You’ve probably heard this, but the first round of the 2015 NHL Entry Draft is tonight in beautiful Sunrise, Florida. If you frequent this website, there’s a good chance you have been following not only the great work Kyle has done in previewing and profiling potential draft picks, but also Michelle’s tireless efforts in covering the Red Wings‘ current prospect pool. With those things said, your thoughts about tonight are probably something along the lines of, “We need a defenseman.” While it’s true that most of the higher end talents in the Detroit system are forwards, taking an offensive player tonight could actually yield a preferable result, because, to put it quite bluntly, forwards are more important than defensemen.
The idea that forwards are more valuable than defensemen isn’t necessarily a new idea. What is new is how a really smart guy named Tom Awad managed to quantify it. In a section of Rob Vollman’s Hockey Abstract 2014 titled “What Make Good Players Good?” Awad categorized players into different “tiers” based up their 5v5 ice time. He chose TOI because, in theory, the best players will play the most. Awad created four tiers of forwards and three of defensemen, thus mimicking the way in which hockey players are deployed. I’m going to relay some of the results to you, but I want to start with a disclaimer. I will be presenting two specific charts from a vast series of situational data analysis, followed by the main takeaways Awad gathered from the entire study. The details of said study span over 30 pages, and charts A and B are not the only things that led to conclusion C.
Chart courtesy of Rob Vollman’s Hockey Abstract 2014
Awad’s conclusions to the entire forward exercise are as follows:
1. Elite (“Tier 1”) forwards get more ice time than weaker forwards, so ice time is a loose proxy for play skill. However, even-strength ice time is reasonably distributed among all players.
2. Tier 1 forwards outscore their opponents through a combination of better puck possession and better finishing ability. Of the two, finishing ability is slightly more important.
3. Shooting percentage is highly correlated with player strength in all manpower situations and does not depend on shot quality.
4. There is little correlation between overall player strength and zone starts. However, there is a high positive correlation between player caliber and quality of opposition as well as quality of teammates.
5. Tier 1 forwards don’t draw many more penalties, or take fewer penalties, than Tier 2 or Tier 3 players. However, Tier 4 players take more penalties.
6. Tier 1 and Tier 2 forwards get the lion’s share of power-play time. Tier 4 forwards get virtually none.
7. The main driver of overall ice time for forwards is offensive ability. Almost all first line forwards also play regularly on the power play. However, a good defensive forward will get a reasonable amount of even-strength ice time and large amounts of penalty-killing ice time.
Awad then examined defensive players. The contrast is stark and easy to spot, even if you know nothing about analytics.
Chart courtesy of Rob Vollman’s Hockey Abstract 2014
The two charts cover different data points, so it isn’t completely an apples to apples exercise. There is one clear trend, however: results change dramatically less between top and bottom tier defensemen, compared to those of forwards. Here’s how Awad explains it:
What are the most interesting observations from this table? First of all, unlike forwards, the goal differential and expected goal differential results are not extremely different by group. While the difference between Tier 1 and Tier 4 forwards was 0.76 goals per 60 minutes and 0.33 expected goals per 60 minutes, the difference for defencemen is only 0.13 goals, six times less, and the difference in expected goals is 0.07. In fact, Tier 2 defencemen have a slightly better expected goal differential than Tier 1 defencemen! The difference in shot differential was slightly more pronounced, with a difference of 1.33 shots per 60 minutes, although at an average even-strength shooting percentage of 8%, that translates to only 0.11 goals per 60 minutes. This all points to an important conclusion: results at even strength are driven primarily by forwards. This is not to say that offensive play is more important than defensive play; simply that, in the NHL, the players who contribute the most to outscoring the opposition at 5v5 are first-line forwards, not top-pair defencemen.
As I stated previously, Awad examines this much deeper than I am presenting it, specifically in regards to situational use. But these were his main takeaways from the defensive study:
1. There is not that much difference in even-strength results among different tiers of defencemen. Results at even strength seem to be driven primarily by forwards.
2. There is little correlation between defenceman ice time and Zone Starts. However, there is a high positive correlation between defenceman ice time and Quality of Opposition.
3. A top-two defenceman will typically play a significant role on the power play. A top-four defenceman will typically play a significant role on the penalty kill.
4. When killing penalties, better defencemen are capable of reducing both the quantity of shots faced and their chances of scoring.
5. Defencemen, as a whole, take more penalties than they draw, but Tier 1 defencemen take fewer of them, despite facing more dangerous opponents.
6. Unlike forwards, where offensive ability was the main driver of ice time, there are many profiles of defencemen. The best defencemen are those that excel at both ends of the ice, but many defencemen in the NHL contribute despite being weaker in one set of skills.
In summary, saying defenseman are less important than forwards is a bit simplistic. But the fact remains that, by a large statistical margin, forwards are the primary results drivers at even strength. Keep that in mind tonight if a forward’s name is called with the the Red Wings’ pick.
Special thanks to Tom Awad and Rob Vollman for the information in this piece. I would strongly encourage all hockey fans to read Vollman’s Hockey Abstract and Hockey Abstract 2014. They can be purchased together here.
