Minnesota Wild and Advanced Stats: Charlie Coyle’s Hidden Game


Mar 30, 2013; St. Paul, MN, USA; Minnesota Wild forward Charlie Coyle (63) celebrates his goal with forward Mikko Koivu (9) during the second period against the Los Angeles Kings at the Xcel Energy Center. Mandatory Credit: Brace Hemmelgarn-USA TODAY Sports

Charlie Coyle’s Hidden Game

What traditional stats don’t tell about the Wild’s standout rookie forward

By Andrew Setterholm and Ger Devine, Gone Puck Wild staff writers

Last season, Charlie Coyle was undoubtedly the Wild’s most effective rookie forward. In highly anticipated seasons for the Wild’s many hot prospects, Coyle had a more consistent and productive season than his fellow first-year NHLers – including Mikael Granlund and Jason Zucker.

For those who didn’t keep a close eye on the Wild youngsters last season, Coyle’s success might not be as apparent – traditional NHL stats don’t do the young man justice. In 37 games played, Coyle’s traditional stat line is not hugely impressive:

GP: 37     G: 8     A: 6     Pts: 14

+/-: +3     PIM: 28

Shots: 50     S% 16.0

GPW staff writers Andrew Setterholm (AS) and Ger Devine (GD) took a closer look at Coyle’s game with the help of some advanced stats. For a guide on how to use “fancy stats,” check out staff writer Ger Devine’s previous article on the subject, or follow along (and join in) on the discussion that follows. With advanced stats, we’ll be putting some concrete numbers to Coyle’s real significance in the Wild offense last year.

What do traditional statistics tell us about Coyle last season?

AS: For me, the traditional stats are not very telling of Coyle’s impact on the game or what he brings to the team. His point total isn’t outstanding. His shot total, for a first-line player, is nothing to brag about. Shooting 16 percent is great, but that’s a consequence of his shooting so infrequently.

GD: I think the 1st thing to remember with Coyle is that we’re dealing with a very small sample size – usually 80 games is considered a sufficient amount to get a reasonably clear picture of a player. However, Coyle really couldn’t have impressed me more with his performance this year as a rookie. I would much prefer to see a young guy like Coyle go out there and put-up great possession numbers and low points totals as opposed to scoring a bunch but being terrible in possession. Good production is nice, but being able to dominate at 5v5 as a rookie means that it’s very likely you will have a long and productive career as an NHLer.

What do advanced statistics say about Coyle?

Coyle’s Corsi On was 10.75 – the third highest on the team for forwards who skated 20 or more games. He trailed only his linemates Mikko Koivu (12.37) and Zach Parise (11.32).

(Corsi stats from www.behindthenet.ca, Wild forwards, 20 games played)

AS: This stat alone tells me that while Coyle isn’t getting a ton of shots himself (realistically, very few compared to his line-mates), when he is on the ice, the team is winning the puck-possession battle by a wide margin.

From watching his game, we know Coyle is doing his most important work near the boards, winning puck battles and helping his line retain puck possession and create scoring opportunities. This doesn’t show up in his traditional stat line, but looking at his Corsi clearly shows his effectiveness as a line mate to the team’s top scorers.

GD: The knock against Coyle is that he played with the team’s two best forwards, which boosted his numbers considerably. This is a fair statement to make about a rookie, but I would argue that it’s very difficult to be a passenger on a line with players like Koivu and Parise because their playing style is based on relentless hard-work and energy rather than effortless skill, therefore, if the 3rd forward is just floating, it would be very obvious. Dany Heatley wasn’t able to last on that line, Coyle was.

What do other stats indicate?

GD: Even just going by the “eye-test”, Coyle looked like he had a great understanding of what Parise and Koivu were doing and he was able to match them every step of the way. His WOWY (With Or Without You) numbers show some evidence of the chemistry between Coyle, Parise and Koivu:

5v5 Corsi For %5v5 in Close Situations, Zone-Start Adjusted Corsi For %
Coyle with Koivu: 59.4Coyle with Koivu: 55.6
Coyle without Koivu: 41.7Coyle without Koivu: 40
Koivu without Coyle: 52Koivu without Coyle: 51.4
Coyle with Parise: 59Coyle with Parise: 54
Coyle without Parise: 44.1Coyle without Parise: 47.1
Parise without Coyle:  51.9Parise without Coyle:  51.1

You can see from those numbers that all 3 players experienced a drop-off in Corsi For % when away from each other. Obviously Coyle’s drop-off is more extreme as his time away from the 1st line was usually spent on the 3rd line, where he was with inferior linemates, playing a different role. But the numbers show that Koivu and Parise both performed better with Coyle on the 1st line, as opposed to when Heatley, Jason Pominville or anyone else was there. Similar trends occur when looking at GF20 and GA20 instead of CF%.

Is Coyle’s success (based on advanced stats) predictive of future performance?

AS: This is where advanced stats can get tricky – predicting the future is, necessarily, a very uncertain business. Some points to consider:

Coyle’s offensive zone start percentage was 64.6 percent. His offensive zone finish percentage was 50.3 percent. This tells us that Coyle was getting a high percentage of his shifts in the offensive zone, presumably after another line got the breakout and created an offensive zone face off. If the Wild choose to skate Coyle on the second line next season, we can expect his offensive zone starts to drop. He might find himself near Heatley’s roughly 50 percent mark. This would mean less time in the offensive zone and fewer scoring opportunities, and a lower Corsi On.

GD: I agree that he will probably see his role change next year (presumably with him switching to C or RW on the 2nd line), and that will lead to a lower percentage of offensive zone starts, but to balance that out, he will likely also face softer competition. I assume coach Mike Yeo will continue to deploy the Koivu line against the other team’s best forwards, so Coyle could thrive away from that extremely demanding assignment. Personally, I would keep him on the 1st as he clearly wasn’t fazed playing a big role last year, and he will be coming back a year older and with a full training camp to build on.

In terms of how predictive of future performance Coyle’s underlying numbers are, I would say that there is enough data there to show that he wasn’t just a passenger on a good line, he wasn’t sheltered and he had a big role in generating solid puck possession for the team, so, unless something goes horribly wrong, we should be happily pouring through an impressive 82-game sample of Coyle’s possession numbers this time next year and he’ll be well established as one of the best young forwards in the league.

AS: Coyle did lead the team in Corsi Relative Quality of Competition (1.614). Hopefully, this indicates Coyle’s ability to play and compete at a high level. If so, it would be reasonable to expect Coyle to either maintain consistency on the top-line, or step up his scoring as a second-line forward. I look forward to seeing where he comes out of training camp, and what he can do with a full season under his belt. I think the Wild and Minnesota fans would be right to expect big things from Coyle.