The Rangers Are Terrible on the Road

All data below provided by Corsica.Hockey

Flaws in the analysis below:
- There is insight missing. These context/input metrics are not enough, and the output metrics are also not enough.
- Other items of interest could be things like Penalties Drawn/Taken home vs away and the following special teams success. Hank/Georgiev splits, and their GSAA home/away. I mean, a million other things. This barely scratches the surface
- Building off above, only 5v5
- I steered away from using relGF% and individual points due to sample size, but if there’s a player who is torching points at home and silent on the road, could be worthwhile digging in there, as well.
- Many more!

At the time I sat down to draw up this blog, the Rangers have played 28 games on the season under David Quinn. Their record during home games, a more than respectable 10-4-1. On the road, however, they fall to 3-8-2. Among NHL teams, this is the most egregious difference between home and away record.

Uncertain as I am that there is a clear way to figure out why this is happening by just using data, I wanted to take a look anyway. We’ll start with context metrics like TOI%, Zone Start Rate, Quality of Competition and Teammates, and then finally output metrics like relCF% and relxGF% - and a brief analysis of what I think each visualization tells us.

First, Time on Ice Percentage.

dashed line is where Home TOI% == Away TOI%

dashed line is where Home TOI% == Away TOI%

Really nothing to dive into here. Quinn is giving basically equal amounts of TOI during both home and away games. There are a couple of players who find themselves ‘off the line’. Chytil seems to play more at home than on the road and Andersson vice versa, but this could very easily just be sample size oddities.

Next, Zone Starts. Hopefully begin to see some changes here where the difference between first change and last change may come into play.

haZSR.png

And differences we see. Names that stand out for more offense on the road than at home: Andersson, DeAngelo, Howden, Vesey, Smith. For the opposite, more defensive zone starts at home than on the road: Zibaenjad, Staal, Pionk, McLeod, Buchnevich. These are names we’ll have to keep in mind when we start looking out more performance related metrics, but, we continue with context.

Quality of Competition. My anticipation is that we see more discrepancies here than we do with Quality of Teammates. Quinn can control his five-man units much more easily home and away than he can who his players line up against. Both metrics of quality we’ll be using are based on Corsica.Hockey’s xG model.

haComp.png

One thing I love that this visualization shows, that admittedly I forgot it would. Look how bunched together a lot of these names are. The differences in Quality of Competition are much, much lower than we expect them to be. It’s not as easy as Staal and Pionk always out there against the other team’s top opposition.

At the same time, these bunched together data points are going to limit the oddities that we do see in differences between home and away, which should’ve been anticipated by me, but alas, wrong again. Of course we have some major outliers here like Andersson, Claesson, Vesey, Zucc, Buchnevich, and McLeod.

Otherwise, I’m not entirely certain if there’s anything we can discern from this besides continuing to realize that the differences in QoC are small.

Now the same, but with teammates.

haTeam.png

This, as anticipated, is very similar to the zone start rate visualization. This is something that Quinn has full control over for both home and road games. It doesn’t matter what the opposition coach does because Quinn controls what five man units go onto the ice.

It should be highlighted though that the bunching of players here is far more spread out than the bunching on the same Competition based viz.

Biggest item that stands out to me here is Pionk, who both on the road and at home gets fed the easiest usage in terms of Quality of Teammates, which will be important to keep in mind in the now coming output visualizations.

Relative shot attempt percentage. Keeping in mind that this metric tells us the difference between an individuals shot attempt percentage and the shot attempt percentage of the team when that individual is not on the ice. This same methodology is used in the second visualization, Relative expected goals percentage.

haCF.png

Neal Pionk: Best teammates, bad output. Gotta love it.

So this is where we have to start using our input visualizations with our output visualizations. What do we remember from those first few charts? TOI and Teammates show no major differences between home and away, totally controlled by Quinn. Quality of competition, while showing differences, is too bunched together to glean anything from. Zone start rate, however, did show some items of interest. Recall:

And differences we see. Names that stand out for more offense on the road than at home: Andersson, DeAngelo, Howden, Vesey, Smith. For the opposite, more defensive zone starts at home than on the road: Zibaenjad, Staal, Pionk, McLeod, Buchnevich. These are names we’ll have to keep in mind when we start looking out more performance related metrics, but, we continue with context.

Andersson: Bad both on the home and on the road, worse at home.
DeAngelo: Positive in both venues, better at home.
Howden: Bad in both, basically the same
Vesey: Positive on the road and negative at home
Smith: Good, even.

Zibanejad: Good both, better on the road
Staal: Horrid on the road and a positive player at home
Pionk: I mean, come on
McLeod: Bad hockey players are bad at hockey
Buchnevich: Great at home, bad on the road.

Quick idea, maybe limit Pionk’s minutes on the road? Let’s see what relative xGF% says…

haxgf.png

Just want to take a random moment and point out that Kreider is absolutely amazing at hockey. Moving on to the same exercise as above:

Andersson: Better on the road than at home.
DeAngelo: Negative in both venues, worse on the road.
Howden: Negative in both venues, worse at home
Vesey: Almost even on the road and negative at home
Smith: Both negative, but about even.

Zibanejad: Great on the road and bad at home - interesting.
Staal: Positive both, about the same, this has been a good exercise for Staal.
Pionk: Better numbers than on relCF, but still a negative player in both venues.
McLeod: Bad hockey players are bad at hockey
Buchnevich: Negative in both venues, about even.

In summary, i t’s probably worthwhile to create a quick data table that looks at TOI% and relCF% and relXGF%.

First, away, where there is clearly something wrong with this team:

awaySummary.PNG

Pionk and Staal get caved on the road. I’m not sure if it’s as easy as just saying: well, play them less! But, it is something to keep an eye on during road games.

homeSummary.PNG

The Staal-Pionk imapct is somewhat mitigated on home ice where they’re, still bad, but not a complete waste of ice. Hayes is really, really good. Imagine what he could do with proper QoT? But that seems to be changing tonight as he suits up at the wing with Kreider and Zibanejad. I’m interested to see how that goes.

Ultimately, there wasn’t a ton of insight here to get some information. The home/road splits for the Rangers were equally as terrible under AV last season, 21 wins at home and only 13 at home. This, a major difference from the year before, with 21 home wins and 27 road wins.

Perhaps it’s just randomness at work and small sample sizes. Ultimately, besides overplaying Staal and Pionk on the road, it is seemingly difficult to explain via context the differences in success.