August 29 2013 11:17AM
One of my biggest complaints with the data currently compiled by various hockey databases is that none of it, at least that I am aware of, allows for a user to restrict components of individual or team averages by date. This becomes particularly problematic when looking for trends in data that don't necessarily manifest in year-long averages.
No one was particularly surprised when Montreal lost to a comparatively strong Ottawa team, and data-friendly hockey fans saw Detroit giving Anaheim a world of trouble in round one. But, splitting the numbers to focus on the more-recent team data would have provided us with a far more definitive look at other match-ups, including that four-game sweep of the Vancouver Cauncks by the San Jose Sharks.
In an effort to identify rising/falling trends, I decided to split each team's Fenwick Close in half -- a possession and scoring chance differential proxy that shows little vulnerability to score effects -- into two twenty-four game samples (generally spanning the beginning of the regular season to early-March, and early-March to the end of the regular season). Because the conferences didn't mingle in play, I separated the Eastern and Western Conference, treating them as two separate leagues. In reality, they were.
On both ends, the data is sort of fascinating.
Let's look at the Eastern Conference first. And, don't be confused by the "split-half" labelling. It holds a literal definition here.
The New York Islanders are the team that should stand out to you most. First twenty-four, an average team. Next twenty-four, a step or two away from holding the best Fenwick Close in the league, trailing only .... New Jersey. Man, those Devils. Fun fact about that team: their save percentage at evens over that final twenty-four game stretch was sub .900. That's how you more or less lose the majority of hockey games while simultaneously dominating pace and controlling the shot-clock.
Ottawa's another team that improved as the season progressed. Part of that is getting a number of guys back from injury. Part of that is calling up fourth-round picks like J.G. Pageau and watching him pick-up more than six of every ten shots at evens. Every button Paul MacLean pushed last year worked, except for that whole 'let's avoid injuries, particularly those to the achilles tendon' one.
Last: Winnipeg. Not a way to finish strong. You know who did finish strong, though? Unrestricted free agent Ron Hainsey, who had the team's best Fenwick Close (53.0%) in the second-half of the season. You ask, why isn't Ron Hainsey signed? Perhaps because he was out-scored 16-8 while on the ice. General managers and coaches can often get caught up in the small samples of goal-based analysis, and mounting evidence certainly suggests that Hainsey -- although doing just about everything right -- is going to be punished because the guys around him weren't up to par. And that includes the guy between the pipes, too.
Now, the Western Conference.
So, by now I think it's been exhausted how great the Detroit Red Wings were coming down the stretch. San Jose, to a lesser extent, too. Interesting thing about that San Jose/Vancouver first-round match-up I alluded to earlier: by a 48-game Fenwick Close sample, you were talking about two teams just 0.5% apart. But, by splitting the sample in half and focusing on the twenty-four most recent games, you had an San Jose team 3.0% better than Vancouver. There's a lot of other reasons why the Canucks lost that series beyond a simple possession/scoring chance proxy, but boy, looking at it from this perspective should've raised a series of red flags.
On the other side, two really excellent attempts at tanking by Dallas and Calgary. What's funny is neither team was extremely successful at doing it. And both teams, stripped of talent and chasing down lottery picks, still posted a better second-half Fenwick Close than playoff-bound Toronto. Bizarre.
Since I touched on an individual above, I guess I'll circle one here. LA Kings F Justin Williams posted a 66.5% Fenwick Close during the second-half of the season, which is absolutely terrifying. And incredible. Equal parts terrifying and incredible. The LA machine is something I'm excited to watch again next season, and I'm really going to make an extra-effort to dial in on what Justin Williams does to tilt the ice like he does. He's not an extra-flashy guy, but man does he dominate.
Final graph showing the change in Fenwick Close from 1H to 2H:
You know, Columbus gave it a valiant effort and watched their numbers jump accordingly, but they truly were an awful team in the first-half of the year. I'm more interested in Anaheim here, a team just about everyone lambasted and labeled as being a paper-tiger masked by unsustainable on-ice shooting percentages. Fair argument, fair criticism. But, they did well clean up their act, unlike their Eastern Conference comparable in Toronto.
Maybe Anaheim is a team to watch next year. I'm skeptical they're a playoff team again, but I'm hesitant to write them off just yet. Brutally tough teams at the top of their division, and pure hockey fodder at the bottom.