Skip to main content

The Toronto Maple Leafs were the first team to pull the trick last week.

Wednesday in Detroit, the Leafs were heavily outplayed, right from the drop of the puck. They were outshot 38 to 14 at even strength. They had a possession rating of a little under 26 per cent, making it the third most lopsided game out of 447 in the NHL this season.

And yet they won.

Two nights later, Montreal did something similar, beating the defending Stanley Cup champion Los Angeles Kings 6-2 despite the fact they were outshot 46-20.

Their possession on the night? Only 37 per cent, their second worst showing of the year.

These are the kind of results that drive the critics of analytics crazy. How can a stat be useful and yet defied so regularly in games? Shouldn't it always be "right?"

It's widely known that hockey is significantly behind sports like baseball, football and basketball in terms of using analytics. And one of the biggest reasons for that is hockey is so much harder to distil down to one number or series of events.

For one, there are far fewer scoring plays in hockey than other sports (think goals versus touchdowns etc.). Fewer scoring plays means more chance is involved, both at the end of a game and the end of season.

Ultimately, that means winning in hockey is more heavily influenced by luck (or randomness) than other sports.

The other big difference between hockey is there is a goaltender – the single biggest wild card in any game, as evidence by James Reimer and Carey Price stealing the show last week.

All that makes hockey harder to measure and quantify. A unique sport, however, requires unique numbers to define what's happening.

Which is where possession comes in.

Gabriel Desjardins, an engineer in Silicon Valley who is one of the pioneers of hockey's analytics movements and has worked for multiple NHL teams, once calculated that puck possession at even strength was responsible for 37 per cent of an NHL team's record in a season.

The combined effect of luck and goaltending, meanwhile, is roughly 43 per cent.

That may seem to minimize how valuable possession stats like Corsi and Fenwick are for analysis, but if you take out the luck factor and goaltending, those numbers make up 65 per cent of what remains.

And that's before special teams – about 20 per cent of games – are even considered.

That's how important simply having the puck more than the other team is.

The main aspect of possession stats that makes them so valuable, however, is they're quite predictive of future success. A team that is good at controlling play in the first 20 or 30 games of the season usually excels in that area over the long term, which is not always the case with other aspects of play.

It's a stable attribute, in other words, and that helps it forecast (better than other stats) how a team will do over the remainder of the season and into the playoffs.

In fact, TSN's Travis Yost has shown that score-adjusted possession over the final 20 games of the regular season has correctly predicted the winner of playoff series 70 per cent of the time in the last eight seasons.

Last season, it was 12-3, and the best possession team in hockey (the Kings) again won the Cup.

For a stat that only accounts for a little more than one- third of winning – and omits goaltending entirely – that's an impressive track record.

Is there room for improvement in hockey analytics? Absolutely, and that will likely come once optical cameras are tracking games beginning in 2015 or 2016 and generating more data.

But Corsi and similar analytics remain important foundation stats that give simple, strong insight into an often difficult-to-measure game.

They may not always be right night-to-night because (a) they aren't attempting to be and (b) that's not likely possible. But don't bet against them over the long haul.

'The trough of disillusionment'

Desjardins calls the period where hockey's analytics are at – and the fact they're so highly controversial – "the trough of disillusionment."

It's a term used in the tech sector and comes from Gartner's Hype Cycle that gauges the adoption and application of new technology.

Or, in this case, analytics.

"The timeline that people expect change is very short," Desjardins said, referencing how previously weak franchises like Edmonton, Toronto and New Jersey made big investments in analytics in the summer. "When there's no immediate, obvious benefit, you have disillusionment."

The fact NHL teams are so secretive about what they're doing with data also doesn't help. It's difficult to know which actions they're taking as a result of using these new ideas and whether or not they're listening to their new analysts. It's therefore sometimes hard to draw conclusions as to whether they're working or not.

Ross D. Franklin/Associated Press

Rising: Most improved goalies

1. Devan Dubnyk, Arizona. Was this the Oilers factor? Dubnyk appeared to have nearly played his way out of the NHL last season, when he cleared waivers and bounced around between three different organizations and even the AHL. The Coyotes took an $800,000 bargain bet on him, and Dubnyk has improved his save percentage more than any other regular starter. He’s only played in 11 games so it’s awfully early, but Dubnyk is 5-2-2 and his numbers are much better than starter Mike Smith.

2. Jonathan Quick, Los Angeles. His team has suffered a bit of a Stanley Cup hangover but not Quick. His .931 save percentage would be a career high if he can keep it up, better than even 2011-12 when he finished second in Vezina Trophy voting. Next to Nashville’s Pekka Rinne, who’s not on this list after missing most of last season due to injury, Quick has the best even strength numbers of any No. 1 leaguewide.

3. Craig Anderson, Ottawa. The Sens big step back last season came as a result of Anderson posting more pedestrian numbers than a year earlier. This year, the 33-year-old is back to being elite, even as his team has struggled considerably defensively.

Honourable mention: Ondrej Pavelec and Corey Crawford.

Save percentage
PlayerTeamThis yearLast yearDiff
1Devan DubnykARI0.9170.8910.026
2Jonathan QuickLAK0.9310.9150.016
3Craig AndersonOTT0.9250.9110.014
4Ondrej PavelecWPG0.9140.9010.013
5Corey CrawfordCHI0.9290.9170.012
6M-A FleuryPIT0.9260.9150.011
7Cam WardCAR0.9090.8980.011
8Brian ElliottSTL0.9310.9220.009
9Jimmy HowardDET0.9180.9100.008
10Jhonas EnrothBUF0.9180.9110.007

Associated Press

Falling: Most disappointing goalies

1. Chad Johnson, NY Islanders. Someone misses the Bruins defence (other than Johnny Boychuk). Johnson came out of nowhere to have a big season at 27 years old, with a .925 save percentage as Tuukka Rask’s backup last year. That earned him a big raise as a free agent (two years at $1.3-million a season) but his play early on has been pretty ugly. As always with goalies, beware of small sample sizes.

2. Ben Scrivens, Edmonton. Another goaltender who hadn’t played a whole lot before this season (72 appearances) and has been hit by the Oilers factor. Scrivens’s career save percentage before this year was .917. It’s already fallen to .911 with how poorly things have gone in Edmonton, and it’s tough to put all of that on his play. But he hasn’t been good.

3. Anton Khudobin, Carolina. Appeared well on his way to taking the starting job from Cam Ward and emerging as the next great Russian netminder, but it hasn’t worked out that way. Khudobin hasn’t won a single game yet and has been given only nine starts as Ward had a stellar November (.927 save percentage) to revitalize his career. This is one crease battle that’s far from over, however.

Dishonourable mention: Mike Smith and Tuukka Rask.

Save percentage
PlayerTeamThis yearLast yearDiff
1Chad JohnsonNYI0.8680.925-0.057
2Ben ScrivensEDM0.8870.922-0.035
3Anton KhudobinCAR0.8920.926-0.034
4Mike SmithARI0.8890.915-0.026
5Tuukka RaskBOS0.9110.930-0.019
6Semyon VarlamovCOL0.9090.927-0.018
7Ryan MillerVAN0.9000.918-0.018
8Kari LehtonenDAL0.9020.919-0.017
9Curtis McElhinneyCBJ0.8970.909-0.012
10Reto BerraCOL0.8820.893-0.011

Getty Images

Penalty plus-minus

One great new stat out there that doesn’t get talked about enough is penalty plus-minus.

What it measures is penalties drawn minus penalties taken, which highlights players that are piling up power plays for their teams.

Leafs centre Nazem Kadri has been an all-star in this department for years and currently is tied for the NHL’s lead with Ottawa’s Kyle Turris with 14 penalties drawn in 29 games.

The average NHL team scores on 18.5 per cent of its power plays; if a player draws 14 of them, that’s the equivalent to giving their club 2.6 goals – or more than 3 per cent of an average team’s total scoring – at this point in the year.

If they take penalties, however, they give back some of that advantage, where is where penalty plus-minus comes in.

Sometimes these players are speedy, shifty types like Kadri and Turris, that defenders can’t help but pull down when they make a move around them. Others who draw penalties are grinding types like Washington’s Tom Wilson, who has done well to get under opponents’ skin in front of the net.

Here are the NHL’s top players this season at giving their teams this boost:

NAMETEAMPlus-minusPer 60 minutes

The league’s weakest players in this regard? Steve Downie, Andrej Meszaros, Scottie Upshall, Nick Holden and Stephane Robidas, who are all sitting in the box far more than they’re sending someone else there.

- all statistics in this column are from prior to Sunday night’s games

Story continues below advertisement

Report an error Licensing Options