«« Prose, Look and Feel Blogs at FanConcert »»
blog header image
Dangerous Metrics in Olympic Ice Hockey and Project Management

The software project management course at the University of Ottawa was pretty good but there were points that stood out. One great point was that people will optimize their behaviour depending on how you measure it.

In the software development world, project managers (PMs) sometimes collect metrics so they can gauge their project's progress. For example, a PM may measure the number of tasks each programmer completes per day. He may also measure the number of defects the QA team reports per day.

These metrics can be quite dangerous if you don't pay attention to the side effects. From the above examples, how big are the tasks in question? Each task can take a different amount of time for a developer to complete. QA personnel could submit many trivial defects in order to increase their daily counts. Quality suffers in the name of quantity. These side effects can have negative consequences for the project.

Let's switch gears to Olympic hockey. The Canadian women's Olympic team beat their three opponents in the round-robin part of the tournament by a total of 36 goals to 1, giving them a goal differential of 35.

Now the women's team is taking some serious criticism for running up the score on poorer teams. Among others, Don Cherry has said that dominance by one or two teams in an olympic sport could lead to its cancellation, like baseball/softball in the summer Olympics.

Why did the Canadian women's team run up the scores? Because the tournament rules state that the team with the highest goal differential gets home-ice advantage[1] in the gold medal game. Home-ice advantage can be very important so the Canadian team is running up the scores.

If people don't want the really good teams running up the scores on poorer teams, they need to remove the incentive to do so. Goal differential in the round-robin is a poor measurement anyway because it's often the case that the two teams in the gold medal game played in different pools against completely different teams. Canada had a large advantage in their pool by playing Italy, a team that is only playing in the tournament because they are the host country. Canada beat Italy by a score of 16-0.

How can they determine home ice advantage in the goal medal game? How about world rankings before the Olympics? The teams are already seeded so that the pools are balanced. Countries could be seeded overall based on past official tournament performance or even the previous olympics.

[1] Home-ice advantage gives the team the last change on face-offs as well as other advantages.

Posted at February 17, 2006 at 07:36 AM EST
Last updated February 17, 2006 at 07:36 AM EST
Comments

Ron Mclean reported today that the IOC is considering not allowing the host country to automatically play in future tournaments (because of this problem).

» Posted by: V. Mardian at February 17, 2006 05:24 PM

It turns out that because Canada had home-ice advantage in their first game against Sweden they do *not* get home-ice advantage again for the final.

Canada wasn't just trying to get home-ice, they were also trying to get into a rhythym of scoring goals. They had a big loss to the US in the last major tournament because they couldn't score. They seem to be more focused on putting the puck in the net this time.

» Posted by: Ryan at February 19, 2006 09:51 AM

I think the best way to decide home ice advantage is with a coin toss. That's how they do it in the superbowl, and actually I think all football games? Anyway, when your on neutral territory, it's probably best to use something that isn't based on past performance. I think the great thing about the olympics is that anyone can win. If the teams that did well before the finals get an extra advantage, then they are more likely to win. The only way to make it fair, is to make it random.

» Posted by: Kibbee at February 21, 2006 08:35 AM
Google
 
Search scope: Web ryanlowe.ca