Sign In
Before we begin, please note that our answer below is only one of an infinite number of correct answers. Furthermore, metrics are not constrained only to athletes and sports, and could be applied in countless contexts.
To begin, let's consider ice hockey as the sport whose metrics we will compare, then assess the quality of some players.
Ice hockey is a game-type sport. As such, the metric that defines which team wins a match is the score — whichever team has more points at the end of the game wins. Therefore, one metric that could be used to compare players to one another is the average number of points each one scores per game. Let's look at some players from a certain local team.
Player Name | Position | Points Per Game |
---|---|---|
Timmy | Right wing | 5.1 |
A.J. | Goalie | 0.1 |
Chester | Center | 1.7 |
It looks like Timmy is a high-scoring player, while A.J. barely contributes to the team's score at all. Is points-per-game an effective metric for assessing the performance of these players?
Points-per-game would not be an effective metric for assessing player performance.
Consider that the opposing team's objective is also to score as many points as possible. Timmy might score a lot of points, but maybe A.J. and Chester contribute defensively as well. Let's look at a few more statistics.
Player Name | Position | Points Per Game | Assists Per Game | Shots Blocked Per Game |
---|---|---|---|---|
Timmy | Right wing | 5.1 | 1.9 | 0.2 |
A.J. | Goalie | 0.1 | 0.3 | 14.7 |
Chester | Center | 1.7 | 4.8 | 6.0 |
An ice hockey team generally has 6 players on the ice at any given time during a game, each playing an equally important role.
As we saw above, a single metric on its own is insufficient for an effective model. In such situations, it is helpful to combine several metrics into one new metric. Let's combine Points scored, Assists made, and shots Blocked into a new metric called P/A/B.
Player Name | Position | Points Per Game | Assists Per Game | Shots Blocked Per Game | P/A/B |
---|---|---|---|---|---|
Timmy | Right wing | 5.1 | 1.9 | 0.2 | 7.2 |
A.J. | Goalie | 0.1 | 0.3 | 14.7 | 15.1 |
Chester | Center | 1.7 | 4.8 | 6.0 | 12.5 |
Combining metrics often paints a more accurate picture of a scenario than observing individual metrics. However, there are many other possible metrics to compare player performance, so even the P/A/B is not necessarily a totally effective metric.