Trading Model Explanation
All WMA models (DGR, S&P 100, & Nasdaq Trading Models as well as the U.S. and European Allocation Models) combine the algorithms in the WMA Trend Model and the WMA Sector Allocation Model. The allocation models place greater emphasis on individual company fundamental rankings. The quantitative model inputs include a relative price vector (the “WMA price”) for each stock, an absolute price vector (the stock’s dividend-adjusted market price), and the daily highs/lows for each stock’s market price. The fundamental rating criteria are presented below. Our U.S. and European allocation model universes contain all stocks investable on Interactive Brokers platform (excluding illiquid stocks with wide bid/asks).
All of our quantitative models place equal importance on the relative and market prices for each stock. In our methodology, if a stock is not outperforming the benchmark on at least an intermediate-term horizon, its score will not be high enough to generate a “BUY” signal.
The model prints show the evolution of signals for the past several weeks. We use a medium/long-term horizon for the allocation models (“long” from a TA point-of-view), limiting our data observations to 40-weeks (daily price data is used in the trading models). The models give “BUY”, “SELL”, and “HOLD” signals. The yellow “HOLD” signals indicate that overall price performance of the stock is mediocre. We do not trade on “HOLD” signals. We recommend referring to the prior signal when a “HOLD” appears. For example, if yesterday a stock was on “BUY” and today we see a “HOLD”, then the “BUY” is not yet cancelled. Only when a “SELL” appears will a long position be closed based on price deterioration. The same logic applies in the other direction.
When a stock’s signals shows all three colors over the past 6 trading weeks (“BUY”, “SELL”, and “HOLD”) this is an indication that the stock price is consolidating in a horizontal trading range or is in the process of a potential reversal.
To better filter our market price-based model signals, we include important fundamental data in our allocation models. We consider fundamental data separately from the quantitative model (the “BUY”, “SELL”, and “HOLD” signals do not incorporate our fundamental scores). In the column preceding the weekly model signals in the allocation models, we post the sector to which the company belongs and the cluster (Growth, Value, Yield) to which the company currently matches. For simplicity, we only show the companies at the top of each cluster — the candidates for inclusion in our Top Picks portfolios. We do not sort on market capitalization, as this is very secondary to the fundamental strength of the company. Fundamental scores (not shown on the Allocation Model prints) range from zero (poor fundamental condition of the company) to 100 (excellent fundamental condition of the company).
In our Techno/Fundamental Allocation Models (see the U.S. Total Market Techno/Fundamental Model and the Europe Total Market Techno/Fundamental Model ), we combine our price momentum scores using weekly data along with fundamental scores derived from Thomson Reuters forecast data for each company covered in the model. We separate companies into the aforementioned fundamental “clusters”, according to our favourite fundamental measures: Valuation (company valuation and stock PER), Growth (rate of increase in sales, positive earnings revisions), and Yield. All our fundamental criteria are described below. The model will “flag” the top 10 to 20 (approximately) stocks in each cluster, provided that the stock’s price action does not slip behind the universe’s benchmark index. Investors can create/maintain a thematic portfolio (eg: a high yield portfolio) based on the model’s weekly recommendations or form a multi-strat portfolio, selecting companies from all fundamental clusters.
Composite Short-Term Fundamental Rating
Based on a weighted average of the following criteria:
- EPS Revisions (short-term),
- Business Predictability.
This rating captures undervalued companies, with good business predictability, and whose EPS estimates have been revised upwards in recent weeks.
Composite Medium/Long-Term Fundamental Rating
Based on a weighted average of the following criteria:
- EPS Revisions (long-term),
- Financial Situation,
- Earnings Quality,
- Business Predictability.
This rating captures undervalued companies, whose business is growing, which have a good financial situation, strong profitability, good business predictability and have not disappointed analysts in the recent past.
In accordance with a techno-fundamental approach, we have stronger conviction in “BUY” signals when the company fundamental score is high, and, conversely, are more inclined to sell/avoid a stock flashing a “SELL” signal accompanied by a weak fundamental score. A summary of our fundamental criteria is presented below.
Growth (Revenue). The growth rating is based on the evolution of the turnover of the company between the last year and the three coming years according to consensus estimates. The higher the growth is (from a relative viewpoint), the better the rating. The goal is to rank companies according to estimated sales and to identify companies with the highest growth.
Valuation. The valuation rating is based on the ratio between enterprise value and its turnover for the current fiscal year. The lower the valuation is, the better the rating is. The goal is to rank companies according to valuation and to identify companies with the lowest valuation.
EPS revisions. The one-week, four-month, and one-year EPS revisions ratings are based on the evolution of EPS (earnings per share) revisions of the company over the current and subsequent fiscal years. Over the previous week/four-months/one-year, the more the EPS estimates are revised upward (from a relative point of view), the higher the rating. The goal is to rank companies according to analyst estimates and to identify companies with the best EPS estimates.
Revenue revisions. The four-month and one-year revenue revisions ratings are based on the evolution of revenue revisions of the company for the current and subsequent fiscal years. The more revenue estimates are revised upward (from a relative point of view), the higher the rating. The goal is to rank companies according to analyst estimates and to identify companies with the best revenue estimates.
Financial Situation. The finances rating is based on the evolution of the net debt of the company (debt or cash) and its EBITDA, compared to revenue. The higher the cash level, the better the rating. The goal is to rank companies according to their financial situation and to identify companies with the highest quality balance sheets.
Profitability. The profitability rating is based on net margin of the company for the current and subsequent fiscal years according to consensus estimates. The higher the ratio, the better the rating. The goal is to rank companies according to the “net income/revenue” ratio in order to identify those which have a high payoff.
Price Earnings Ratio (PER). The price earnings ratio rating compares the company’s current share price to its per-share earnings for the current and subsequent fiscal years. The lower the PER, the better the rating. The goal is to rank companies according to their earnings multiples and identify those which are cheap.
Earnings quality. The earnings quality rating is based on the quality of past earnings released by the company compared to analysts’ estimates. The better the earnings release, the higher the rating. The companies closest to the consensus will have an average score. The goal is to identify companies that publish regularly above consensus.
Business Predictability. The business predictability rating is based on the dispersion of analysts’ estimates concerning the evolution of the company business in the coming years (range estimates). The more the estimates are concentrated, the higher the rating. The goal is to rank companies according to the predictability of their business and identify companies whose business is highly predictable.
Potential. The potential rating is based on the average target price fixed by the consensus from Thomson Reuters. The higher the target price, the better the rating. The goal is to identify companies who have, according to analysts, the strongest upside potential.
Consensus. The consensus rating is based on analyst recommendations. It provides an indication of the position taken by most analysts polled by Thomson Reuters. The goal is to identify companies that benefit from the maximum of buy (or sell) recommendations.
Yield. The “yield” rating is based on the dividend relative to its share price. The higher the dividend yield is, the better the rating is. The goal is to identify companies that can supply a significant dividend return to their shareholders.
Dynamic Trading Models
The Dynamic S&P 100 and Dynamic Nasdaq 100 take a shorter-term outlook, using daily price data and targeting a 1-3 weeks horizon (short-term) and a 1-3 month horizon (medium term). For each stock, the upper line shows the model’s signal for the short-term horizon, while the bottom line provides the model’s recommendation for the medium-term horizon. The past 10-days signals are shown. The company’s fundamental score is then shown for both a short-term and long-term horizon. For example, the short-term score includes EPS revisions for next week and upcoming four months while the long-term score contains 1-year EPS revisions.
We recommend trading stocks with fundamental scores above 50 on the long side and looking to short stocks with fundamental scores well below 50. Based on the investor’s trading frequency, the short-term model signals (upper lines) can be used as simple alerts, used to reduce (SELL) or add (BUY) to positions, or ignored for those who position trade (focusing on the intermediate term signals in the bottom lines).
The last four columns in the dynamic trading models add additional features that we find useful in trading:
- The RSI column gives alerts if a stock becomes overbought (OB) or oversold (OS) based on the 9-day and 14-day RSIs. The “Rel OS” and “Rel OB” signals that the stock has become oversold or overbought on a relative basis. Remember that overbought and oversold levels can persist for several days or weeks.
- “Break-outs” alerts if the last closing price of the stock closed above it recent high over the past 5-days (one arrow up, ↑) or over the past 10-days (two arrows up, ↑↑). Same logic for price break-downs. While the ST model signal integrates break-ups or break-downs, we find it very useful to visualize which stocks are taking-off (or falling out of bed). In addition, a “CCI ↑” will appear if a stock sells-off then generates a rebound signal based on our Commodity Channel Indicator levels. These signals don’t occur frequently, but are generally profitable when we get one.
- “21-day Resistance/Support” shows where a stock is positioned within its trading range over the past 21-days. The top percent is the distance to the 21-day session high while the bottom percent is the distance to the 21-day session low.
- “52-Week” will flash a “HIGH” or “LOW” when a stock is within 1% of its 52-week closing high or low.