Like NHL Hockey Teams, StarMine Models Shoot — And Score

Here’s a wintertime riddle: what’s a good method for a hockey fan to predict which NHL teams will make the Stanley Cup playoffs – and what relationship could it possibly have with equity analysis? Let’s see what kind of puck each is chasing.

When hockey season is underway, many fans use statistical analysis to get a quick clue for the Stanley Cup contenders. These measures are relatively new — Fenwick For Percentage (FF%), Corsi For Percentage (CF%), and Goals For Percentage (GF%).  Fenwick tracks shots on goal + missed shots. Corsi  adds blocked shots and GF% is goals for divided by the sum of goals for and goals against.

In a similar manner, quantitative models can be used to predict the top performing equities.  We’ll demonstrate the similarities between predictive sports statistics and quantitative analytics used by our StarMine models to predict top and bottom performing equities.

Let’s look at the stats for the past five seasons, using the three metrics just mentioned to create a combined score.

Exhibit 1:  NHL top and bottom performers, last five seasons

Source: Five-on-five close game FF%, CF%, and GF% provided by Hockey Analysis

Who will make it to the playoffs?

Creating a combined score to rank NHL teams based on their performance in the prior season is shown to be have been successful 89% of the time over the past five seasons for picking teams that will (and won’t) make the playoffs. (As a homage to my alma mater, five- time NCAA men’s hockey champion Boston College, I’ll call it the Eagle Score.)

The idea behind the Eagle Score is that teams with greater possession and tendency to win in close game situations are more likely to make the playoffs. In the chart above, five-on-five close game FF% and CF% are used as a proxy for possession, while five on five close game GF% is used to measure the tendency to win games. A combination of these measurements is then used to produce an Eagle Score of 1 to 100, with 100 representing the highest rank. The Eagle Score makes it possible to easily identify stronger and weaker NHL team candidates for the playoffs.

Create your portfolio of next season’s Stanley Cup playoff candidates, along with those teams likely to be on the golf course in early summer. For winners, choose the top decile teams, or those with Eagle Scores of 90 or above. The table in Exhibit 1 shows this has been 100% accurate over the past five seasons!

Next, pick the also-rans by selecting the bottom decile teams, or those with Eagle Scores of 10 or below. As with shorting equities, there are no guarantees and teams, like companies, will take on more risk when pressured. Therefore, it is not a surprise that a few bottom decile teams managed to make it to the playoffs; however, it should be noted that within the past five seasons, none advanced beyond the first round.

So roughly 20 to 25% of the teams get tossed out of the picture strictly based on last season’s performance. For those looking to increase their number of picks that will make it to the playoffs, more in-depth analysis can be performed on the remaining teams where attention to recent events, individual players or matchups may be critically important to determining who will make it to the championship rounds.

Strong ARM returns

In a similar fashion to this methodology, we can use our StarMine Analyst Revisions Model (ARM) to identify companies that are likely to beat or underperform the market.

ARM is a percentile (1-100) ranking of stocks based on changes in analyst sentiment, with 100 representing the highest rank. This model is highly predictive of both the direction of future revisions and price movement. It looks at revisions in financial estimates –typically EPS, EBITDA and revenue — across multiple periods. Also, the StarMine quant research team examined the predictive power of analyst recommendations, and found that a change in recommendations mattered more than the absolute level, and ARM takes that into account as well.

Exhibit 2: Bombardier Inc. Analyst Revision Score vs. Price

Source: Eikon

Bombardier in gear

For example, look at airplane and railcar maker Bombardier Inc. (BBDb.TO). Analysts have raised both earnings estimates and recommendations, resulting in an ARM score of 93 that places it within the top decile of North American companies. Additionally, Bombardier’s share price has increased by 15% to C$4.36 from C$3.80 on Aug. 1, when the company’s ARM score hit 92.

Here are ARM scores for some northern companies:

Exhibit 3: Select Canadian equities with ARM scores of 90 and above or 10 and below

Source: Eikon

The Stanley Cup of Canadian equities

Similar to the Eagle Score methods used above to pick NHL teams for this year’s playoffs, we can use ARM to create a long-short portfolio of Canadian equities.

The results are worth noting. The StarMine quant research team’s October 2014 StarMine Performance Report shows that hypothetical purchases of large cap Canadian equities with an ARM scores of 90 or above (top decile) while shorting those with an ARM score of 10 or below (bottom decile) would have produced a trailing 12 month return of 43.6%.

Looks like the ARM model might help put the puck in the net.

Receive stories like this to your inbox as they are published. Subscribe here and follow us @Alpha_Now on Twitter or check out the Eikon blog. If you are looking to access our data or analytics, register for a free trial.