Former NHL goaltender Steve Valiquette hosted his first-ever hockey analytics conference today in Tarrytown, New York, attended by collegiate and professional players and coaches from across the country.
Valiquette, a current analyst for MSG Networks, introduced his “Hockey Needs Data” symposium by presenting data collected by his shot-tracking company, Clear Sight Analytics. In creating his company, Valiquette has tracked hundreds of hockey games at the U16, collegiate and NHL level to assess what makes a quality high-danger scoring chance in today’s game.
“Analytics” is a controversial word in hockey nowadays. People have a natural resistance to what they perceive to be over-analysis and needlessly complicated statistics. However, Valiquette not only breaks down his findings in easy to compute ways, he has video backup to illustrate his point.
Attendees filed into the Marriott conference room as episodes of MSG Networks’ web series “Vally’s View” played on the video screen. Former Ranger Andre Dore manned a table at the back where players could repair broken composite sticks. Former San Jose Shark Nolan Schaefer, who assists in tracking shots and assembling video for Valiquette, mingled with the guests.
I was genuinely surprised by the age range in the room. Coaches and collegiate players were the biggest demographics represented, but there were some junior, high school and even middle school players present as well.
The lights eventually dimmed and a brief video began to play. A title flashed stating that the stat “save percentage” was introduced in 1982. It then cut to highlights of goaltending in 1982- more accurately, 1982 goalies getting lit up. Goaltenders came out to the hashmarks to challenge shots. Dropping down on a shot meant flopping on your backside. It will never not astound me that there were not more goals scored pre-1990.
Eventually, the highlights shifted to shots from the same angles and locations in 2017, each resulting in routine saves. The style has changed. The pads are bigger. Yet the same basic statistic of save percentage is used to evaluate goaltender performance. This, Valiquette began, was in need of an update.
Merely discussing total quantity of shots on goal and shots saved is not enough. Instead, Valiquette used his vast amounts of data and game film to categorize shots on goal by the level of danger they pose to the opposing goaltender. In short: what makes a high-percentage, mid-percentage or low-percentage scoring chance.
“It’s not just about ripping numbers off of NHL.com,” he explained. “It’s about watching every game.”
High-danger scoring chances included breakaways (1 out of 4 convert), shots off passes that cross the slot-line once (1 out of 3) and shots off passes that cross the slot-line twice (2 out of 3). Though they happen infrequently, he explained, they are a very high-danger chance with very high shooting percentages.
Mid-percentage chances included shots with a deflection (1 out of 6.5) and shots with a screen set (1 out of 7). Valiquette brought up videos of screens being set, both by the attacking team and the defending team.
“Watch the goalie’s response to every [goal in which his own man screens him],” he joked. “You’re going to see a hairy eyeball.”
The optimal shot allowed for a defending team would be a clear-sight shot, the ultimate in low-percentage scoring chances. Valiquette used the 2014-2015 Rangers’ playoff run as an example. New York fired 252 shots on goal in the postseason that year that the opposing goalie had clear sight on. They scored only two goals. Against the Tampa Bay Lightning, they sent 96 shots on goal that Ben Bishop was able to track cleanly.
They did not score once.
On average, only one goal is scored every 34 clear-sight shots in the NHL. Clear sighted shots that produce rebounds only result in a goal once every 110 shots.
Other low-percentage chances, which came as a surprise to the room, included net plays like wraparounds and drives to the crease. Net plays typically only convert 7% of the time- an astonishingly low number when one considers just how often “driving to the net” is considered to be a hallmark of prototypical power forwards. Even rebounds off net plays only go in once every 55 opportunities, according to Valiquette.
“Driving to the net, it’s very sexy looking,” he began. “But it doesn’t create create chaos that people think…the goaltender is so big.”
This inspired me to ask a question later in the program. Valiquette was eventually joined by MSG Networks studio host Al Trautwig, former teammate Scott Gomez and head coach of Quinnipiac hockey Rand Pecknold. Fans and pundits alike seem to think that driving to the net is the true mark of a good power forward. It is a central talking point of Rick Nash’s game in particular, that if he is not getting to the net on his own, he’s called “marshmallow soft” by Mike Milbury on NBC Sports.
If fans and media (Milbury being a former player, coach and general manager) are both calling for the same thing, how does a role factor into one’s decision making process? If Rick Nash knows that he is going for a lower-percentage shot but that he’s fulfilling his pre-determined “role” assigned to him as a label, is this a good thing or a bad thing?
Gomez, being a former forward, said that it is the job of the coaches to block out outside voices in the first place. Pecknold noted that the decision to drive the net is all about reading the situation in terms of quality of opposing goaltender and whether or not the forward thinks they can beat them in a matchup. And besides, Gomez added, Nash is one of the premier scorers in the league for over a decade. Critiquing his play is almost a moot point with how successful he’s been.
Valiquette paused and said, “What we teach, we’re not saying ‘don’t take the shot.’ It’s ‘how do we enhance the shot?'” He elaborated, saying that it’s not enough to simply crash the net in a power move, but organizing a play where two other forwards can come in and join the attack creates more traffic and boosts the likelihood of scoring.
“It’s all about re-working the plan,” Valiquette finished.
This, in essence, is the motivation behind the Clear Sight Analytics conference. Players do not need to be overwhelmed with data to the point where they are thinking about numbers while skating around on the ice. However, using the numbers and percentages provided by Valiquette and company can shape the overall gameplan, so that in execution, players are still thinking pure strategy, but that strategy is based on what will maximize shot quality.
The panel specifically praised the defensive system of Scott Stevens, currently an assistant in Minnesota. Valiquette worked with goaltender Devan Dubnyk in the offseason, and Dubnyk reportedly much preferred Stevens’ method of facing more shots of a lower quality than limiting total shot quantity but facing more dangerous chances.
The crowd was very receptive to what the panel presented in terms of numbers and video. Valiquette mentioned one of the lowest percentage shots is from the “dead angle zone,” or the slice of ice from the post to the outside hashmark. He showed a clip of Sidney Crosby’s goal on Henrik Lundqvist from Friday night in which he banked a shot off Lundqvist’s head and into the net, sending titters of excitement and awe throughout the room.
When Pecknold began breaking down the puck movement that made the Washington Capitals’ powerplay so effective, the tic-tac-toe work of T.J. Oshie, Nick Backstrom and Alex Ovechkin drew audible ‘ooh’s’ and ‘aah’s’ from the crowd. It was as if everyone, young and old, was transformed into a little kid first discovering the beauty and precision of hockey highlights.
What helped Valiquette engage his audience was his application of his analytics and statistics to real names and faces. For every concept on how to defend, there was a video to back it up. Mats Zuccarello’s shooting habits were discussed at great length. Scrutinizing an individual player’s tendencies made the stats seem more real.
The panel also introduced a few new statistics to evaluate player performance. The first was “Goalie Rating,” a stat intended to be comparable to that of QBR in football. Valiquette explained that this metric measured a goaltender’s save percentage based on shot quality.
Towards the top of the league in this category through January 1st: Jimmy Howard, Anders Nilsson, Matt Murray, Roberto Luongo and Sergei Bobrovsky.
At the bottom: Michal Neuvirth, Kari Lehtonen, Antti Niemi and yes, Henrik Lundqvist, a revelation that drew an exasperated, “oh, come on!” from one attendee. But it makes sense, given that Lundqvist has taken a big step back this season in terms of number of clear-sighted shots saved. Last season, Lundqvist allowed only one goal every 41 clear-sighted shots. This season, that number is one every 29.
Valiquette wasted no time mocking that both Stars’ goaltenders appeared on his naughty list.
He then went on to explain metrics that evaluate skater performance. The first stat he introduced was “Adjusted plus-minus.”
“I don’t think anyone in this room likes using plus-minus alone as a tool,” Valiquette said. Thus, he introduced his new tool which tabulated all quality mid- or high-danger shots at 5-on-5, excluding empty-net goals.
The top three in this category? Patrice Bergeron, Brad Marchand and David Backes. It’s nice to play with the best defensive player in the league.
At the bottom of this list: Phil Kessel. He ranked 782nd in the league. I always knew Amanda was the #bestKessel. Also towards the bottom were Radim Vrbata, Martin Hanzal, Nick Leddy and Richard Panik.
Finally, Valiquette introduced a stat he perfected with Scott Gomez, one of the better playmakers of the 2000s. Fittingly, the Playmaker Stat quantified primary passes that lead to mid- or high-danger shots within three or fewer seconds. It’s a way of calculating the best passers in the league without solely relying on the assist category.
The stat came to fruition out of a conversation between Valiquette and Gomez, in which Gomez complained that he was setting up teammates beautifully but teammates were not burying.
At the top of the league, in order: Derek Stepan, Mats Zuccarello, David Krecji, Nick Backstrom and Connor McDavid. Valiquette pointed out that Alex Ovechkin sat 217th in the league in this category, purely because he is a shooter and not a passer. Being towards the bottom of the league in this category does not necessarily detract from a player’s value.
Can’t hurt to be towards the top, though.
Marian Gaborik was the lowest rank that Valiquette mentioned, 399th in the league. It’s been a rough year for Gaborik, who is not too far removed from setting the Rangers’ single-season even-strength assist record.
The names on these lists are not surprising. They are a good measuring stick without upsetting the apple cart too much on what was already preconceived. But these simple adjustments to already existing statistics modernize the game and help shape the way players can be compared and judged.
“When I was a player, plus-minus was a big thing,” Gomez said. “And they counted empty netters. So I remember being out there for a few empty netters to start the year, all of a sudden I’m a -6. That’s not really fair.”
Al Trautwig, in his introduction to the panel, gave a brief explanation of his introduction to analytics. “When I covered the Yankees in the 1990s…analytics was a weird word. It wasn’t used. Then the Red Sox hired an executive who specialized in analytics and they won their first World Series since 1918. And then they did it again. And then that executive went to Chicago and they win their first World Series since 1908.”
“I thought, okay, there’s something to this.”
The analytics movement is usually met with skepticism initially. Gomez told one anecdote in which he presented a Devils coach with a tape created by Valiquette that gave a certain breakaway strategy that could be used to convert more shootouts. Gomez became more successful in the shootouts following the tape’s advice, but it took two weeks for his coaching staff to show the rest of the team (and when they did, Gomez was not invited to the meeting).
Valiquette said that when he presented his data to Rangers assistant general manager- and former teammate- Chris Drury, Drury admitted that he “couldn’t unsee” the validity of it.
“Old school coaches will tell you, ‘go out and play,'” Gomez said. “‘What do you mean, just go out and play? I’ve never been on the fourth line before. Do I dump it in? Get to the net? How does this style work?'”
Gomez admitted that Valiquette’s data stuck with him while on the ice and that when it occurred to him while playing, he would alter his play to go for higher percentage shots. More often than not, he said, it benefitted him and his scoring improved.
Pecknold and Valiquette rolled practice footage of the Wild and Red Wings in which shooters rapidly approach unimpeded and shoot from the circles, while the goaltender lazily stands in the way.
“That particular shot happens once every 1.4 games,” Valiquette said. “It’s a pointless warmup. It does nothing.”
Valiquette implored the coaches in the room to practice shots from angles and cross-slot passes to properly engage goaltenders and warm them up. The practice of high-danger chances will also help the development of forwards.
Said Pecknold, “[Analytics] have certainly got me a lot more wins. It’s made me think differently, especially what a goalie needs in practice.”
A range of other topics and stories were exchanged, from Pecknold’s recruiting process to Gomez’s general disdain for coaches disinterested in connecting with players, to Michal Roszival’s love of egg sandwiches.
Though Valiquette admitted that hockey “isn’t ready” for a full-on analytics revolution, and that stats need to be standardized, he was optimistic that it would not be long before teams caught on.
“We don’t have resources we count as exact or factual…every team has their own way of calculating scoring chances,” he stated. And he should know- Valiquette played for many coaches, joking that his Hockey Database page was “embarrassingly long.”
“But it’s there,” he continued. “It’s undeniable at some point. If we showed every shot with video, it couldn’t be argued.”
Herein lies the crux of the analytics movement. Old school hockey minds, be they in the print media or front office management, swears off statistical analysis from the keyboards of bloggers from mom’s basement (yes, this is a real analogy used insultingly frequently). Watch the game, they will say.
Well, Steve Valiquette has watched the game. He’s played the game. He’s micro-analyzed the game in every way at every level, and the data remains the same. At the end of the day, why resist what could ultimately benefit teams in the long run? It’s already proven the data is pretty universal regardless of skill level.
“Hockey is hockey,” Valiquette said.
For more information or clarity on Steve Valiquette and Clear Sight Analytics, check out MSGNetworks.com’s series, “Vally’s View.”
