio to achieve that return. The results are interesting. There indeed was a wide dispersion in the performance of the various investment strategies.

The results cover the 2 ½ month period from July 15th to September 30th inclusive.

The Winners

Here is a table showing some key performance metrics for each of the portfolios, sorted by return. This table comes from the Value over Time window of the Stock Rover Portfolio Analytics facility. The normal column order has been rearranged to show only the pertinent columns for this blog.

Value Over Time

You can download this table in Excel format here

There are a number of return columns in the display. The differences lie in how inflows and outflows are handled in the period, whether to include dividends and interest in the returns or not, and whether the returns are for the time period or annualized. To simplify things, we will work with the Period Rate of Return Column, which is defined as follows:

The money-weighted or personal return over the selected period including both price appreciation and dividends also called an Internal Rate of Return (IRR). IRR is calculated on a daily basis using every day in the reporting range.

Let’s re-sort the display by this column, best to worst, as shown below.

Value Over Time Sorted

Given the terrible market environment, it is somewhat surprising that 4 out of the 13 screeners were able to post positive performance swimming against such a strong negative market current. Let’s take a look at the winners.

CAN SLIM

The CAN SLIM screener posted the best overall performance with a 4.8% return. If you factor in the negative market of 7.2%, using the S&P 500 as the benchmark, that is an overall Alpha of 12%. So how did the screener do it? Well it did it by having only three stocks, and one of them Clearfield (CLFD) appreciated over 38% in the period. Clearfield is a small company that manufactures and distributes passive electronic connectivity products. One of the other companies that passed the screener was Encore Capital (ECPG), which dropped 30% in the period.

The return of each of the three holdings can be found in the Holdings Detail Tab of the Stock Rover Portfolio Analytics facility as shown below.

CAN SLIM Holdings Detail

So while CAN SLIM won the return derby, I doubt it will win the risk reward derby with only three highly volatile stocks.

Strong Buys

The Strong Buys screener finished in the #2 position just behind CAN SLIM at a 4.6% return vs. 4.8 for CAN SLIM. However Strong Buys has a major advantage of containing 50 stocks.

The Strong Buys Screener is a relatively simple screener, relying on only four criteria. The screener is a ranked screener, such that the top 50 stocks that pass the screener are the ones that are used. The screener description is as follows:

Finds stocks with a high margin of safety that are also in favor with the market as shown by a sentiment score in the top quartile and a recent buy sign from the MACD technical indicator.

Below is a screenshot of the actual criteria of the screener.

Strong Buys Screener Criteria

The performance of the individual passing stocks is shown in the screenshot below. The screenshot comes from the Holdings Detail tab of the Portfolio Analytics Facility.

Strong Buys Holdings Detail

Stock, Industry and Sector Momentum

In third place with a 2.3% return is the Stock, Industry and Sector Momentum screener. The screener description is as follows:

Find US stocks that are outperforming the S&P 500 over the last month and year, operating in industries and sectors that are also outperforming the S&P 500 over the last month and year. An additional criteria is the industry must be outperforming the sector in the 1 month and 1 year periods. All passing companies must have a market cap of at least 50 million dollars.

So this is a pure momentum screener and it did well despite the negative momentum of the market as a whole. However like the CAN SLIM screener, this screener will suffer in the risk rewards measurement to follow because only 7 stocks passed this screener when it was run on July 15th.

Piotroski High F-Score

In fourth and final screener with a positive return is the venerable Piotroski High F-Score screener. This screener generated a 0.2% positive return against a market that dropped 7.2%. Not bad at all. The screener description is as follows:

This screener uses 9 criteria that look for companies that have solid financials that are getting better. The original 9-point system was developed by Joseph Piotroski, a professor of accounting. Passing companies must have a score of 9.

Similar to the Strong Buys Screener, the Piotroski High F-Score is a more diversified portfolio consisting of 54 overall positions.

The Losers

We have spent time looking at the four winning screeners. Let’s take a quick minute to look at a few of the worst performing screeners, so we can get a sense of the screeners we may want to avoid in this market environment. The good news, is that out of the 13 Stock Rover screeners, only two performed worse than the market in the period.

Top Stocks

By the numbers, the worst screener by far, generating a 16.2% loss with 21 stocks was the ironically named Top Stocks screener. This is also a relatively simple screener focusing on low EV / EBDITA valuation, High Return on Equity and Return on Assets and price momentum as defined by being within 15% of the stock’s 52 week high.

However the story isn’t quite as bad as it seems for this screener because two of the companies that passed the screener stopped pricing on the exchanges, so they were calculated with 100% losses. Factoring that out, the screener still lost 7%, which is around market average. So Top Stocks wasn’t terrible, but perhaps Average Stocks maybe a more accurate name for the screener.

Buffettology Inspired

Warren Buffett isn’t going to like this, but the worst screener was the Buffettology Inspired screener with a loss of 7.8%, slightly worse than the market at 7.2%. This screener contained 50 stocks. The description of the screener is as follows:

This screener is based on criteria described in the bestselling Buffettology book. The company should have a 10-year track record of generally increasing EPS with no negative earnings years; long-term debt not more than 5 times annual earnings; average ROE over the past ten years at least 15%, average ROIC over the last 10 years at least 12%, and earnings yield should be higher than the long term Treasury yield.

A bit surprising perhaps, but remember the screener is only looking at 2 1/2 months of returns. Warren Buffett is a long-term kind of guy who thinks in decades.

Large Cap Value

The final screener that slightly underperformed the market was the Large Cap Value Screener, which contains 50 positions. The screener returned a loss of 7.4%, only slightly worse than the market at 7.2%. The description of this screener is as follows:

Find large companies (greater than 5 billion in market cap) that are inexpensive by traditional measures such as low price to earnings, price to sales and price to book. These companies should still be growing sales and earnings.

Risk and Reward Analysis

Now its time to take a look at the risk side of the equation and see what looks good once we factor in the risk required to achieve the reward. One thought to keep in mind is that portfolios with fewer stocks will inherently tend to have more volatility and risk.

Here is a table showing some key risk metrics for each of the portfolios, sorted by risk adjusted return. This table comes from the Risk and Reward tab of the Stock Rover Portfolio Analytics facility, and it shows a number of metrics concerning the risk taken to achieve the return.

Risk and Reward

You can download this table in Excel format here

There are a number of risk metrics here, but if we focus on Risk Adjusted Return vs. the S&P 500, we can see that Strong Buys rises to the top and Top Stocks falls to the bottom. Again Top Stocks is not quite that bad because of two stocks that failed to price and are shown as 100% losses. But even factoring that out, it would be the worst risk adjusted screener during this time period.

Risk and Reward Sorted

In a small consolation for Warren Buffett, the Buffettology Inspired screener passed the Large Cap Value Screener when risk adjusted metrics are included.

Conclusion

This has been a very interesting investment exercise. I have learned a lot while writing this blog and I hope you feel the same while reading it.

Different investment strategies tend to work (or not work) during different times. There is no one strategy that is always the best (or the worst). Given the current times, which feature rapidly rising interest rates, high inflation, fear of recession and global disruption from war and the pandemic, these are certainly not ordinary times. And in the current times, it seems that paying attention to Margin of Safety (a Stock Rover proprietary metric), sentiment and the MACD indicator pay off. Whereas the more traditional valuation metrics such as Price to Earnings, Price to Book, Price to Sales as well as Return on Assets, Equity and Invested Capital seem to matter less.

Why? Only Mr. Market knows and he isn’t talking.

I’ll close by stating that like all things, both the macro environment and the strategies that work in a given environment can shift on a moment’s notice. The blog post highlights what is working and what is not working today. If the macro environment remain the same, the likelihood is that the strategies that are working today will also be working tomorrow and the day after, but it is far from guaranteed. The market always seems to confound all attempts to understand it, and especially to predict it.




Comments

John Carman says:

Being somewhat newer to Stock Rover I’m wondering if this type of comparative analysis of the different screeners has been done over longer time frames eg 1, 3, 5, 10 & 20 years. If not I’m curious why? It seems like the results would be the single most valuable use of Stock Rover.

Thanks for the great article.

Mel Turetzky says:

I agree with John with some additional caveats. It would be well to compare screeners during at least two periods and certainly over longer periods. I feel a screener should be examined during a one year period of market growth and another during market downturn, the larger the growth and downturn the better. Currently the year 2021 would provide an example of the former and the current year (even the 10 months current) would be good for the latter.
But this is still a very good beginning for a facility I have advocated for a while. If there were more history available for some of the equity properties then we could do a bit of this on our own. It is an excellent tutorial, very thoroughly analyzed, and one of the reasons I have been a Stock Rover cheer leader since day one.
Melvin Turetzky

Howard Reisman says:

This is the first of what we expect will be a continuous project to measure the long term performance of a variety of screeners. As to why we didn’t do it earlier, there were other things on our agenda that were scheduled ahead of it. I guess the saying “better late than never” applies here.

John says:

Thanks for the reply. Indeed better late than never– so thank you for all the other great things too!

Howard Reisman says:

Thanks you Mel and John. I appreciate your kind comments.

Ihuoma says:

This was interesting to read. It will be interesting to see if the strong buy signal improves the result if we consider miving average crossovers as well. Can’t wait to see what you publish in the next edition. Kudos 👏

Jim says:

Great article and interesting results. The more I use the tool, read the articles and increase understanding of the information, the more I realize how limited my views were before.

Now, if only you could accurately predict the future….! Along these lines, one extension could be adding to the news sites we can access through the Insight tool.

Katharine says:

Hi – this was quite interesting and I look forward to future articles along these lines. When choosing time periods for comparison, is Stock Rover able to do “historic screening” where it would run a screen and get the results it would have gotten on, for example, 1/1/2016 or 7/1/2019? Or will you need to use the current date as a starting point, and then let time elapse to see results over various time periods?

Katharine

Howard Reisman says:

Stock Rover has a limited backtesting capability in that you can look at past data on a quarterly basis or yearly basis and use it as part of the screening criteria.

It does not have a generalized ability to run a screener as of an arbitrary date in the past. Hence this test, which runs the screeners on a specific date and watches the results as time passes.

JJia says:

I agree with John. It would have been much more insightful to run the screeners as of 1/1/2022 and show YTD performance comparison. In another word, a back test comparison. I notice the winners mostly selected small/micro cap companies. It would also be interesting to apply a cap based criteria across all screeners to make the comparison more apple-to-apple.

Howard Reisman says:

Both changes are coming in the next round. The screeners will be re-run as on 1/1/23 and they will be changed to put lower limits on the market cap so the passing companies aren’t too small.

Bill Bowen says:

First I would like to thank Howard and associates for providing the best investing analysis tool I have ever encountered.

Second, if future back testing of screeners is contemplated, I recommend Richard Tortoriello’s book “Quantitative Strategies for Achieving Alpha” on how to do back testing rigorously .

Third I must ask: do these back tests include any criteria for rendering them practical investment strategies? e.g. turnover, trading volume, tax consequences, bid-ask spreads, etc. In other words, could the typical Stock Rover client actually invest in these strategies and achieve these results?

Howard Reisman says:

Thank for you for your kind words.

Certainly the results could be achieved for a practical investment strategy. The only fly in the ointment would be the really small stocks that pass may have too small trading volume and too large bid / ask spreads to make buying at an efficient price difficult.

Next round I will address these issues by putting minimums on cap size and trading volume to ensure the passing companies can be efficiently purchased.

Stan says:

First of all, thank you Howard and your team for this insightful analysis.

When I first joined Stock Rover in 2021 I did precisely that prior to using any for my real world investing; I chose 20 screeners, customized each one to also include my secondary criteria of passing a short-term financial liquidity test, and having a market cap greater that $1 billion. Over the 6-month period of the test, the “Stock, Industry and Sector Momentum” came out on top by a wide margin. The result was not surprising given the strong upward trajectory of the market during that time frame. But CAN SLIM, on the other hand, ended at the bottom-half.

So, I’m very surprised that CAN SLIM screener came out on top here, whereas the “value” type screeners were at the bottom. This is in stark contrast to what all the experts in financial media have been suggesting to do during this bear market, basically for investors to focus on companies having more traditional investment metrics similar to your losing screeners: Fair Value, Dividend Growth, Large Cap Value, Buffettology, GARP, etc., etc. However, this empirical data suggests otherwise. Perhaps, as suggested by others, that’s due to the short time frame of the test?

Second, why do you conclude the following?
“…it seems that paying attention to Margin of Safety (a Stock Rover proprietary metric), sentiment and the MACD indicator pay off…” None of these criteria are a part of CAN SL or even Momentum.

Third, I really appreciate and enjoy using Stock Rover’s screeners because they are so flexible to customize to fit one’s own investment criteria.
Again, thank you.

Howard Reisman says:

Thank you for your detailed comments. CAN SLIM only selected three stocks and one of them had a huge appreciation in the period. I consider that an anomaly.

As far as the “financial experts”, it has been my experience that empirical data often diverges from expert opinion. As far as predicting the future, financial experts have no particular advantage over anyone else, or darts for that matter. Because no one can accurately predict the future or what will definitely work.

The best risk adjusted screener in the period was Strong Buys and that was the criteria it used. The indicators paid off over the 2 1/2 month period of the test. Will they pay off over a longer time period? We shall find out.

PM says:

Good stuff. Thanks.

Very few people would buy every stock in a 50 stock screener result – so one’s results would of course vary greatly based on what one picked to buy from the screener results.

So, it would be useful to look back at the list of stocks from initial screener output and compare with future results to see if any specific sub-features (fundamental, sentiment) could have helped pick more of the ultimate winners.

Howard Reisman says:

Agree. Of the course the idea with a screener is to get a small set of interesting candidates and then from there, do additional research to whittle the list down to an even smaller group that you might actually want to buy.

The benefit of the overall performance of the screener is to determine whether the initial set of candidates provided by the screener is on the right track for market beating performance or not.

Melvin Turetzky says:

This exercise establishes how valuable the lookback process is. So wouldn’t it be wonderful if Stock Rover provided a function that would permit us, today, to go back and determine the results of a screener six months, a year, two years ago? Or at least some limited part of that history?

John Follansbee says:

I enjoyed the exercise. I agree with your conclusion that strategy success varies with the time periods involved. One strategy that I have successfully employed in this bear market is selling calls with particularly high implied volatility ratios against my stock portfolio. Calls are generally just beyond the expected move with a Delta of ~22 about 45-50 days prior to expiration, rolling ~21 days prior to expiration and placing GTC’s at 50% of expected profit. One can adjust these parameters depending on the tax status of the portfolio and whether or not one cares whether or not the stock is called. Selling calls on speculative holdings far above the current price enforces selling discipline. If a stock pops too far, we can always consider rolling the call for a credit if the stock meets our Stock Rover analysis criteria. This is often a rinse and repeat strategy.

Howard Reisman says:

Agree this is a good strategy if you don’t actually mind getting the stock called and the ensuing tax consequences will not result in a big tax obligation.

Dave R says:

Interesting as a review of what each screener provides; however, a 2.5 month back test is completely irrelevant for serious investors. It should be simple to provide 5 and 10 year back tests in a blog post using each possible subset of these period lengths over a given study period.

Comments are closed.




Top