How our Quant ratings work
Our Quant ratings give quick and powerful insight on stocks by addressing five factors that have historically been highly correlated with stock returns: Growth, Valuation, Quality, Revisions and Momentum.
Five factors, scored against sector peers
For each factor, we compare a variety of quantitative metrics for each company with the rest of its sector, and assign a grade from A (best in sector) to D (worst in sector). The overall Quant Rating aggregates the five factors.
How factor scores are calculated
Within each sector and geography, stocks are evaluated using a broad set of quantitative inputs. Each metric is ranked on a sector-relative basis, so companies are always compared with their direct peers.
For example, in the U.S. Technology sector, the Quality factor incorporates more than 80 quantitative data points. For each metric, such as return on equity, companies are ranked from best to worst within the sector: the highest-ranked company receives a score of 100 for that metric, the lowest-ranked a score of 0. These percentile ranks are then aggregated across all metrics within the factor to produce a sector-relative factor score for each stock.
The same method is applied consistently across all factors, as well as to our proprietary qualitative signals derived from content analytics. Each stock therefore receives standardized, sector-adjusted scores on every factor.
Quant Picks
Our Quant Picks are the stocks with an overall Quant Rating of A+: approximately 30 stocks with the best quantitative ratings in their sector. Each pick receives an equal weight, and sector weights closely match the index, so that performance comes purely from stock selection, not from sector allocation.
Quant Picks are the very best in every sector based on an aggregate score that considers the five core factors (growth, valuation, quality, revisions, momentum) and six complementary factors: capital return, sentiment, EPS surprises, analyst opinion, other quantitative data, and qualitative data from content analytics.
Machine learning
The metrics we consider and their relative importance differ from one sector to another, based on their historical relationship with stock returns in that sector. As these relationships change over time, our scoring algorithm adapts to focus on the most meaningful indicators.
Every month, the algorithm is refined with an expanded dataset. Machine learning lets the model adapt to a changing world where investment styles and criteria evolve. Because the algorithm is always optimized using only the data available at each decision date, the process avoids the typical biases of factor investing and improves the probability of continued alpha generation.
How factor weightings are determined
Factor weightings vary by sector and geography. The algorithm continuously evaluates the historical relationship between each factor and subsequent stock returns within a specific sector and region. Factors that have historically shown stronger explanatory power receive higher weights in the overall Quant Score.
Because these relationships evolve, weightings are dynamic: the model regularly recalibrates factor importance, allowing it to adapt to changing market regimes, style rotations and sector-specific conditions. The same dynamic framework was applied throughout the historical backtest, with factor weights recalculated using only the information available at each point in time.
| Factor | Tech. | Comm. | Discr. | Staples | Energy | Real est. | Utilities | Financials | Industrials | Materials | Health | Average |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quality | 9% | 29% | 13% | 24% | 7% | 53% | 4% | 35% | 0% | 15% | 15% | 19% |
| Valuation | 6% | 2% | 9% | 15% | 15% | 19% | 4% | 18% | 11% | 9% | 14% | 11% |
| Growth | 7% | 5% | 6% | 6% | 0% | 4% | 0% | 5% | 8% | 5% | 11% | 5% |
| Revisions | 5% | 7% | 7% | 1% | 10% | 0% | 10% | 3% | 0% | 19% | 5% | 6% |
| Momentum | 9% | 5% | 9% | 0% | 7% | 4% | 38% | 10% | 11% | 15% | 2% | 10% |
| EPS surprises | 4% | 6% | 4% | 0% | 0% | 1% | 1% | 1% | 6% | 17% | 2% | 4% |
| Analysts opinion | 17% | 9% | 0% | 18% | 7% | 5% | 1% | 4% | 0% | 4% | 4% | 6% |
| Sentiment | 5% | 18% | 14% | 1% | 21% | 0% | 0% | 15% | 14% | 4% | 16% | 10% |
| Capital return | 5% | 2% | 11% | 4% | 23% | 1% | 8% | 6% | 22% | 6% | 27% | 10% |
| Other quantitative data | 15% | 8% | 3% | 3% | 4% | 4% | 3% | 2% | 2% | 1% | 2% | 4% |
| Qualitative data (content analytics) | 18% | 10% | 24% | 29% | 5% | 10% | 30% | 2% | 25% | 4% | 2% | 14% |
| Factor | Tech. | Comm. | Discr. | Staples | Energy | Real est. | Utilities | Financials | Industrials | Materials | Health | Average |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Quality | 11% | 16% | 9% | 7% | 7% | 11% | 5% | 16% | 4% | 0% | 19% | 9% |
| Valuation | 4% | 26% | 0% | 8% | 2% | 24% | 6% | 22% | 7% | 26% | 18% | 13% |
| Growth | 5% | 0% | 4% | 3% | 0% | 0% | 19% | 13% | 5% | 0% | 0% | 4% |
| Revisions | 6% | 4% | 9% | 0% | 7% | 0% | 10% | 0% | 9% | 6% | 8% | 5% |
| Momentum | 32% | 16% | 30% | 0% | 37% | 9% | 11% | 0% | 18% | 7% | 18% | 16% |
| EPS surprises | 4% | 5% | 23% | 0% | 3% | 2% | 4% | 13% | 7% | 24% | 2% | 8% |
| Analysts opinion | 3% | 8% | 4% | 0% | 0% | 0% | 30% | 3% | 0% | 0% | 7% | 5% |
| Sentiment | 17% | 0% | 8% | 0% | 27% | 31% | 6% | 5% | 15% | 6% | 7% | 11% |
| Capital return | 1% | 5% | 1% | 55% | 10% | 0% | 1% | 5% | 20% | 15% | 11% | 11% |
| Other quantitative data | 4% | 2% | 6% | 4% | 3% | 4% | 1% | 11% | 3% | 16% | 0% | 5% |
| Qualitative data (content analytics) | 12% | 18% | 7% | 24% | 5% | 18% | 6% | 12% | 13% | 1% | 9% | 11% |
Content analytics
Using Python, we read thousands of press releases every day and convert words into quantitative data. A press release announcing a stock split or a share repurchase, for instance, becomes a data point in our dataset.
Financial databases tell you the date a stock split and the number of shares repurchased in a quarter. By analyzing press releases, we capture the announcement, which is the most important input for a trading strategy. With over 500 qualitative signals translated into quantitative variables, we use a much wider range of data than most investors.
Sector neutral
Many quantitative services show great historical returns, but they have been overweight technology for years. Is it skill, or luck? We believe there are great opportunities in every sector, and that sector picking is hazardous. The best long-term approach to keep a reasonable level of active risk is to stay sector neutral and generate alpha across all sectors. That is why our Quant Picks always cover every sector, with weightings closely matching the index.
Sampling and universe
For U.S. and Canadian equities, our stock sample consists of the S&P 500 Index and the S&P/TSX Composite Index. As a result, our Quant Picks are mostly large-cap stocks. Equal weighting creates a small size bias compared with the market-weighted indexes.
We also rate internationally domiciled companies listed in the U.S., using the same methodology as for U.S. stocks. See the FAQ for how this coverage differs.