• Value In The Margins: Rangers and Shotblocking

    by  • June 22, 2013 • Hockey • 10 Comments

    I sat in on Marek vs. Wyshynski along with The Score’s Ellen Etchingham. (Yahoo’s Greg Wyshynski made occasional appearances by way of a spotty internet connection from Chicago.) It’s the second time I’ve done it and it’s an absolute blast – just three or four people having fun talking about hockey. Plus you get to be in a studio which is kind of neat. And, if you’re on a creaky chair and a fidgety sort of guy but sound fairly relaxed and calm when you talk, the internet will just figure that Ellen’s nervous and blame her. So that’s cool.

    One of the things that we talked about was the idea that the Bruins are a big, bad team. I’m kind of skeptical. When I looked at hitting and getting out-hit earlier this year, the B’s weren’t a team that really tended to out-hit their opposition last year – if you don’t want to get hit a lot, play a really good team that has the puck more than you. That’s not precisely what Jeff was talking about – he was talking more about the fact that if you play a team like the Bruins you always know, for example, if you’re a defenceman going back for a puck that you’re always going to get pasted – but it’s relevant, I think. Greg brought this up too – when he talks to players on opposing teams preparing for a series against Boston, they’ll mention that they know they’re going to get hit.

    We don’t have really solid hitting data that would enable me to evaluate Jeff’s point or Greg’s thought, as relayed by the players – with location data on hits, I could look at hits on defencemen that occurred below the goal line, for example. My objection’s a little more broadly based though – I think hockey’s more homogenous than people realize. Does Milan Lucic smoke defencemen when he’s forechecking on a dump-in? Sure. So do lots of guys.

    To me, the issue really is the marginal difference between the teams that are most and least likely to hit. When we talk about this stuff, that’s really what we’re interested in. Nobody would say “The Bruins are a team known for using hockey helmets” because everyone wears helmets. If we’re going to talk about some trait being present or lacking in a team, it should be something that they really do significantly more or less than anyone else. Talking about a team having some defining trait that all teams have is silly.

    This came to mind today with the news that John Totorella is probably going to be the new coach of the Vancouver Canucks, which prompted a round of comment on Twitter about whether Vancouver has a team that wants to block shots. Tortorella has become indelibly linked with shot blocking by virtue of his time in New York. Leaving the question of whether people are forgetting about Tortorella’s Tampa Bay teams to one side for the time being, there still seems to me to be a critical question here: is shot blocking really a defining trait of Tortorella’s New York Rangers?

    I wrote about this at some length last year, but from a different angle than I’m going to take now. A boring word first about a technical point. One of the great difficulties from which many people who like using data to make points suffer is an inability to choose the data that most tightly relates to the point that they’re making. There’s a kind of intuitive sense that if you want to talk about shot blocks, you just count shot blocks. The NHL goes a step further and presents the data all mashed together on its website: EV/PP/PK shot blocks.

    I think that this is wrong. To me, when we’re talking about how much a team blocks shots, we need to have a sensible sort of measure of this. The first step is talking about specific game states; in this case, 5v5. The next is recognizing that we need to talk about shot blocks by reference to the volume of opportunities that a team has to block shots. Consider two teams: Team A has a 60% Corsi% and blocks 50% of the shot attempts against it and Team B has a 40% Corsi% and blocks 35% of the shot attempts against it.

    Team A is going to block 20 shots out of every 100 shot attempts in games that it plays. Team B is going to block 21 shots out of every 100 shot attempts in games that it plays. I’d still call Team A more of a shot blocking team though – they’re more likely to block a shot, once one has been attempted. If we want to talk about shot blocking, let’s talk about shot blocking. Getting a lot of shots fired at you and piling up a lot of shot blocks simply by virtue of that fact does not, in and of itself, make you a shotblocking team. A lot of people don’t seem to get this.

    That blathering aside, let’s talk about the data. I’ve put together a table with the percentage of 5v5 shot attempts that were blocked overall, at home and on the road for the 2012-13 season. As you’ll see, most teams have fairly similar numbers at home and on the road but it’s probably fair to wonder if the scorer in San Jose just LOVES counting things and isn’t going to be stopped by technicalities like “nothing happened” and whether the guys in Buffalo and New Jersey just don’t care. Outside of that, teams tend to be pretty similar, home and away.

    I’ve sorted it by the road percentage because I’m more comfortable that the road percentage provides a measure of what a team actually does – you’re basically getting a mix of “objective” opinions if you approach it this way. Sure enough, the Rangers are at the top of the list. The list of teams at the bottom is awfully interesting to me – it’s kind of a mix of teams that I think of as smart (LA, Vancouver, Phoenix, Ottawa, San Jose and Detroit) and teams with goalies who have big reputations (LA, Dallas, Vancouver, Phoenix and Ottawa), goalies who you might figure don’t need the help. Hmm.

    It does seem fair to call the Rangers a shot blocking team. They’re top of the pops. It’s kind of funny that they have such a big shot blocking number with an excellent goalie – I can understand the Blues launching themselves in front of pucks far more easily, much like the pre-Roloson 2005-06 Oilers. As I said above though, what we’re really interested in is the difference between the Rangers and, say, the average team and then maybe the difference between the Rangers and a team that’s bottom of the barrel in terms of shot blocking – the Kings.

    Before I get into the arithmetic, a final point: not all shots that are blocked are going to be shots on goal. Sometimes you’ve got guys hurling themselves in front of shots that were going to miss the net anyway. For that reason, I think a better number to use here to evaluate things is the percentage of shot attempts that turned into shots on goal. Blocking a shot that was going to miss the net is just needless pain (subject to something I’ll touch on below).

    The table at left summarizes the percentage of shot attempts against that turned into shots on goal. Again, it’s sorted by the road data. Unsurprisingly, it tracks the shot blocked data pretty closely – teams that block a lot of shots like the Rangers aren’t going to allow a high percentage of shots on goal, although adding missed shots into the equation blurs things slightly.

    There were 1075 5v5 shot attempts against the Rangers this year on the road. As they were excellent at preventing those from turning into shots on goal, just 48.9% of them became shots on goal – 526 shots against. The average team allowed 53% of shot attempts to become shots on goal. That would be 570 shots on goal. There were 24 road games, so the difference is about 1.8 shots on goal prevented per game at 5v5 between the Rangers and an average NHL team.

    What if we look at a team at the other end of the scale? Take Los Angeles, the team which was least likely to block a shot. 56% of 5v5 shot attempts against the Kings turned into shots on goal. If the Rangers had allowed 56% of shot attempts to become shots on goal, they would have allowed 602 shots on goal. It’s a difference 3.2 shots per game.

    Over the course of an 82 game season therefore, the Rangers going about their business as they do is going to save them 150 shots relative to the average NHL team and about 260 shots relative to a really shot block averse team like the Kings. If you convert that into goals at the rate of 8%, you come up with the Rangers’ shot blocking saving them twelve goals against per 82 relative to an average team and 20 goals against relative to a team like the Kings. Those are fairly substantial savings. We’re talking about two wins more than an average shot blocking team and three more than the Kings. Buying a player who would make that kind of a difference isn’t cheap. In a capped league, that could be an incredible benefit if you can get your players to buy into it.

    However, I think that those are kind of out of outer limit estimates. To start with, while I’ve assumed that 8% of shots are going to go in, I have a strong suspicion that that’s not true of shots that get blocked. Just from watching hockey, it feels like those shots are more likely to be shots that are taken from the blue line – those shots probably have a less than 8% chance of going in. Second, perversely, a team that is dedicated to trying to block shots is probably going to re-direct more shots past its own goalie than a team that doesn’t. That’s not going to show up in the data that we currently have available. Third, there’s a cost to being a shot blocking team in terms of injuries to valuable players. I threw this question out on Twitter and got a list of Rangers who’ve been hurt blocking shots.

    Here’s Ryan Callahan breaking his leg:

    Here’s Ryan Callahan breaking his wrist:

    Chris Drury, Brandon Dubinsky, Mats Zuccarello and Marc Staal have all suffered injuries blocking shots too. In addition, it was suggested to me that Dan Girardi is kind of perpetually banged up. There’s a cost in wins to all of these injuries and, while some of them would have happened anyway on teams that don’t make shot blocking a priority, there’s a risk every time a skater steps in front of a shot and the more your skaters do it, the more likely that risk is going to come to pass. The injuries slice away some of the value from all of the shot blocking.

    There’s another angle to this as well, more difficult to quantify with the data available but certainly quantifiable by a team that put the effort into doing so. We talk about (and in coming out with an outer limit, I’ve assumed) shot blocks as if it’s a shot that doesn’t happen. There would have been a shot, a guy stepped in front of it, the shot didn’t happen.

    What happens next though? A thought experiment: what if every time you blocked a shot, the other team recovered the puck and got a shot on goal from that exact spot two seconds later? If that were the case, shot blocking would be entirely useless. “Here: try and break one of my bones and then, after you’ve maybe done that, take a shot at the goalie.”

    I don’t think that’s the case. I do think that if we’re trying to evaluate shot blocking as a talent, we need to understand how it changes the game. I suspect that it may well be worth less than we’d intuitively think. At least sometimes, the team just recovers the puck and generates another shot. When I look at the teams that don’t block shots, I’m suspicious: they’re generally pretty smart teams, teams that think about hockey and the fact that they aren’t doing it to near the extent of some other teams raises some red flags for me.

    What does all of this mean? Well, before we label teams as having some sort of defining characteristic, I think it’s worth it to try and determine whether it actually exists. In the case of the Rangers and shotblocking, I believe it does. I’m less sold on the Bruins and hitting, although I tend to believe Greg when he says that players certainly do believe it does. The rest of this is more about how we should think about and evaluate whether such things matter. I’m not at all sold that, once you took everything into account, the Rangers’ shotblocking’s all that significant.

    Email Tyler Dellow at tyler@mc79hockey.com

    About

    10 Responses to Value In The Margins: Rangers and Shotblocking

    1. Triumph
      June 22, 2013 at

      I know that Martin Brodeur and the Devils generally are against guys leaving their skates to block shots; I’m a little surprised to see them this high up in road blocks in light of that. I think years ago I thought being anti shot blocks was attributable to Brodeur’s vanity (why I thought this I can’t figure out, but I also can’t figure out why Brodeur was getting 77 games a year when the Devils were winning division titles). In light of data like this, it seems generally shrewd.

      I’d be curious about the data over the last few years – I don’t remember the Rangers shot-blocking becoming a media narrative until they were horrible territorially and were leading the division last year.

    2. Pierce Cunneen
      June 23, 2013 at

      I have done some research and found that shot blocking does not have a positive or negative effect (not accounting for injuries). Link to that research can be found here: http://www.broadstreethockey.com/2013/6/11/4419850/the-value-of-blocking-shots

      What I found was that blocking shots at a higher rate has a decent correlation with giving up Corsi against at a higher rate, but no correlation at all with Fenwick against. So what we are seeing here is that teams that are good at blocking shots give up more Corsi events against, but not Fenwick events against.

      So blocking shots at a higher rate is not having an effect on the number of shots that get past the defenders and either hit the net or miss the net. So shot blocking is having neither a positive nor a negative impact.

      I then looked at Fenwick shooting percentage, to see if blocking shots at a higher rate had any effect on how likely shots that got past the defenders were to go in. Again, no correlation.

      So when not considering injuries, we are seeing neither a positive nor negative impact of blocking shots at a higher rate, but with the increased risk of injury, I would say shot blocking is a net negative.

      The link at the beginning of this reply explains my research a little better.

    3. Pingback: Spectors Hockey | NHL Postseason Blog Beat – June 23, 2013.

    4. Mr DeBakey
      June 24, 2013 at

      One game this past season the Oilers Sportsnet play-by-play guy mentioned, after an Oiler had blocked a shot, “The Oilers are 4th overall in the NHL shot-blocking on the road.” [He’s actually quite good at digging up little factoids and seasoning his broadcast with them.]

      He presented this fact like it was a good thing, but I had my doubts. I was thinking it just meant the other team always had the puck in the Oilers end on the road.

      But after moving things around on a spreadsheet, it seems there is no correlation between blocked shots and position in the standings. Some good teams block lots of shots, and so do some not-so-good.

      The Oilers ended up 9th overall. I assume that once a play-off spot disappeared over the horizon, “No Gain, No Pain” kicked in.

    5. June 24, 2013 at

      “To start with, while I’ve assumed that 8% of shots are going to go in, I have a strong suspicion that that’s not true of shots that get blocked. Just from watching hockey, it feels like those shots are more likely to be shots that are taken from the blue line – those shots probably have a less than 8% chance of going in.”

      Cam Charron showed some evidence for that (not proof, but strong circumstantial evidence): http://nhlnumbers.com/2013/5/29/lets-talk-about-shot-blocking

    6. Tom Benjamin
      June 24, 2013 at

      I think the idea that the Rangers should be defined as a shot blocking team is another example where the very collection of the statistic makes the act seem far more important than it is. I’d define the Rangers as a team that defensively collapses to the net (as opposed to aggressively pursuing the puck in their own end.) The objective is to keep the puck out of the slot and prime scoring areas.

      Tortorella isn’t trying to get his players to block shots. He is designing his defensive zone coverage in such a way that lots of blocked shots are almost inevitable outcome of the strategy.

      However, I think that those are kind of out of outer limit estimates.

      Besides the three points you made, Tyler, there is an offensive cost to the Tortorella (Boston, St. Louis etc) strategy. Its a lot harder to get the puck moving quickly the other way when the defense has five players clumped in front of the goalie. Goals saved by the strategy is only half the equation. What about the goals lost because there is a slower transition to offense when the puck is recovered?

      On the other issue, the big bad Bruins is a valid description but I don’t think hit data will ever show it up. It is a strategy based on two assumptions. One, the rederees can’t call everything, and two, the officials almost invariably even up the penalties. Like the Flyers of the 70′s the name of the game is physical intimidation by always pushing the envelope with hacks, whacks, cross checks and face-washing.

      If the refs call it tight, the opponent had better score on the early power plays because the Bruins won’t change their game. Eventually the refs start letting all but the most egregious fouls go. The Bruins are advantaged because they get away with a lot more fouls. When opponents start retaliating there are fights the Bruin usually wins or the Bruins start getting their turn on the PP. The best part about the strategy is that you can usually force a more skilled team to try to play the Bruins style.

    7. Bruce McCurdy
      June 25, 2013 at

      not all shots that are blocked are going to be shots on goal. Sometimes you’ve got guys hurling themselves in front of shots that were going to miss the net anyway

      One of many great points in this piece, however for this one I would offer the counterpoint that often a player filing the shooting lane forces wide shots for which the defender gets no credit. I’d be hard-pressed to guess which happens more often — missed shots that were blocked, or would-be blocks that forced misses — but as a first order guess I’d suggest they cancel each other out to a large degree.

      Second, perversely, a team that is dedicated to trying to block shots is probably going to re-direct more shots past its own goalie than a team that doesn’t.

      Yes sirree bob, those shots that do make it through the thicket are way more likely to be tipped or screened. The success rate of that smaller number of shots should be a little higher, one would surmise. Again, something of a trade-off.

      What happens next though?

      Aye, that is the question. One of the things that is truly appealing about some of these new lines of study is the sequencing of events. You’ve been doing a lot of this sort of thing, Tyler, as have those doing Zone Entries and the like. What happens next?? A better idea of where an event “fits” in the flow of play might give a better idea of its actual value.

      In the case of blocked shots, I wonder what might be gleaned if you and your mad number-crunching skills were to look at (all) the PxP event that follows each shot block: What % of the next events are still in the same end zone as where the block occurred? What % show the original shooting team back in possession of the puck (a witches brew of Fenwick events and giveaways credited to that team, or a hit, takeaway, penalty, or second shot block for the defensive team, or more arguably, anything that results in a whistle at that end)? What is the average interval between blocked shot & next event, & is it shorter than typical interval between two random events? Is it any different for Rangers than other teams? Probably a whole lot of noise & 50/50 stuff, but I wonder if there might be some flow of play patterns that could be teased out of the big data.

      Lots of food for thought here.

    8. Tyler Dellow
      June 25, 2013 at

      I’d be hard-pressed to guess which happens more often — missed shots that were blocked, or would-be blocks that forced misses — but as a first order guess I’d suggest they cancel each other out to a large degree.

      Yeah, that’s a fair point. At the same time, the second table deals with percentage of shots that hit the goal. I’m implicitly taking that into account I think – to the extent that the Rangers force shots wide, it’s included in the analysis.

      In the case of blocked shots, I wonder what might be gleaned if you and your mad number-crunching skills were to look at (all) the PxP event that follows each shot block:

      Yeah, that’s kind of the obvious thing to do, isn’t it? ;)

      • Bruce McCurdy
        June 25, 2013 at

        the second table deals with percentage of shots that hit the goal

        Agree, this seems like a case where shots on goal, Fenwick and Corsi are each telling us slightly different things, but each has value.

        Ultimately, the ideal is to find the link between whatever type/sequence of events and “Goal”. Pretty complex task, I reckon.

        Yeah, that’s kind of the obvious thing to do, isn’t it? ;)

        I’ll look forward to your next post(s)!

    9. Pingback: Devils in the Details | All New Jersey Devils News

    Leave a Reply

    Your email address will not be published. Required fields are marked *