Friday, November 02, 2007

Detailed special teams statistics

I've been working with Gabe Desjardins at Behind The Net, helping feed him ideas for new stats, and his latest release includes some great data as it relates to teams playing in different situations.

There's a ton of information there (didn't you always want to know the rate of missed shots while shorthanded for Los Angeles?), but some of what I think is the most useful data is in the power-play and penalty-kill figures, where we can now determine the effectiveness of teams based on minutes played in that situation.

The Canadiens, for example, have the league's best power play, with 14 goals scored in 71 minutes of 5-on-4 play (as of Monday), which works out to a rate of scoring 39.4% of every two minute stretch.

Columbus, meanwhile, has the top penalty kill, with only three goals against 4-on-5 over 70.7 minutes played. That works out to a 91.5% kill rate.

There's a lot of other good info regarding shots on goal, for and against, in different situations.

Examples: Overall, Carolina has the most shots on goal per game, and Vancouver has the least. Detroit allows the fewest shots against, while Philadelphia allows the most.

But if you look at the shots on goal, for and against, by situation, things change considerably. Ottawa is the top team for shots on goal, both for and against, at even strength. On the power play, Florida is generating an incredible 57.4 shots per 60 minutes played, while the Islanders are putting up just 30.4.

Shorthanded, Los Angeles netminders are seeing by far the most shots, 55.8 per 60 minutes, while Nashville penalty killers are allowing just 27.5.

From that, you can determine things like team save percentages in certain situations, such as .824 for Toronto while shorthanded, but .893 at even strength.

Or how about the fact the Penguins lead the league with 28.2 missed shots per 60 minutes on the power play?

There's a ton of stuff there, and while the numbers are now a few days old, it's likely to be regularly updated all season. It's a real shame the league doesn't produce these numbers, but at least now there's a source for fans and bloggers to consult.



At 3:06 p.m., November 02, 2007, Blogger saskhab said...

I've been wanting SV% on special teams for a long time. That's awesome. How about some pressure stats for goalies? Like SV% in 1 goal games, 3rd periods of these games, last 5 minutes of these games, OT, etc. I want to see some "clutch" stats.

It should be noted that Columbus has not only allowed only 3 goals while down 4 on 5, but have also scored 2 goals themselves. That's an incredible PK. People should talk about GF/GA ratios in special teams more often.

At 3:43 p.m., November 02, 2007, Blogger James Mirtle said...

I know I've heard that a few times, but I tend to think of penalty kill goals as anomalies: There are so few of them every year, and generally they come against a team that makes a mistake and coughs up a breakaway.

Maybe I'm wrong, but I don't think shorthanded goals are a good indicator of a team's ability to win (or play well on special teams).

At 4:50 p.m., November 02, 2007, Blogger saskhab said...

It largely depends on who your personnel is, but it can at times be the result of great point pressure and exposing weaker point players (like forwards who play the point). It also depends on what kind of system you employ on the PK.

Nonetheless, if Columbus were to continue to have success at a similar rate on the PK, it almost makes taking a penalty not a concern for them.

I'm serious about "clutch" stats. They're considered so important in baseball, I don't know why they're never discussed in hockey beyond GWG's (which can be very arbitrary - the first goal of a game can be the winner, or the goal that puts a team up 5-1 could end up being the winner, just as a 3rd period tiebreaker can) and OT goals.

At 8:37 p.m., November 02, 2007, Blogger Hawerchuk said...

How do you define what's 'clutch'?

If you use Win Probability, OT goals are the clutchiest, followed by game-winning or -tying goals in the last two minutes, followed by the shootout.

I spent quite a while working on this and ultimately decided I didn't like clutch numbers. Dallas Jokinen led the league in clutch two years ago because of his penalty shots - and yet you'd never put him on the ice in the last two minutes of a game.

At 5:41 a.m., November 03, 2007, Blogger Bruce said...

If you use Win Probability, OT goals are the clutchiest, followed by game-winning or -tying goals in the last two minutes, followed by the shootout.

To me game-tying goals (which the league no longer even tracks on its official stats) are the most important. A GTG -- defined as the goal which sends a game to OT -- guarantees one point for the team, with a 50% chance at a second point through OT or shootout. So each one is effectively worth 1.5 standings points.

Whereas a winning goal, even in overtime, is merely worth the second point; and without that winning goal, the team would still have a chance at the second point in the continuation of overtime or the shootout, since it is guaranteed to be awarded at some point anyways. So it could be argued that the value of the winning goal is less than one standings point. For sure it has less value than the game-tying goal.

At 12:31 p.m., November 03, 2007, Blogger Hawerchuk said...

yeah, sorry, brain freeze. GTGs late in the game are the most valuable in the standings.

GTG, 19:59 3rd = 1.5 pts
GWG, 19:59 3rd or later = 0.5 pts
SO goal to go up by 1 = 0.33 pts

The GTG drops quite substantially in value the more time is left in the game.

At any rate, clutch stats easily get skewed by 1-2 goals or a lucky year in OT (see St. Louis Blues 2004). If anything, "clutchiness" is less repeatable than other skills.

At 12:33 p.m., November 04, 2007, Anonymous Mogen_david said...

TOI/game in each situation would be a great addition. Teams that spend less time short hand even if horrible at the PK can be just as effective as good PK.

At 12:35 p.m., November 04, 2007, Blogger James Mirtle said...

Exactly. That's something that he'll have up there quite soon, too.

At 7:33 p.m., November 04, 2007, Blogger Hawerchuk said...

TOI/game by situation is now there...Sorry for forgetting it before - too many stats to keep track of!


Post a Comment

<< Home

Links to this post:

Create a Link


Free Page Rank Checker
eXTReMe Tracker