(Brian Babineau, Getty Images)

(Brian Babineau, Getty Images)

I’m tired of being told that hits and blocked shots don’t matter. The best advanced stats bloggers generally refrain from saying it outright, but I’ve heard the sentiment too many times. I’m also tired because it’s hard to get sleep when you’ve got a newborn and a toddler, but that’s a side issue.

As I spent some time wondering why the idea of hits and blocked shots not mattering bothered me so much, I started to relate it to the shot quality debate. Many proponents of advanced stats will tell you that shot quality doesn’t matter and that shot quantity is far more important. It’s understandable why many traditionalist take issue with hockey analytics when they’re told that shot quality, hits, and blocked shots are unimportant when all three can play a vital role in the outcome of a game.

It’s clear to me that all three do, in fact, matter and that looking at hits and blocked shots in the same light as the shot quality versus shot quantity debate can shed some light on why.

We know that shot quality exists because we can see it on the ice. Some players are just better at shooting the puck than others and shots from certain areas on the ice are more likely to result in goals than shots from elsewhere. If Steven Stamkos is taking a one-timer from the slot, it’s a safer bet to hit the back of the net than a Zenon Konopka wrist shot from the boards. Of course, the low-percentage wrist shot from the boards could potentially deflect off a defender’s skate and go in — it’s a low-percentage shot, not a no-percentage shot — but it’s clear to observers that shot quality matters.

And yet, the most often-cited advanced statistics in hockey are Corsi and Fenwick, which measure shots for and against, including missed shots, and, for Corsi, blocked shots, ignoring shot quality altogether. Here’s where the misunderstanding comes in: proponents of advanced stats aren’t saying that shot quality doesn’t exist or that it doesn’t matter when it comes to the individual shot. It’s just that when you look at the big picture, shot quality more or less disappears.

The two elements of shot quality are location and skill. A shot taken from closer to the net — aka. a scoring chance — has a better opportunity to be a goal. Possession statistics like Corsi and Fenwick, however, have been shown to have a high correlation to scoring chances, so, since it’s easier to track the quantity of shots than it is to track scoring chances, it’s much easier to use Corsi and Fenwick.

As for skill, that generally comes out as shooting percentage, which fluctuates so wildly from season-to-season that it’s essentially useless in predicting future performance. Skill certainly plays a part in shooting percentage but so does luck. Players with abnormally high shooting percentages in one season are unlikely to repeat it the next. Corsi and Fenwick are far more useful as they’re far more repeatable.

Here’s an example: Patrik Berglund had the highest shooting percentage in the league last season among players who played all 48 games. His shooting percentage was 23.0%, scoring 17 goals. That brought his career average up to 13.0%. How much do you want to bet that his shooting percentage will be far closer to 13.0% than 23.0% next season?

The quantity of shots is what ends up mattering the most when it comes to predicting future success: Fenwick Close, which tallies shots for and against when the score is close, including missed shots but not blocked shots, is the most reliable of our current statistics as it correlates well with future success and is far more repeatable than something like shooting percentage.

So, while we see shot quality matter on the individual level, it’s shot quantity that ends up being more useful for analysis.

We can look at hits and blocked shots in a similar way. We know that a hit or blocked shot can have a major impact on the course of a game because we can see it on the ice. We see a hit on the forecheck that frees up the puck, leading to a scoring chance. We see a defender lay out to block a shot, taking away a good goalscoring opportunity. We see the positive effects of hits and blocked shots every time we watch a game.

What we’re really seeing are quality hits and blocked shots. It’s very easy to ascribe a higher value to those hits and blocked shots that seem to really matter, erasing from our memory the ones that had a negligible impact on the end result of the game. The same is true of shots: the ones that stick out in our memory are the quality shots, the scoring chances. We tend to forget the bad angle shot that was easily saved or the weak wrister from the point.

Quality hits and blocked shots obviously matter but, unlike shots, the quantity ends up being a negative. A high quantity of shots indicates that a team was in control of the puck, whereas a high quantity of hits and blocked shots means the other team had the puck, which, all things considered, is not ideal. In fact, hits and blocked shots correlate better with losing than they do with winning.

The virtue in a high quantity of hits is that maybe, just maybe, it’ll wear down the other team, even though that team is composed of players who have spent the vast majority of their lives getting hit by opposing hockey players. The virtue in a high quantity of blocked shots is that media members will praise you for your blue-collar effort and grit.

There’s one other virtue to both, however: there is the chance of throwing a quality hit or making a quality shot block at the right time.

The quality hit that creates a turnover leading to a scoring chance or that wipes out an opponent to prevent a scoring chance — that’s the whole point of throwing hits in the first place. The quality shot block that bails out a goaltender or, even better, starts a rush the other way that creates a scoring chance, makes the bruised shins worth it.

Hits and blocked shots, as individual events, can be very positive. A high quantity of them, however, is bad news.

The oddity comes in the emphasis: in order to get quality hits and blocked shots that lead to scoring chances and save goals, coaches emphasize quantity, at least for teams that focus on such things. Players are encouraged to finish every check and block shots at nearly every opportunity, in hopes that one of those hits will lead to a turnover or that one of those shot blocks will prevent a potential goal.

With shots, however, quality is emphasized more than quantity. The emphasis is on scoring chances, not just shots, excepting those times when a team goes into desperation mode and is encouraged to shoot from anywhere. The search for quality shots, however, inevitably leads to quantity over time for the better team. In attempting to create scoring chances, shots on goal, blocked shots, and missed shots will be created as a by-product.

In some ways quality hits and blocked shots are a by-product of emphasizing a high quantity of hits and blocked shots. A team that’s good at possessing the puck, however, won’t end up with many hits or blocked shots, even if quantity is emphasized.

The key to remember is that shot quality, hits, and blocked shots do matter, if only when it comes to the individual event on the ice. It’s only when those individual events become small pieces of data in a bigger picture that shot quantity becomes more important and hits and blocked shots become a net negative. The individual scoring chance, big hit, and blocked shot still have meaning and purpose, no matter what some may say.

Comments (31)

  1. “The individual scoring chance, big hit, and blocked shot still have meaning and purpose, no matter what some may say.”

    Nobody has ever said this, or anything like this. Of course players should be trying to get as many scoring changes, hits, and blocked shots as possible. Trying to. As you said, scoring chances have shown to largely be a by-product of more shots on goal; or vice versa depending on the narrative you want to follow. Negatively, teams with more blocked shots and hits tend to not be so good. It’s not that they’re bad teams because they block so many shots, or hit so much; it’s because they’re a bad team that they’re constantly without the puck and being forced to block more shots and make more hits.

    You’ve got your correlations and causations all screwed up here. The “advanced stats” guys you’re drubbing here aren’t so confused when then talk about such things.

    • Colin – I have to strongly disagree with you on whether there are people in the hockey blogosphere that discount hits in the way you mention. Watch what happens at say, certain sportsnation websites, if someone notes that the teams need to be more physical (or that certain players don’t hit enough).

    • The problem is that there are people who say the things I’m arguing against. Most advanced stats guys who know what they’re talking about don’t, you’re right. But some do and it’s incredibly frustrating.

      As for having my correlations and causations all screwed up, I think instead you’ve completely misunderstood what I said in this article. Obviously teams that block a lot of shots and throw a lot of hits are generally bad teams. I said as much in this article. If you missed that, then you didn’t read closely enough.

      I like advanced stats. I use them regularly and I’ve tried to add my own contributions to the conversation. Don’t think that I’m on the other side of this argument.

      • Toronto Maple Leafs were 2nd last in Fenwick Close and 1st in hits. We’ve been told, ad nauseum, that the Leafs are a really bad team, with a bad coach and will be worse this season, if we use Fenwick Close as the gold standard for successful teams.

        The LA Kings were 1st in Fenwick close—it’s safe to say they are a great team—and 2nd in hits.

        I’m confused. Please hold me.

  2. What’s the argument here? Those advanced stats measure correlation of events to winning. No one is telling hockey players not to hit or block shots because it will hurt their numbers.

    This is a rant but I don’t even know what you’re angry about.

    • I’m angry about those that discount the value of hits and blocked shots just because having a large quantity of them is a bad sign. If you haven’t encountered that sentiment, then feel free to ignore this post.

      • Here’s what you’re not getting:

        A) Hitting and blocking shots are good things at both the individual and team level.
        B) Having high team totals of hits and blocked shots is indicative of a poor possession team.
        C) Poor possession teams don’t win as often as good possession teams.

        That’s it. If you’re arguing that people think that B&C invalidate A, then you’re arguing to incredibly stupid people who aren’t worth the effort. You’re essentially pissing in the wind here.

        • You can identify ‘shot quality? Most of your reference here is to shot “location” not quality.

          As for “hit quality” or “blocked shot quality”. Find some evidence of it somewhere as a useful skill. Logically guys that throw more hits or block more shots should have more “quality” hits or blocked shots, and yet they don’t seem better at minimizing scoring chances against as a result.

          The problem with arguments like this when you use no data to support it is that you aren’t offering any evidence beyond anecdotal “eye-test” statements. You’re using the same logic that has led to the original belief this stuff is important that most evidence we have refutes.

          The issue isn’t that hits never matter… It’s that they are secondary and don’t generally lead to goals for or against. Sure a guy that is good who also hits and intimidates is nice to have on your team, but the important part of his game is being good at possession and scoring… The hitting part is just sort of icing on the cake.

  3. I would like to see hit stats kept with their zone: forward zone, neural zone and defensive zone hits. I bet a team outhitting in the forward and neutral zones is playing better than a team amassing hits in the defensive zone.
    Team A 25 total hits-15 fwd, 7 neutral, 3 d zone
    Team B 35 total hits-8 fwd, 5 neutral, 22 d-zone
    I would think Team A is controlling play and generating more chances

  4. Leading the team in blocked shots just means you never have the puck and are always defending. Understandable for players who play a lot of PK minutes, but a bad stat for anyone else.

  5. I’m with Colin and jmikelittle on this one. Reminds me of the old arguments made when advanced stats started to gain traction in baseball.

  6. Sure a blocked shot could be effective at stopping scoring chances but it could also take you out of the play (or out of the game). You could also create bad deflections. I think it creates value but not as much as some coaches believe (especially out at the point where the player has such a low success rate to start with).

  7. Everyone’s already said this, but I’m really not sure who you’re arguing with. You’ve built a bit of a weird straw man.

    • It’s not a strawman because it is an argument I’ve encountered. Again, most advanced stats guys who know what they’re talking about don’t say that hits and blocked shots are useless, so they’re not who I’m talking about. But there are some who take the idea that a large number of hits and blocked shots are a negative and discount the value of hits and blocked shots entirely.

      • Find me one example of someone saying that hits and blocked shots, at the individual level, are a bad thing.

        Large quantities of hits and blocked shots at the team level are indicative of a poor possession team; you can’t use hits and blocked shots at the individual level in an attempt to disprove that theorem.

        • I can certainly find examples of people sharing the sentiment that individual hits aren’t important because we can’t statistically show their importance. For instance, Steve Burch’s comment at 9:38 above, on this blog.

        • “in fact if we spent LESS time looking for hits or fights or hot blocks we’d probably IMPROVE”


          • Stephen Burtch ‏@SteveBurtch 27 Mar
            ok so the #Leafs lead the NHL in fights (useless), hits (uselessStephen Burtch ‏@SteveBurtch 27 Mar
            ok so the #Leafs lead the NHL in fights (useless), hits (useless) and blocked shots (basically meaningless)) and blocked shots (basically meaningless)

          • Stephen Burtch ‏@SteveBurtch 27 Aug
            @dmgoodhart NHL GM’s/Coaches/Fans like to think hitting/fighting/etc. is important… it’s not.

      • I see far more people talking about hits and grit being all-important than that discount them altogether. THE HAWKS AREN’T GONNA GET PAST THE KINGS AND THE BROONS, THEY’RE TOO SOFT, THEY NEED TO TRADE KANE FOR SOMEONE THAT HITS PEOPLE. Oh wait.

        • I can’t argue with you there, Lee. My issue is that some people do tend to discount hits entirely and it doesn’t do anyone any good to ignore the fact that such a sentiment exists online.

  8. Hits, as defined by current tracking, are recorded events in which player-contact results in a turnover…It kind of renders the notion “quality hits” a bit of a misnomer.

    ANY takeaway (whether by contact, or not) is a GOOD takeaway – The difference between one takeaway and the next is (perhaps) the zone in which it occurs, and the potential outcome it prevents (something we may never definitively know – Such as “the goal-against that never was”) or initiates (such as the hit that springs the odd-man rush for the game-winner.)

    We don’t begrudge Pavel Datsyuk’s consistent and repeated picking of his opponent’s pockets with his deft stick-checking by saying: “Oh, takeaways aren’t worth shit because his role as a ‘takeaway wizard’ is only made possible by reason of the Wings’ inability to maintain puck-possession.”

    A team can (obviously) have incredible possession-numbers, and STILL repeatedly strip the puck from the opposition during what scant time they manage to even get a stick on rubber.

    It’s kind of a moot point.

    Too often (for my tastes) those engaged in the compilation and analyses of advanced stats make the self-fulfilling error of justifying that which they engage in, and is measurable, and discount or dismiss that which is not (as though it played no role whatsoever.) Just because one hasn’t understood how to quantify an event or attribute doesn’t mean it’s influence is irrelevant or (worse) non-existent.

  9. August sucks eh?

  10. I think that a better stat than straight blocks would be %blocks… divide blocks by all the player’s negative Corsi events (shots + misses + team blocks) and see which guys stop the largest percentage of opposing attempts. That should, in theory, tell you more than just total blocks, which is biased in favor of guys who permit a ton of attempts in the first place.

    It will never be a full substitute for relative Corsi because great defenders are more than just guys who knock stuff down… you can simply never let the opponents into your end, you can break up their rushes and knock them off the puck before they ever get started. This is also where “eyes” are still superior, because there’s no way to measure “rushes frustrated” with any hope of accuracy or objectivity, not with the tools we have now.

    Hockey is just too fluid to easily divide into a series of discrete events like pitches, at bats, or plays from scrimmage. The next big breakthrough in hockey analytics is going to be the statistical equivalent of fluid dynamics – a way to break down trends in play that doesn’t rely on individual events as much, since there are only so many of them.

    • Good point here. You don’t want to be in a position to have to block shots but getting down and blocking a shot rather than forcing your goaltender to make a play helps the team. As a stat it’s kind of similar to saves. Goaltender X, who gives up more rebounds per game than the average goalie (generally a bad thing) is likely going to end up making more saves in a given game than the average goalie. A percentage system may be a smart statistical addition.

      But it still suffers from all of the same problems as individual Corsi: How much of it is team dependent and how much is a result of something the player himself did. Is the shot block opportunity created because a teammates made a turnover or because of something the player himself did? Does the player’s position or team system provide him more opportunities to block shots than another player?

    • I agree. I’ve actually written about percentage of shots blocked in the past: http://blogs.thescore.com/nhl/2012/01/06/who-is-actually-the-best-shot-blocker-in-the-nhl/

  11. Aren’t both quality hits and quality blocks are in a limited way already built into the Fenwick stat? If you make a quality hit and prevent a shot attempt, you just denied the other team a Fenwick event, which brings up your own personal Fenwick%. If you make a quality block, the same thing happens. Since blocked shots aren’t counted as Fenwick events, you again prevented one Fenwick event for the other team. You could hypothetically count quality blocks (on a line by line basis anyway) by subtracting Fenwick events against by Corsi events against.

    • Fair point but because hockey stats are so interconnected, most stats are already built into each in a limited way. For instance, if you allow less shots, your team should give up less goals against. I’d be pretty surprised, though, if someone said we should stop counting shots because we already count goals.

  12. Blocking shots is a good thing. Having high amounts of blocked shots means the other team is getting a lot of shots on your team, which is a bad thing. It is better to block a shot than to allow it through, but it’s best to not let the other team get the shot off in the first place, which is done by not letting them have the puck. It’s a hierarchy of sorts: no shot>blocked shot>shot on goal/missed shot.

    Hits are in a similar spot. It might be better to hit a player than to not hit them (I’m not sure whether this is even true, but let’s assume it is) but it’s better for them to not have the puck in the first place.

    It’s a similar thing with takeaways, but even more pronounced: on the individual level, a player who has a lot of takeaways (and few giveaways,) is probably really good (see: Jonathan Toews,) but it’s best for the other team to not have the puck in the first place. A high number of takeaways AND good possession numbers are almost certainly better than just the latter (and certainly better than the former.) It’s the exact opposite situation with giveaways.

    I believe you know all this, but I figure anyone who reads the comments might as well get an idea of why this article is addressing a strawman of sorts.

Leave a Reply

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