Shattenkirk and McDonagh: So Happy Together

Friend of the Blog NaturalStatTrick introduced a new “WOWY” tool today. WOWY, for clarification, shows how players and teammates perform together and apart.  NaturalStatTrick’s new WOWY tool is unique in that it allows you to choose up to 5 players at one time.  While this can be fun (and provide a lot of useful information for developing pairs and skating units), my interest was mostly in the Ryan McDonagh/Kevin Shattenkirk pairing.  Going into the season, most “experts” expected Shattenkirk to slot in to McDonagh’s right and finally give McDonagh the “top pairing” partner he’s been missing since [Anchor_Img_4346.jpeg] ascended to the top pair. 

Well we know how that worked out. The Rangers started the season terribly, and the McDonagh-Shattenkirk  pairing was over before it really got started.  NaturalStatTrick says the pair has played 43:07 minutes at “Even Strength” in the 14 games so far this season (3:08 per game) – and that would include situations where 27/22 may have been on the ice late in a game the Rangers trailed (but not empty net situations). Since game 3, they’ve barely played together, and their results should be taken very lightly. 

BUUUUUUUTTTT, maybe separating them was a little rash and might be worth revisiting?  

click to enlarge 

The first lines in each chunk are the McDonagh/Shattenkirk pair results. The second lines are McDonagh’s results without Shattenkirk. The third lines are Shattenkirk’s results without McDonagh. The yellow highlights show the “best” results in each of the stats included. [Note – NaturalStatTrick, as of now, doesn’t include some of the other metrics that are all the rage now, like an Expected Goals metric or a WAR-stat]. 

You’ll see the obvious, in the limited minutes played, McDonagh/Shattenkirk’s together results are better than their separated results.  In some cases, the difference is substantial. 

click to enlarge

Now we look at the goal-based results, and we see more variation. This makes sense, because this is a really small sample, and even in a larger sample, goals are subject to more randomness than shots. We see that here. For example – would you have expected Shattenkirk/Non-McDonaghPartner to have the best Goals Against/60 of these different options?  [Maybe you would, when you consider some of the other defensemen Vigneault has given McDonagh to work with this season].

We see the McDonagh/Shattenkirk pairing has the worst Goals Against/Rel of all the pairings, and it’s not even close. But we know that goals, both for and against, are also subject to something else that can fluctuate wildly – goaltending. While the McDonagh/Shattenkirk pair has a pretty good shooting percentage (11.11%), it’s had a horrific save percentage (80.95% or .809).   

So what difference does it make?

These numbers are of limited value over short time periods. There could be any number of things influencing these results, beyond just goaltending, that a deeper statistic- and game-tape-focused review could provide. But these surface-level numbers make me think McDonagh/Shattenkirk deserve some more time as a pairing.

…Again, I’m right in my analysis.

PS – Support NaturalStatTrick here.