Winning the Game, Losing the Score – Why Analytics Matter

Image Credit: Michelle Gordon | Thunderstruck Sports

Four straight seasons. Four straight first-round exits—cue all the armchair general managers in the Tampa Bay area... the window has closed! The core has aged out! Cooper has lost the room! It's time to tear it all down! But let's slow down and take a breath. Yes, the Lightning lost to the Montreal Canadiens in round one, and yes, results matter. However, results don't always tell the whole story in hockey. This is where underlying analytics start to tell a different story. They can't replace the scoreboard, but they can help us answer a real and difficult question: Is this team actually getting worse, or are we allowing a small sample of outcomes to shape our perception of reality?

The Analytics Behind Game 7

Sites like MoneyPuck run simulations of every NHL game. In 500 simulations of this game, the Lightning won 83.6 percent of the time, while Montreal came in at a measly 16.4 percent. Shots on goal favored the Bolts 29-9. Expected goals favored them 2.65 to 0.97. High danger chances, while low for both teams, still swung in Tampa Bay's favor by a 6-3 margin. In his post-game comments, Jack Adams Award Winning Coach Jon Cooper poignantly explained, "Sometimes, you win the game and not the score." This game stands as a classic example of the gap between analytics and what ultimately shows up on the scoreboard.

The Skill-Luck Continuum in Hockey

There's a concept in analytics that helps explain this. In The Success Equation, investment strategist and author Michael Mauboussin describes a continuum between pure skill (like chess) and pure luck (like a coin toss). He argues that every activity falls somewhere along that continuum. Of every professional sport, hockey falls most heavily on the luck-dominant side. The low-scoring nature of hockey, along with its unique physical variables (like the bounce of the puck), increases the underdog's chances of winning on any given night—which is exactly why analytics matter when we try to evaluate performance over time.

The Time for Change?

If Game 7 had been based on skill alone, the Lightning would have won that game. The underlying analytics make that case convincingly. But if hockey leans heavily toward luck, what does that mean going forward? Do we make significant changes to the team that won every skill-based analytic? To a team that, by every underlying analytic, won the game—but lost the score? And what changes need to be made by the team that wins that Game 7 in 84 percent of simulations?

But the Lightning's fourth straight first round exit was not the result of game 7 alone. And although the Bolts did dominate the underlying metrics in the final game of the series, they did not dominate them in the previous six. Perhaps we need to change our question. Instead of asking what changes need to be made to the team that won the game but not the score in Game 7, we should be asking, "What changes need to be made to the team that took seven games to finally play to this capability level?"

The Strengths and Limitations of Analytics

Until the final horn in Game 7, many betting sites still had the Lightning as a strong Stanley Cup favorite. The analytics tell us that the Lightning were good enough to win round one of the playoffs. They show us that the window hasn't closed, that Cooper hasn't lost the room, and that it is NOT time to tear it all down.

But what the analytics cannot do is excuse patterns. Too often in the final month of the regular season and in the playoffs, the Bolts gave up the first goal, and found themselves chasing the game. Their power play struggled to produce. They couldn't seem to win any key face-offs. Too often, they did not look like the same team that went on a 20-1-1 tear around the Olympic break.

The Final Verdict

The analytics are clear in one respect—this is still a good Lightning team with potential for a deep playoff run in the coming year. But they are also a team that does not consistently play to their potential. So although it's not time to "tear it all down," it certainly is worth looking into some tweaks that could improve their consistency, and possibly tip bounces and puck-luck in their favor. A Luke Glendening type of forward who can reliably take key face-offs, a different power play structure with a strong shooter (this is where Steven Stamkos is still missed), another experienced depth defenseman. The underlying metrics in Game 7 were excellent. If only that Lightning team had shown up in games 1-6.