• Big Oilers Data III

    by  • May 1, 2013 • Hockey • 11 Comments

    This is part of a series looking for reasons for the Oilers Corsi% collapse in 2012-13 by examining things on a shift-by-shift basis. Part 1 can be found here. Part 2 can be found here.

    I’m going to move now from discussing Shot Attempts Allowed (SAA) to discussing Shot Attempts For (SAF). The observant and statistically literate reader will note that these are the constituent elements of Corsi – by Shot Attempts” I’m referring to shots, blocks and misses. In this post, I’m going to look into how the Oilers defencemen changed in terms of SAF on a shift-by-shift basis in 2011-12 and 2012-13.

    In 2012, 35 pitchers pitched enough innings to qualify for the American League ERA crown. The most likely to walk a batter was the Toronto Blue Jays’ Ricky Romero, who walked 12.7% of the batters that he faced. At the other end of the spectrum, Scott Diamond of the Minnesota Twins walked just 4.3% of the batters that he faced. Amongst hitters who qualified for the batting title, Adam Dunn walked in 16.9% of his plate appearances and Alexei Ramirez walked in just 2.6% of his plate appearances. This is the historical norm: hitters at the extremes in terms of walk more frequently and less frequently than pitchers at the edge of walk rate.

    Bill James had a theory about this which has always struck me as pretty good. His theory was that the wider spread for hitters meant that walking was more within the control of the batter than within the control of the pitcher. Plate discipline and ability to throw strikes are both skills; it’s just that the spread of talent on plate discipline is wider than that of the ability to throw strikes. Keep that idea in mind as we delve into the SAF numbers for defencemen. I’ll start with the percentage of shifts on which Oilers defencemen were on the ice for an SAF. First, defencemen who were on the team in 2011-12 and 2012-13.

    The graph is sorted from left to right by increase in 2012-13 from 2011-12. You’ll notice that this is a very tight spread – in 2011-12, the Oilers defencemen who played at least 100 shifts saw an SAF on a given shift between 36.1% and 42.8% of the time. This year was even tighter – it ranged between 38% and 41.2%. I’m not going to get into the forwards in this post, but I can tell you that the range with them was much broader – from 30.8% to 49.8% this year.

    Intuitively, that makes a certain amount of sense. It also suggests to me that, on the Oilers at least, the forwards are the ones who drive the SAF volume, not the defencemen. I suspect that this is a principle of more general application, subject to some exceptions. A guy like Erik Karlsson, for example, I suspect that he helps drive SAF numbers. Take a look at the fifteen Senators forwards who’ve played at least 240 minutes with him over the past two seasons.

    There looks to be a significant impact. Like I say, I doubt that that’s the norm but I bet if I ran this analysis for the Ottawa Senators, things would look a bit different. Here are the guys who played in 2011-12 or 2012-13 but not both.

    The Oilers leader in shifts with at least one SAF this year was, unsurprisingly, Justin Schultz, with the Oilers posting at least one SAF during 41.2% of his shifts. We might be seeing some of his influence on Nick Schultz’s numbers – Nick isn’t exactly known as an offensive player but he posted the second best number here for the Oilers and was up significantly over his number from last year.

    Let’s move on to looking at multi-SAF shifts as a percentage of all shifts on which there was an SAF. Again, I’m starting with the guys who were on the team in 2011-12 and 2012-13. (A word about Peckham that I assume is obvious but perhaps isn’t: he barely played in 2012-13. I’m including him as a returnee but he played only 91 5v5 shifts so his presence is more for the sake of a way of splitting up the D than a sign I’m hoping we can learn anything from looking at him other than if you’re a marginal performer, don’t come into camp out of shape.)

    Again, we’re talking about pretty tight ranges. In 2011-12, the range was from 36.1% to 42.8%. This year (excluding Peckham), it was from 32.4% to 37.7%. I’m not going to go into the forwards yet but the spread for them was much bigger last year (21.7% to 38.5%) and this year (23.6% to 39%). For me, when I see spreads like that, I’m inclined to think that the forwards are exerting far more control over these numbers than are the defencemen. When we use something like Corsi%, these things are baked into it; pulling everything apart and spreading it out this way lets us get a better sense of this.

    Defencemen who didn’t play for the Oilers in both 2011-12 and 2012-13:

    You will note that the Oilers were more likely to turn a shift with a SAF into a multi-SAF shift with Mark Fistric on the ice in 2012-13 than they were with Justin Schultz. If you asked me to guess, before I did this work, whether I thought defencemen would have more control over whether there was at least one SAF on a shift or whether a shift was a multi-SAF shift, I would have guessed the former. I would think that whether or not a shift becomes a multi-SAF shift turns more on the ability of the forwards to retain possession and generate multiple shots than it does the ability of the defencemen at the blue line. On the other side of the coin, a defenceman who can facilitate the movement of the puck up the ice seem to me to be a guy who’s more likely to have more shifts with an SAF than a guy who can’t. I’m spitballing here but, logically, there seems to me to be more of a scope for a defenceman to turn a non-SAF shift into one with at least a single SAF than there is for him to move it from a shift with one SAF to a multi-SAF shift.

    Finally, the percentage of all 5v5 shifts that were multi-SAF shifts. As explained in previous posts, I prefer the SAF/no-SAF and multi-SAF/SAF measures because I think that they do a better job of separating out what we’re looking for but I seem to have developed a format here, so I’ll stick with it.

    OK. What does all of this tell us? In some ways, this is the least interesting post in this series. (Trust me, the one on forwards and SAF-shifts is nuts.) On the other hand, it kind of suggests what you’d believe to be true if you watch a lot of hockey – the impact of most defencemen on the volume of SAF shifts and, particularly, the volume of multi-SAF shifts may not be that high and almost certainly isn’t near the level of importance of forwards in the creation of SAF and multi-SAF shifts. There’s some value in having data confirm things that you already believed going in – it should give you some more confidence when you hit a surprising finding that you’re on to something.

    Up next: Forwards and SAF and amazing facts.

    Email Tyler Dellow at tyler@mc79hockey.com


    11 Responses to Big Oilers Data III

    1. May 1, 2013 at

      Another excellent post, Tyler. Would it be fair to say that your research in this area is showing that forwards are significantly more important than defensemen? If they aren’t generally primary drivers of SAF (this is probably a helpful change in nomenclature incidentally) and they aren’t generally primary drivers of shifts with none or only one SAA, that only leaves a small sliver of daylight for them to have an impact on the game.

      • Tyler Dellow
        May 1, 2013 at

        I am starting to wonder outside of the really, truly elite defencemen, the difference making that they provide is pretty small. However, I’m a bit hesitant to say that all that definitively at this stage. I’d like to get a bit further into the SAA thing, start looking at some components of it. I’m not sure that the data for that exists presently.

    2. Tyler Dellow
      May 1, 2013 at

      To put it another way – I’d be interested to see what these four pairings of five would do against each other:



      Mike Brown-Brown-Brown

      Mike Brown-Brown-Brown

      I know who’d be best and worst. It’s the middle that’s curious.

      • Showerhead
        May 1, 2013 at

        I imagine you have a sense of the middle two as well but it will be about trying to find the degree to which forwards carry the play.

        *Makes popcorn, pours beer, waits patiently*

        • Bank Shot
          May 1, 2013 at

          I’m not sure those are fair comparisons. For one thing, Mike Brown level defencemen barely exist in the NHL. I can think of John Scott and that might be it. Basically every defenceman in the NHL takes a regular shift.

          Also, there are three forwards versus two defencemen. Consequently, forwards as a group are going to have more of an impact. There are more of them on the ice.

          Cancel out the extra player advantage, and throw Mike Brown equivalents to the Crosby clones:



          Can Brown-Crosby-Brown hold their own in that circumstance?

          • Bank Shot
            May 1, 2013 at

            Rather, can Lidstrom-Lidstrom hold their own against the Crosby Clones in the second scenario?

    3. Trenton L.
      May 1, 2013 at

      I don’t think you should be so quick to write off individual D-man contribution to SAA.

      Bruce McCurdy recently did an article on Whitney’s poor performance and showed on player usage chart that Whitney, Potter and Fistric all essentially had the same OZS% and QoC.


      So with that in mind look at how these guys performed on your SAA analysis yesterday. For % shifts with at least one SAA you show Potter ~44%, Fistric~44.3% and Whitney ~48.5%.

      If we assume each shift is 40seconds and about 10.5 mins non 5 vs. 5 time per game you come to the average d-man playing ~24 shifts per game. Times 82 games is 1968 shifts in a season times each players % of shifts with at least one SAA gives respectively 865, 872, and 954 SAA events for the three amigos.

      So Whitney, under similar circumstances, would be on for a pace of an additional 84 shots against over the course of a full season. Lets assume an 8% avg shooter means that Whitney has cost the Oilers 7 goals against in comparison to using Fistric and Potter.

      I agree the high variation YoY in the at least on SAA suggests the bulk of it is out of the d-man’s control but this nice little example shows where dman quality can make a difference.

      • Tyler Dellow
        May 1, 2013 at

        Trenton -

        Yeah, I think defencemen have some impact on the SAA shift number. I guess what I’m saying is more that I suspect that there’s kind of a baseline number of shifts you’re going to have on which you might face an SAA and that that’s probably driven by the forwards. The question is really the size of the margin – how well you play things coming up the ice, how well you break up the play once it gets to the defensive zone.

        I don’t have any difficulty with the argument that Whitney was the weakest Oiler defender at this.

        • Triumph
          May 2, 2013 at

          Could also the kind of SAA be controlled by the defensemen? I’m thinking here of a 2 on 2 – I feel like on most 2 on 2s, a shot attempt is made, but if the play is well-defended, it turns into a block or a miss. Most shot blocks I’d consider a win for the defense, and Sunny Mehta showed long ago that shot blocking is a skill. I’d be curious if there was a reasonable gap between the best and worst shot blocking defenseman – i.e. if there are D whose actual shots to shot blocks ratio is substantially better than other D. Of course, since shot blocks are counted weird the league over we’d need road shot block stats…

    4. Sliderule
      May 2, 2013 at

      You and MacT are onto the problem with the oilers.

      Forwards who have no idea of what to do away from the puck are the biggest problem with the oil.

      The forwards have speed and can be seen making a solid back check but in their own zone they don’t cover anyone especially in the slot.

    5. csmccart
      May 8, 2013 at

      I think it would be interesting to see what percentage of shifts each DEF played with each FWD . I think that may provide insight into the “forward factor” in the statistics. It would also be interesting to see SAA and SAF in terms of successful breakout attempts.
      On another note, I happen to think a big part of the defensive game is the ability to generate turnovers and takeaways. I agree that size and physicality are do not necessarily characteristics that generate a strong defensive game, take Nuge for example. I feel like this might provide insight into which DEF contribute to regaining possession.
      Overall, I find these stats very interesting. Especially when it seems that the oilers tend to struggle against teams who employ a cycle game.

    Leave a Reply

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