Drilling Down: Columbus Is Playing .500 Hockey, But Do The Analytics Point To Overachieving?

By Ed Francis on January 3, 2022 at 2:45 pm
The Columbus Blue Jackets are a .500 team through heading into the new year. The analytics, however, suggest they should not be.
Russell LaBounty-USA TODAY Sports

It's a weird time to be a Columbus Blue Jackets fan.

The organization wasn't expected to do much this season, and even the most optimistic fans thought the playoffs were only slightly more than a possibility. 

The Blue Jackets had an impressive training camp and preseason though, then picked up wins in seven their first ten games — tying a club record for their best start. Through 18 games, Columbus won 12. They had won all five of their overtime games, beaten several Stanley Cup contenders (including Colorado twice in four nights), and things were starting to feel different.

Enter, December.

Columbus had a putrid final calendar month of 2021. A two-week pause in games as a result of COVID-19's omicron variant came in the midst of the Blue Jackets' toughest stretch of the season. They finished 3-7-1 in December, and 3-8-1 one if you start the month one day earlier. 

It was a tough, slow dose of reality that the Blue Jackets are still a young and inexperienced team with key weaknesses. 

Another tough dose is that, even counting the recent struggles, a deep dive into the analytics show that the Columbus Blue Jackets have radically overperformed this season. 

In fact, they should be ... quite a bit worse. 

The hockey analytics site MoneyPuck does a highly in-depth dive into each play of each game. This is anything from what shots went in that shouldn't have, to what percentage a puck would go in the net based on the angle from which it was shot.

During and after each game, MoneyPuck calculates the odds of a team winning a game based on how many goals a team should have scored — and shouldn't have allowed. They call it the "Deserve-To-Win O'Meter", and here's how they define it:

"A feature on our live game pages which shows the each team's chance of winning the game had the game been re-enacted, except with average goaltending for both teams. The meter works by simulating all of the individual expected goals that occurred in the game. For example, consider (a) scenario where the home team has taken one shot worth 0.5 expected goals and the away team has taken one shot worth 0.1 expected goals."

Let's look game by game. In the chart below, dark green games are the best: Columbus won, and the analytics say they deserved the win. In light green are games Columbus won but probably shouldn't have, based on the metrics. Light red returns — and there aren't many of them — are games Columbus lost, but probably should've won. The dark red returns are deserved losses for the Blue Jackets, and yellow-shaded games are games that were just too close to call.

10/14   vs ARI W 8-2      20.8% No, should've lost
10/16   vs SEA W 2-1       67.7% Yes, should've won
10/19   @ DET L 4-1      24.6% Yes, should've lost
10/21   vs NYI W 3-2      57.3% Yes, should've won
10/23   vs CAR L 5-1      13.1% Yes, should've lost
10/25   vs DAL W 4-1      39.6% No, should've lost
10/29   @ NYR L 4-0      40.9% Yes, should've lost
10/31   vs NJ W 4-3      34.2% No, should've lost
11/3   @ COL W 5-4      46.7% No, should've lost
11/6   vs COL W 4-2      47.3% Toss-Up
11/12   vs WSH L 4-3      80.9% No, should've won
11/13   vs NYR L 5-3      68.4% No, should've won
11/15   vs DET W 5-3      72.2% Yes, should've won
11/18   @ ARI W 5-4      83.4% Yes, should've won
11/20   @ VGK L 3-2      11.1% Yes, should've lost
11/22   @ BUF W 7-4      36.8% No, should've lost
11/24   vs WPG W 3-0      36.5% No, should've lost
11/26   vs VAN W 4-2      25.6% No, should've lost
11/27   @ STL L 6-3      37.5% Yes, should've lost
11/30   @ NSH L 6-0      13.1% Yes, should've lost
12/2   @ DAL L 3-2      18.8% Yes, should've lost
12/4   @ WSH L 3-1      22.5% Yes, should've lost
12/5   vs SJS W 6-4      32.5% No, should've lost
12/7   @ TOR L 5-4      48.3% Toss-Up
12/9   vs ANA L 2-1      66.5% No, should've won
12/11   @ SEA W 5-4      58.0% Yes, should've won
12/14   @ VAN L 4-3      37.9% Yes, should've lost
12/16   @ EDM L 5-2      26.7% Yes, should've lost
12/30   vs NSH W 4-3      45.8% No, should've lost
1/1   vs CAR L 7-4      20.1% Yes, should've lost

It's far from a perfect system, as nothing is perfect. But the results are fairly clear: the Blue Jackets should not be 15-14-1 right now. They have won just five games this year that the analytics say they should have won: one-goal wins at home over the Seattle Kraken and New York Islanders (both went to overtime), wins over Detroit and Arizona in back-to-back games in the middle of November, and the rematch with the Kraken in Seattle. 

Just two games have been a true "toss-up" — where the odds of within were within 3% of a 50/50 chance. The Blue Jackets are 1-1 in those games, winning at home against Colorado in November and losing in Toronto in December. 

The most common result are deserved losses. Eleven of Columbus' 30 games have fit that category this season. Behind that though, is where the devil comes into the details: nine of Columbus' 15 wins (60%) have been games where metrics would typically say the Blue Jackets don't walk away with two points. 

Some games may seem a little funky — no better example than the opening night win over the Arizona Coyotes. Columbus won 8-2, and Nationwide Arena could've belted out "The Hockey Song" in the second period to signify the Blue Jackets win. 

Look deeper though, and the game shouldn't have been that close. Arizona let in all kinds of goals that shouldn't have counted — the Blue Jackets had an expected goal tally that game of just 2.01, while the Coyotes expected goal count was at 3.95. 

The 5.99 difference between goals and expected goals for Columbus was by far their biggest of the season, and one of only three times this season they've bested their xGF by three or more goals. Consequently, the Blue Jackets allowing 1.95 goals less than expected was their fourth-best performance of the season.

This isn't to say the Blue Jackets got "lucky", per say. This was an outlier, but the game should have been closer than it was. Ultimately, all this just puts merit to the proverbial "that's why they play the game." 

Columbus is going to be really good, probably sooner rather than later. This year was supposed to be growing pains and the 5th Line is seeing those pains in all their splendor right now. 

If everything that was "supposed" to happen on the ice, the Blue Jackets would be roughly 8-20-2 right now. The 18 points they "should" have would tie them for second-worst in the league with the Montreal Canadiens, and would put them three points ahead of (ironically) the Coyotes.

Instead, they're at 31 points. There's a little snag in the number of games played, but they're just three points out of a potential (though unlikely) playoff spot. Maybe 15-14-1 isn't so bad, after all.

At least not this season.

View 2 Comments