Unconventional Wisdom - Career HART

August is the most hockey devoid month of the long hot summer... June has the playoffs and the Draft, July has free agency and the Prospect Development Camp, and even September has the Rookie Tournament and Training Camp. But here in August we struggle to find hockey related information to discuss. If you want to talk baseball, Natasha can give you a rundown of the Little League World Series, which features her local Taney Dragons and their star pitcher, Mo'ne Davis, who just so happens to be a girl. But if you want to discuss Pittsburgh Penguins hockey, there just isn't an awful lot going on right now.

So in an effort to find something hockey related to pass the time in this long boring month, I am once again taking a look at fancystats. This time around we will be looking at career numbers for Hockey Analysis Rating, which as one of the advanced stats career numbers only go back as far as 2007-08. There are a couple of things to keep in mind with these results. The first is Drazenovic, his numbers only include the most recent 3 seasons as he still occupies two separate entries, one as Nick and one as Nickolas, so his 2010-11 numbers from his time in St. Louis are not included, although suffice to say they suggest his numbers would be even worse.

The other issue is that the numbers happen to be somewhat heavily influenced by sample size, so anybody that has played less than 750 minutes might be somewhat suspect, and those that played less than 500 minutes are quite questionable. So when looking at the numbers I let those that played less than 500 minutes get cut off if they fell outside the bounds of the regular players, and placed their names in red so that you can be aware of who they are. I also put those who played under 750 minutes in blue, but they all fell within the bounds of regular players, so their results may actually be fairly accurate.

Our players who have under 500 total career 5-on-5 or Close minutes are Samuelsson and Dumoulin on the blue line as well as Payerl, Drazenovic, Sill, Leblond, and Megna up front. In addition, Bennett is under 750 5-on-5 and under 500 Close, so there is a chance his numbers may be a little inaccurate. Lastly we have 4 more D who have under 750 Close minutes, they are Chorney, Bortuzzo, Despres, and Maatta, which means they are most likely accurate results but still something to keep an eye on.

The numbers utilized come from David Johnson's HockeyAnalysis, featuring the metrics he created that take the numbers we are used to, in this case Goals and Fenwick differentials, and weight them based on that player's teammates and opponents faced.

I set them up like the classic bubble charts. The X-axis based on Zone Start percent, those who get more sheltered/beneficial O-zone starts are further right whereas those who have tougher D-zone starts are further to the left. With most of the numbers, but especially possession, one can expect to have the advantage when taking most of your starts in the O-zone and to have a tendency to get outworked when you take most of your starts in the D-zone. The Y-axis is based on QoC, but which QoC is dependent upon which metric we are attempting to view. The higher the QoC, the better the opponents you face are at that particular metric.

And lastly the bubble sizes are based on the player's individual "relative" metric, dependent upon what that particular chart is examining. HARO is based on Goals For or Fenwick For, with players who are higher up tending to not only have higher individual numbers but also better in comparison to their teammates and the competition they face. HARD is based on Goals Against and Fenwick Against, with players who are higher up not only tending to have lower individual numbers but also better in comparison to their teammates and the competition they face.

HARO

As mentioned, the X-axis is ZS%, so the further right the player is the more often they get cushy O-zone starts. The Y-axis is the inverse HARD QoC, since it fit together more aesthetically with the other charts to present it that way. Since it is an inverse, the lower the player is, the lower their opponents' respective Goals/Fenwick Against were, i.e. facing off against better defensive opponents, whereas those higher up face off against those opponents who are more likely to give up Goals/Shot Attempts. The bubble size is based on HARO, so the larger the bubble is the bigger the better the player's offensive numbers are, with white bubbles representing negative values. The axes cross at the league-wide median values, and I also included additional axes for forwards in blue and defense in red.

5-on-5 Goals HARO - Median ZS% 50.5 (50.5 F and 50.3 D), Median HARD QoC -0.6 (-0.6 F and -0.5 D)

<a href="https://pbs.twimg.com/media/BvVUu3RIgAE4QHB.jpg:large" target="HARO"><img src="https://pbs.twimg.com/media/BvVUu3RIgAE4QHB.jpg:large"></a>

Close Fenwick HARO - Median ZS% 50.3 (50.3 F and 50.2 D), Median FHARD QoC -0.5 (-0.6 F and -0.4 D)

<a href="https://pbs.twimg.com/media/BvVU2CWIYAABYi9.jpg:large" target="FHARO"><img src="https://pbs.twimg.com/media/BvVU2CWIYAABYi9.jpg:large"></a>

So looking at the chart, the upper right quadrant are those that get the most beneficial O-zone starts while matching up against the opponents who are most likely to give up goals against, which could be seen as sheltered minutes, but could also indicate the team is good at adjustments and getting those players in the most beneficial situations. We expect them to succeed, so if you manage to be negative it is quite bad. Those in the lower right quadrant are used in offensive situations but also have to face off against better defensive opponents. Being that they start in the O-zone we expect them to succeed, but based on the opponents they face we aren't overly upset if they are slightly negative. Those in the upper left quadrant are getting tougher zone starts but aren't facing particularly effective defensive opponents. Starting in the D-zone we don't expect much from them, but based on the lesser QoC we would like to see them at least close to positive. Then lastly those in the lower left quadrant have the toughest zone starts and face the toughest defensive opponents. These players are expected to be negative, and if they somehow managed to come out ahead they would be amazing, of course if they were good enough to do so they would then get more minutes and wind up playing a bigger role which would shift them to a different quadrant.

HARD

Once again the X-axis is ZS%, with those further right receiving more cushy assignments. The Y-axis is HARO QoC, so the higher the player is the higher their opponents Goals/shot attempts tend to be, i.e. facing tougher offensive competition, whereas those lower on the chart face less effective offensive options. The bubbles are based on HARD, so the larger the bubble the better the player's defensive numbers are, with white bubbles representing negative values. The axes cross at the league-wide median values, and I also included additional axes for forwards in blue and defense in red.

5-on-5 Goals HARD - Median ZS% 50.5 (50.5 F and 50.3 D), Median HARO QoC -0.1 (0.0 F and -0.2 D)

<a href="https://pbs.twimg.com/media/BvVUy9vIUAAzILQ.jpg:large" target="HARD"><img src="https://pbs.twimg.com/media/BvVUy9vIUAAzILQ.jpg:large"></a>

Close Fenwick HARD - Median ZS% 50.3 (50.3 F and 50.2 D), Median FHARO QoC0.1 (0.1 F and 0.0 D)

<a href="https://pbs.twimg.com/media/BvVU5NOIcAE3yqG.jpg:large" target="FHARD"><img src="https://pbs.twimg.com/media/BvVU5NOIcAE3yqG.jpg:large"></a>

So looking at the chart, the lower right quadrant are the most sheltered, they have easier zone starts and face lesser offensive opponents. Those in the upper right quadrant still get easy O-zone starts, but they also face off against tougher offensive opponents. Those in the lower left have more difficult D-zone starts, but face off against weak offensive competition. And lastly those in the upper left have the toughest zone starts while facing the highest calibre of offensive opponent, they are the true shutdown players.