• A Tale of Two Power Plays

    by  • December 30, 2013 • Hockey • 5 Comments

    There are events that become part of the fabric of the discussion of a team. They’re signposts, markers for when something changed. Dallas Eakins’ decision to wipe out the five forward power play in the 25th game of the season, after they gave up a shorthanded goal to Chicago looks like it might stand as one of those this season. Patrick Sharp broke his stick trying to clear a puck, the puck handcuffed Yakupov because it had no speed, Jonathan Toews jumped on Yak who was in an awkward position and it was 1-0 Chicago.

    The clucking of the hens in the press box was vindicated by history’s first shorthanded goal and the Oilers abandoned the five forward power play. Fast forward 17 games and Andrew Ference is a regular sight on the PP and things are awful.

    If you split the Oilers season into two parts: the first 24 games and then from the Chicago game forward, the power play looks like it belongs to two different teams. I’ve put together the data as both totals and rates:

    I tend to think of things in rates, although I know that the ranges maybe aren’t widely enough known yet that they can be presented without some discussion about context. That shot rate for the Oilers through 24 games was excellent: 60 S/60 at 5v4 is a tough, tough thing to do. That should be what the Oilers are looking to become – a team that puts up that number over and over and over. It was, however, a little bit misleading. The Oilers were only taking 97.8 SAF/60. That’s still an improvement – their previous high in the six years of data was 82.9 SAF/60, so that’s an 18% increase or so. It looked like a serious step forward.

    I mention this because the shot number was a bit soft. Through 24 games, the Oilers were turning 61.5% of shot attempts into shots. Not a lot of blocked shots or misses. That was something we could have expected to kind of ease off. Over the course of a year, the vast majority of teams will be between 57% and 50% or so, in terms of the rate at which they turn shot attempts into shots at 5v4. The Oilers have kind of crashed hard here – they’re at 49.7%.

    The thing is – it’s not just the rate at which shot attempts turn into shots normalizing. The shot attempt rate is also off dramatically. They’ve fallen from 97.8 SAF/60 through 24 games to 81.8 SAF/60 over the last 17 games. The old Oilers’ PP that doesn’t generate any shot attempts is back! This is not a good thing because great PPs over time are the ones that generate a ton of shot attempt volume at 5v4. Shooting percentage comes and goes; a team that keeps generating shots can protect itself from that to a certain degree.

    I get interested in the hows and the whys of these things, so we’ll talk about it over the course of the week. I’ve re-watched the Oilers PP this year to generate data relating to how they enter the offensive zone and how much time they spend there. I can blend this with the NHL’s PBP data and do some cool stuff. I don’t like to make these posts too dense so I’m just going to do the zone entry/zone time data in this one and then leave it there.

    You’ll note from the table above that the Oilers played virtually identical amounts of 5v4 TOI in the first 25 and last 17 games: 114.7 minutes and 112.2 minutes. I classify zone entries into seven different types: OZ faceoff wins, OZ faceoff losses, successful carries, unsuccessful carries, successful dump-ins, unsuccessful dump-ins and miscellaneous. Miscellaneous includes zone time where, for example, the other team just pulls the puck into their own end or the Oilers are in their end when a PP starts due to a penalty expiring.

    Things are actually pretty similar, in terms of success, when comparing the good part of the season and the ugly part. There have been fewer offensive zone faceoffs in the ugly part, presumably because there have been fewer shots, which means fewer goalies freezing pucks. That’s reflected in the non-faceoff entries as well – 221 in the good part of the year, 246 in the bad part. I suspect that what that tells us is that the Oilers are gaining the zone and then seeing the puck cleared rather than generating a shot which leads to a faceoff.

    If we focus just on the carries and dump-ins, we see that things are very similar. 79.9% of the Oilers’ attempted neutral zone entries in the good part of the season are carries. 82.8% of the attempted neutral zone entries in the bad part are carries. Virtually identical. They succeeded on 91.8% of their carries during the good part of the year and 89% of their carries during the bad part of the year.

    Two points here. First, during the Oilers-Flyers game on Saturday night, Twitter’s @dawgbone98 and I were kicking around the concept of zone entries. I was in the middle of this project at the time and, to my eye, teams had been kind of sitting on the Oilers preferred zone entry of late. The Oilers love to drive the offensive zone with the puck carrier five or ten feet off the boards. The puck carrier then kicks it out to a stationary player on the boards. In theory, this results in the stationary player having a few seconds to get the puck moving and permit the team to set up.

    In practice, it looked to me like teams have been sitting on it lately, waiting for the Oilers to do it and then cutting it off, either before they hit the blue line or just as they did. I thought that I noticed, watching the Calgary and Philadelphia games, that the Oilers were dumping the puck in more frequently at 5v4. I don’t trust my eyes – all of us have brains that are constantly trying to fool us – but the data seems to back that up. I’d guess that someone on the coaching staff was watching some games over Christmas and figured that the Oilers needed to start making teams pay for cheating on the blue line and anticipating that the puck would be carried in.

    The second point is a bit more nuanced. Hockey analytics people love controlled zone entries. They tend to lead to more shots. The thing is…that “truth” depends on the circumstances in which it was established continuing to exist. If teams suddenly start stringing four men across the blue line and forcing turnovers, it may be that dumping more pucks in becomes the smarter play. If you’ve got four opposing players flatfooted or moving slowly on the blue line and you can dump it in with your guys attacking with speed, you’re going to have a much better chance of recovering the puck.

    I’ve never published this for some reason, but I did some detailed tracking of the Sharks and Oilers head to head during the Todd McLellan era at 5v4. The Sharks are, of course, basically the 5v4 gold standard over this period. The Oilers have attempted zone entries by way of carrying the puck 77.7% of the time against San Jose; San Jose has done so 71% of the time against Edmonton. Edmonton’s a little above that – about 80% – but nothing worth getting too worked up about.

    This isn’t to say that that’s necessarily a bad thing. Edmonton’s success rate at gaining the blue with possession by carrying the puck this year at 5v4 has been phenomenal. The Oilers are clicking along at about 90% in both portions of the season under review. The Sharks were at 76.6% in the games I reviewed. Edmonton was at around 80%. I don’t think you can reasonably make an argument that the Oilers have struggled in terms of their success rate at gaining the offensive zone this year. That said, I’m also of the view that there’s nothing wrong with a little game theory and pushing opponents back with dump-ins on occasion. It would be interesting to go back and look at those Shark games and see if they were more likely to dump it in during blowouts.

    All told, it’s hard to find a difference between the good and bad parts of the Oilers at 5v4 here and, given the comparison with the San Jose data, it’s hard to conclude that this has been where the problem lies. I’ll cut this one off here. Up next: zone time and time between zone entry attempts.

    Email Tyler Dellow at tyler@mc79hockey.com

    About

    5 Responses to A Tale of Two Power Plays

    1. woodguy
      December 30, 2013 at

      1) Thanks for kicking the shit out of my narrative that their successful zone entries were declining.

      2)In terms of SAF that turn into SA, do you have SJS’s rate from when you were examining it?

      You mention that they were due to regress to 57-50%. I’d imagine a PP with moving players and fast passes would tend to end up on the high side.

      Would be interesting to see where SJS ends up due to their historically excellent PP and also where WAS ends up as their Ovechkin Overload strategy seems to be producing ludicrous SH%

      As an aside, how much of WSH 5v4 SH% do you attribute to the PP design that Oates brought in?

      They went from middle/lower end to lead the NHL last two years.

      Mostly variance or is getting Ovechkin the puck in space driving most of it?

      Awesome post, look forward to the next bit.

    2. Bruce McCurdy
      December 31, 2013 at

      The thing is…that “truth” depends on the circumstances in which it was established continuing to exist.

      This is an outstanding point in an excellent post. Why change what’s working? Because it doesn’t work in a vacuum, and other teams adapt. So you must adapt as well.

    3. chris
      December 31, 2013 at

      Excellent piece. I felt like the oil have been trying to pass the puck into the net on the PP for a couple months and this supports it. Funny how spector called one of our PP goals under the bad PP regime a ‘repeatable play’ recently, as I feel like passing the puck into the net is less repeatable than aiming for a 60/60 PP shor rate and scoring consistently that way.

      Encouraging entry performance though.

    4. godot10
      December 31, 2013 at

      So, in other words, the power play disaster is really a coaching failure, since it takes the coach 20 games to adjust to a change in defensive tactics by the opposition on the power play.

      A good coach can’t wait 20 games for something to become obvious in the statistics.

      In the AHL with NHL-cailbre goaltending and a stacked roster (Eakins’ Marlies team), maybe you could wait 20 games to make an adjustment, but doing that in the NHL with an undermanned roster, well you lose the season in 20 games.

    5. Tom Benjamin
      December 31, 2013 at

      Hockey analytics people love controlled zone entries. They tend to lead to more shots. The thing is…that “truth” depends on the circumstances in which it was established continuing to exist. If teams suddenly start stringing four men across the blue line and forcing turnovers, it may be that dumping more pucks in becomes the smarter play. If you’ve got four opposing players flatfooted or moving slowly on the blue line and you can dump it in with your guys attacking with speed, you’re going to have a much better chance of recovering the puck.

      I don’t think this is a coaching choice. If the defense is strung across the blueline, the player with the puck does not have a choice – the correct play is to fire it in. Obviously it is best if the team can step over the blueline with the puck, and that’s the first choice of every team and every player. Sometimes the defense won’t let the offense do that.

      What other choice does the puck carrier have? Pass to a covered man? Try a deke? In fact all the forwards will read the defense, realize the puck is going to get dumped in and get on the horses. If the puck carrier does try a deke, the best result is usually an offside because the wingers get fooled.
      The PP has to get the puck up the ice fast enough so the defense doesn’t get set up across the blueline to block the easy entry.

      I look forward to the next post. I’m going to guess that the biggest difference between the good power play and the bad one will be the time between entry attempts.

    Leave a Reply

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