Turris line defies zone start logic

If you’re new to “zone starts”, a player’s offensive zone (OZ) start % is a simple measure of how often a player starts in the offensive zone versus the defensive zone.  (For a primer on corsi (a nuanced version of tracking shot attempts) check out this article)

Offensive Zone Start % Impact on Corsi

You don’t have to look very closely to see a correlation between OZ starts and better stats.  The Puck Stops Here blog recently looked at the Top 20 Players By Offensive Zone Starts and unsurprisingly these top 20 are also players with high shot attempt numbers.  So it appears that these players are being used in a manner that maximizes their shot attempt numbers and normally do have higher shot counts.  HockeyAnalysis.com posted a similar conclusion here.

The general thinking goes like this:  More OZ starts = more shot opportunities = better corsi numbers => more production over time.

But the actual impact of OZ starts is still up for debate.  The best argument I have seen so far is by Eric T. from NHL numbers who looked at this very issue. He challenges some of the earlier conclusions which found that each OZ start would, on average, increase a player’s corsi by 0.8 to 1.1. He states the obvious problem in assessing the impact of OZ starts:

“…taking a correlation of every players corsi with their zone start stats doesn’t take into account that it is the top line players that usually get the offensive zone starts and thus this likely over estimates the impact as these players do take more shots regardless of their zone start. “

In other words, top players who get more shots also get to start in the OZ more frequently which may skew the impact we think OZ starts are having on corsi.

Eric T. corrected for the issues he found and concluded that:

“offensive zone faceoffs increases your Fenwick For by about 0.31-0.32 shots [per OZ start], which means it probably increases your Corsi For by about 0.4 shots [per OZ start]”  

So, as far as I can tell, there is an impact, but it might not be as dramatic as some initially claimed.

Turris Line Stats

In the case of the Ottawa Senators this season, the second line has much higher relative corsi (their own corsi compared to to that of the team while they are not on the ice) than anyone else on the team.

relative corsi

(not even close)

Therefore you would expect that this line would be starting in the OZ more than the other lines on the team.  But this is not the case.

Below is a Player Usage Chart from HockeyAbstract.com.  (Make your own here)  For the purpose of this article, we are really just looking at relative corsi, denoted by the colour of the bubble (blue = good, orange=bad) and OZ start % (bubbles to the right = high OZ start %, bubbles to the left = lower OZ start %).

Sens_Player Usage Chart

Chart courtesy of HockeyAbstract.com

You can see here that even though the second line has the best relative corsi by far, they have the lowest OZ start % of all the forwards.  

Typically, when players are on the left-hand side of the chart, analysts refer to them as “defensive minded players” or say they are being played defensively by their coach (see Hockey Abstract by Rob Vollman, bottom page 125 or the summary of how zone starts are used by Panther’s blogger Alexander Calloway. for some examples).  We also generally expect their corsi to be lower since they should be taking more shots against in the defensive zone.  Given the above we’re left with some questions…

How good is the Turris line given that they don’t get favourable OZ starts and still dominate all Sens players in terms of corsi?

– and –

Is the Turris line being misused and could we see even superior corsi results if they saw a significant increase in OZ starts?

With respect to the first question – damn good.

With respect to the second question – yes, most certainly, but it’s less certain that the net effect of more OZ starts for the Turris line would be positive for the team as a whole. Clearly, Turris & MacArthur & (Conacher/Ryan) aren’t having any issues getting substantial shot attempts despite less OZ starts.  Further, if the Turris line gets more OZ starts then by necessity the Spezza line (and possibly others) will not.  So, while it could be good for the second line to get more OZ starts, more analysis is necessary to determine if a change would be good for the team overall.

Feel free to comment below or send your thoughts.