By: Dan Grant
On the eve of Canada’s first Olympic Men’s Ice Hockey match-up, against the terrifying powerhouse of… Norway? I thought it a pertinent time to continue our advanced stat guide series. We’ve already tackled baseball and football and you can bet basketball is coming sometime in the future.
I’d been putting off the advanced hockey stats guide because the stats are very ‘new’. Not new in terms of how long they’ve actually been around – teams have used them internally for years – but they’re not as commonly used by the hockey public, the way ‘sabreheads’ are common in baseball or virtually any stat is common in football. Hockey is a fast-paced team game that was one of the first to embrace non-traditional statistics. It’s been publicly tracking plus/minus for players back to the 1960’s; two assists have been awarded in hockey since forever, leading to other sports secretly tracking ‘hockey assists’ in their game; in other words, who makes the pass before the pass that creates the basket/goal.
This kept in mind, hockey is truly a team game. Six players vs. six players all with defined roles – but those roles are a matter of strategy, rather than of actual rules, except for the goaltender. If a team wanted to play five forwards or five defenseman, there wouldn’t be anything in the rules that said they couldn’t. The positions are made up. That is why advanced stats have been harder to apply to hockey players – what their ‘goal’ on the ice is isn’t always clearly definable. A player who scores 65 points in a season might seemed more valuable than one who scores 30, but the one who scores only 30 might be relied on for top notch defense, while the one who scored 65 points might have woefully underperformed his own ‘goals’, based on scoring chances and ice time. Still with me? Good, because I’m a bit lost already.
The point is, there are now a few readily consumable stats that I feel can be readily applied to the coming Olympic competition. It’s hard to be predictive in such a short tournament; obviously anything can happen in any game (ask Tommy Salo). That said, we can look at past performance and style of play and try to figure out which teams will excel at what aspects of the game. Then we can determine which aspects are more important and who has the best chance at success. Sound good? I don’t care, I’m doing it anyway!
An important new word for you to learn is ‘Corsi’. Named after former Buffalo Sabres goaltending coach Jim Corsi, the Corsi style of statistical metric is one of two that is widely used to evaluate stats at an advanced level in the NHL. The other is Fenwick, named after some guy named Fenwick. The difference between the two is literally only that Corsi includes blocked shots in its calculations and Fenwick doesn’t. There are billions of arguments for this and against this but I personally prefer Corsi because it sounds cooler. And because blocked shots are important.
CORSI – The original Corsi stat was meant to be an improvement over the plus/minus statistic we discussed earlier. Namely, that stat measured only whether a player was on the ice for a goal for or against but it ignored everything else that happened. That player could have had a horrible defensive shift, allowed five scoring chances that were unsuccessful and then had a puck bounce in off his ass and he’d still be credited with a plus 1. Corsi attempts to measure the amount of shots for and against, correctly arguing that this is a better indicator of how effective a shift the player has had. This can be applied on both the individual level, the team level and applied to different five man units. It’s a simple yet effective tool.
Once Corsi was established as a new standard for measuring players effectiveness on a shift per shift basis, other wrinkles were added. CORSI-QOC attempts to measure the quality of competiton (QOC), basically stating the fact that elite checking lines are bound to give up more shots, since they’re often playing against tougher opposition. Again, this works on an individual level, through the example I just mentioned, and also on a team level. Some teams pad their overall Corsi stats by blowing out terrible teams but struggle against tougher opposition, making Corsi-QOC an important statistic, similar to strength of schedule in the NFL.
The term for an individual players Corsi score is Corsi Relative or CORSI-REL. It compares a teams overall Corsi score when an individual player is on the ice, to when that player is off the ice. We can thus quantify how an individual player affects his teams overall effectiveness.
Take a breath. That’s basically it for Corsi. There are various other tweaks and wrinkles but that’s the majority. Other important stats include Zone Start/Finish which tracks what zone the puck is in when players change, the theory being if it’s in the offensive zone, that means they’ve ‘won’ their shift and PDO, which, similarly to BABIP in baseball, tries to eliminate the spectre of ‘luck’ from shooting percentage. I think it’s less effective in hockey, where any random number of factors can affect a goal, but it’s still an interesting metric.
So what does all this and a bag of hammers have to do with the Olympics you ask? Well if we can figure out which teams will dominate the Corsi statistics through their style of play/personnel, we might be able to make a fairly accurate prediction of who will get to the medal round.
Obviously Corsi statistics are only widely available for NHL players and a number of the European teams in the tournament feature lineups that are less than 50% NHL staffed. So looking at raw data can’t be our only avenue.
Let’s take a look at the pools:
Pool A: USA, Russia, Slovakia, Slovenia
Pool B: Canada, Finland, Norway, Austria
Pool C: Czech Republic, Sweden, Switzerland, Latvia
On first glance, it looks like defending champion Canada has lucked out in POOL B. Let me explain. After round robin play, four teams get byes into the quarterfinals and the other eight teams will go into an elimination game to determine the other quarter final opponents. That means each pool winner will advance and one ‘wild card’ team will go forward, based on overall record and points.
Of the 12 teams competing, since the NHL has participated in the Games, Canada has won gold twice, Czech Republic has won gold once and bronze once, Sweden has won gold once, Russia has won a silver and a bronze and the USA has two silver medals. Somehow, Finland has won two bronze and a silver in that time. While Slovakia and Belarus have both appeared in, and lost, Bronze Medal games. Germany and Switzerland have reached the quarterfinals (Switzerland twice), while Austria, Norway, Latvia and Slovenia have never advanced past the initial elimination round. Germany and Belarus failed to qualify for this Olympics. Austria and Slovenia fall into Canada’s initial draw, meaning their third game against Finland will likely determine who comes out of the pool with a bye into the quarterfinals. Cue Austria and Slovenia doing their best Rodney Dangerfield.
Teams that handle the puck more generally dominate most Corsi statistics. If you have the puck more, you can create more scoring chances and more shots on goal. Canada has loaded up on speed at these Olympics (Check out my preview here) and should use that speed to control the puck in the offensive zone. Finland has probably the best goaltending depth in the tournament with Tuuksa Rask and Kari Lehtonen between the pipes and Stanley Cup winner Antti Niemi ready to step in, in case of injury. Regardless of who the Finns start, they’ll need a great goaltending performance to best Canada. With a mobile defense core, led by Duncan Keith, Drew Doughty and Shea Weber, the Canadians are going to try to limit turnovers, make quick passes and control the zones. They’ll have to deal with the All-Time leading Olympic scorer, but he’s 43 years old and might be the Finnish Flash no more. I think Canada comes out of this pool fairly easily, as long as their own potentially shaky goaltenders hold up.
POOL A appears to the be the ‘Pool of Death’, with the USA looking to avenge its Gold Medal loss from 2010 and Russia having everything to prove with the home crowd behind it. Slovakia is always a talented team, led by the behemoth Zdeno Chara and Marian Hossa but an injury to Marian Gaborik and thin goaltending (a duo of Peter Budaj and Jaroslav Halak: yikes!) leads me to believe that Russia and the USA are in a two dog fight for the top spot in the pool. This could also be where our ‘Wild Card’ team comes from as well, unless Slovakia can catch lightning in a bottle and knock off one of the big dogs.
So who will win between Russia and the USA? I’ll go with Russia because they’re at home, but both of these teams are big and fast up front, with questions between the pipes. The USA gets credit for it’s goaltending depth because of Ryan Miller’s performance in Vancouver (it was legendary) and Jonathan Quick’s Conn Smythe trophy win, but both have had rough seasons; Miller is now 33 and Quick has looked extremely shaky at times. As for Russia, they have youth in net in Sergei Bobrovsky and Semyon Varlamov; if one of the two gets hot, the whole tournament is theirs for the taking. Few teams can ever match their talent up front with Alexander Ovechkin, Evgeni Malkin, Pavel Datsyuk and the exiled Ilya Kovalchuk (he’s back!) and if they can control the offensive zone, Corsi will love them and everyone else will fear them.
Pool C should be a battle between the Czech Republic and Sweden but I honestly don’t see this being much of a fight. The Swedish side boasts more talent in virtually area of the rink, even if Henrik Sedin doesn’t play. The Czech’s won their sole gold medal on the back of Dominik Hasek at the height of his powers, paired with Jaromir Jagr at the height of his. This time around, Jagr is still playing at 40, while Hasek is long gone and they haven’t been able to replace him. Not surprising, as at his best, Hasek was pretty much the greatest the game has ever seen. The Swedes have long played a puck possession game; the majority of their talent, NHL players included, came up through the Swedish Elite league, and are quite used to the big Olympic ice. The Czech’s have this advantage as well but boast one of the tournaments oldest rosters (beyond Jagr, they also sport Petr Nedved – no really – at 42, Patrik Elias, 38, Marek Zidlick, 37, Tomas Kaberle, 36 and Michael Rozsival, 35). The Swedes will come out of this pool easily and I could see the young Swiss team giving the Czech’s a run for their money.
That’s as far as I’ll go for now. Trying to predict beyond the quarterfinals is a fool’s errand at this point. I hope this was helpf–
“But Dan”, you’re saying, “Your link between the Olympics and advanced stats was tenuous at best. You really just talked about one thing and then the other. You basically led off by saying that the advanced stats could only be applied over wide swaths of data, while these teams have never actually played together! How are we supposed to take you seriously?”
“Well jerkface,” I reply, “It’s my column and I can do whatever I want! The advanced stats themselves don’t exist yet but the principles behind them are totally applicable. The teams that dominate the puck and are able to control the passing lanes will win the tournament”.
“So,” you retort, “You’re going out on the huge limb of saying that the teams that play better will win?”
“Hey” I sputter, “You can’t talk to me like that in my own column! I’m leaving!”