PPP: How is your summer going?
CHID FINDS: The summer has been a roller coaster, from the tough loss in game seven to prepping for the draft, then qualifying RFAs, targeting UFAs. It’s a lot to prepare for, there’s not much of an off-season anymore
PPP: Can you explain your role with the Leafs?
CF: I'm the assistant to Chris Dennis. Chris was recently appointed to be an assistant coach to Randy Carlyle -- the club made the decision that they want to have as many men behind the bench as they have in the executive suite. Chris does a statistical analysis after each and every game. He’s built a system to track and evaluate at least half a dozen data points. I help collect and maintain that data. For instance, during the season, I’ll log into NHL.com and look at the stats page a couple of times a day. Maybe change the sort function or run a new query.
PPP: So the Leafs are doing statistical analysis?
CF: Oh, absolutely. Some older hockey guys may not want to admit it, but it’s a segment of the game that we’d be foolish to ignore.
PPP: The Leafs have seemed "anti-stats" to a lot of fans, I’m surprised to hear this.
CF: We may be late to it, but we’re fully on-board. Right after Mr. Leiweke was hired, he asked me for two things - a good road atlas of Toronto and to quantify our home record when Alan Frew sings. It turns out we lose a lot of games when we play certain songs. It's these types of causal relationships that are worth exploring and can assist with executive decision making.
PPP: What are some of the metrics that you track?
CF: I don’t want to get into too much detail about our proprietary methods, but I can tell you that we delineate between shots on net and scoring chances. A scoring chance is not a shot on net and a shot on net is not a scoring chance. We track all kinds of data -- wins, losses, goals for and against, who is first on the ice at practice, first off, first into the weight room and more. We also look at what time during the game goals are scored and we’ll try to make sure we’re paying extra attention to defence during those moments. In Game 1, Boston scored at 16:48 in the first period. If you check the boxscores for the rest of the series you’ll see we didn't let them score at 16:48 again - for six straight games. That’s the power of big data.
PPP: What about what people refer to as "advance stats", are you tracking them?
CF: Look, Corsi is a great proxy for possession, but you can’t do much by proxy. You don't need me to tell you that it's very difficult to score a goal by proxy. The only way I know of is if the other team puts it in their own net. I've crunched the numbers and that’s an exceedingly rare event. It's certainly no way to build a team. For those fans who don't watch the games and just want us to have better Corsi numbers, they are going to be very pleased with the addition of David Clarkson. There’s a player who’s all about quantity of shots over quality.
PPP: I've read quite a bit about the Leafs and shot quality. Do you believe the team was effective at suppressing shot quality in favour of shot quantity?
CF: This is what our data shows. We gave up what – 32 to 35 shots a game -- you could look it up at NHL.com -- but how many of those shots were goals? Maybe 3 each night. That shows we weren't giving up quality shots. These NHLers can really fire a puck. If we hadn't been suppressing quality shots, we’d be giving up six, seven, maybe more goals every game.
PPP: And you track this? You can measure this?
CF: Yes. I have spreadsheets that tracks this on a game by game basis. Shots on goal. Saves. If you divide one by the other you can create an emerging stat that I’ve suggested we call "Save Percentage" but the assistant coaches call it "Asymmetrical Shot Suppression." We also track each and every zone exit, we measure how hard someone banged the puck out off the glass. I can assure you it’s an incredibly rare occurrence for a team to score a goal on a puck banged out off the glass. That's why it's a great hockey play. Much safer than carrying the puck up ice, which can lead to a turn over.
PPP: What’s been the most surprising thing the numbers have revealed to you?
CF: I think stats are strongest when they confirm or complement what you’re seeing on the ice. One of our favourite stats is like the NBA's hustle stat, it combines Shots, Hits, Intagibles, and Takeaways. Another stat is ArFat/60 – which measures arm fatigue over a 60 minute game. We think if we can give up a lot of outside shots, keep teams to unproductive possession on the perimeter they’ll tire out. When they get the puck near what Coach Carlyle calls the "home plate area" their arms will be too tired to get off a quality shot. It’s this type of thinking that might reveal that it’s far better to play without the puck and give up lots and lots of low quality shots.
PPP: What do you think you've helped the Leafs accomplish for next season?
CF: I think we've done a really good job of helping management find new market efficiencies to make us more competitive.
PPP: You mean market inefficiencies?
CF: No, efficiencies. The NHL dresses about 900 players a year. As I've explained to Mr. Nonis, over 48 games that provides us with a huge sample size. Enormous. A lot of good work has been done by other teams in finding inefficiencies in the market. So much good work that we now believe there are no more inefficiencies left to exploit. In such a marketplace, we can take advantage of the one thing the Leafs have and that's cash. That's why we're willing to pay above market value for things that used to be inefficiencies. That's efficient. Other teams can't use their cap space like that.
PPP: Can you give us an example?
CF: We’re striving to find places that no other team values like fighting or goaltender stick handling skill. Bernier is going to be a big upgrade in this regard. He’ll be able to quickly move the puck to our d-men who can move the puck off the boards or bang it off the glass -- we support creative hockey, it's their choice to bang it off the glass or blast it off the boards. We've also looked at how we can overpay for limited returns such as face-off wins or for marginal 7th or 8th d-men.
PPP: The Leafs have taken a lot of heat in some quarters for their off-season moves. Many people can't see the reasoning behind it. How do you think the Leafs compare with other teams in the use of statistical analysis?
CF: I think we're in a very good spot. I was talking to a Western Conference club who have a dedicated statistical and analytics unit. Really smart guys. I shared some of our plans with them, they said we should prepare for a Black Swan event. I think he was having me on. I've thought about this and there's no way Natalie Portman and ballet could factor into our season. If he was being literal, I want to assure him that it’s next to impossible to get waffles into the ACC, so nobody is sneaking in a large black bird to throw on our ice. It's just not going to happen.
PPP: Did you have a hand in the decision to buy-out Mikhail Grabovski?
CF: A decision like that happens higher up the food chain and is way above my pay grade. I will say that it was clear, based on his D-zone starts and overall utilization, that Grabovski didn't fit in with Coach Carlyle’s approach to the game. I should add that I checked NHL.com this morning and Grabovski still hasn’t signed. If you look at UFAs that have signed, it’s pretty clear that Grabovski isn’t as good as say Nathan Gerbe or Rob Klinkhammer. Those two guys have SPCs, something Grabovski doesn't have. That's a pretty clear indication that they’re NHLers and Grabovski isn't. You can't argue with that, it's just fact.
PPP: Does your work overlap with Claude Loiselle and his duties managing the Leafs salary cap?
CF: No, but I recently showed Claude capgeek.com. He was blown away. That site has some sweet calculators.
PPP: If you run WOWYs, the data shows that Tyler Bozak has not been a great match with Kessel. Did you support his signing as a 1C?
CF: One of our proprietary stats, which I can’t speak about with too much detail, is Face Level Offensive Wardrobe (FLOW), which shows Bozak is among league leaders and definitely a top C. I don't want to suggest that our FLOW data is predictive, but I can tell you that every year around Halloween Bozak really hits his stride.
PPP: What are you most looking forward to this season?
CF: HBO’s 24/7 should shed a light on how hard everyone involved in this team works. From Liles being the first guy in the gym each morning to Carlyle listening to the audio of each and every game – he really breaks down that tape – to Mr. Nonis' work in media relations with Dreger. On a more personal level, I can't wait for the first 48 games. That's more than enough of a sample size to show where this team is headed. It's going to be wild.