Disappointing season proves that analytics don’t work. #mycolumn:

There are two games left for the Maple Leafs this season, but I don’t need to see them to know what I know. You see, this was a season of new beginnings. Brendan Shanahan was hired, and with him came a vision for a new way of thinking, utilizing analytics heavily. He signed supposed wunderkind Kyle Dubas, known for assembling a powerhouse OHL team with the help of analytics. It’s obvious that that was a fluke.

And this isn’t just my opinion – I’m going to use the nerds’ fancy stats against them to prove that Kyle Dubas is a fraud and analytics don’t matter. 

How come, with this new genius braintrust and foolproof plan built on analytics, the Leafs still finished in the bottom five in even-strength corsi %? And if corsi is that important, are Carolina and Dallas in the top 10? How is Edmonton ahead of the Leafs in corsi, but behind them in the standings? How is Los Angeles first in the league and still not guaranteed to even be in the playoffs? How is Montreal bottom 10 and so good? How is Calgary worse than Toronto but in the playoffs? It makes no sense.

The crazy thing is I, like many smart people, pointed all of this out last summer when these “genius moves” happened and no one listened. Analytics don’t matter. The crazier thing is what I saw with my eyes I just backed up using numbers. Eat that, nerds.

The worst part about this bad hiring was that it lead to an even worse firing. People blamed everything on Randy Carlyle, the one coach that proved he could get the Leafs to the playoffs. Then management replaces him with a more “analytics-friendly” coach in Peter Horachek and look what happens. The team plays some of the worst hockey it has ever played. I need no more evidence – analytics are bad. 

Look, we all love our science experiments. And that’s what this season was – an experiment. I don’t think it’s responsible to use a phonebillpayer-funded team as a science experiment, but at least now we know whose theories were unsurprisingly correct and whose were embarrassingly wrong. Even though it basically cost me $200 a month.

But maybe Shanahan became aware of the absurdity of basing your team on analytics along the way, which is what fuelled the best trade of the year – acquiring Zach Sill – a move that did was not particularly “analytically friendly”. What it did was use Shanahanalytics. More on that in a future column. 

We better hope that moves like the Zach Sill trade are a sign of things to come because if that’s the direction the management team wants to take, I’m all in. If it’s building around guys like Nazem Kadri and Jake Gardiner, we are in trouble. And I don’t need data to support that because I already know it fits what I see – those guys just don’t pass the eyetest.  

And just to be safe, Brendan Shanahan should fire Kyle Dubas.