NHLNumbers Predictive Power Rankings

Derek Zona
February 23 2013 10:51AM

Is this the leader of the best team in the league?
Photo by Michael Miller via Wikimedia Commons

Power rankings are no fun for any writer or editor.  No matter the amount of effort or honesty involved in creating the rankings and slotting the teams, oversights happen and someone in every fanbase except the team ranked first overall will feel slighted and react like a toddler.

So we've set out to eliminate the effort by building a power rankings model based on underlying statistics and their predicative ability.  Our rankings aren't based solely on the current NHL standings, though points earned are a portion of the model, they are based on a number of underlying metrics so the rankings are constantly in flux. 

We're still tweaking the model, so we're not quite ready to unveil it, but consider this NHLNumbers' current best effort at predicting the standings and playoff pairings at the end of the season.  One item of note - 38% of the game is luck, and we don't attempt to predict or model that 38%, and we don't plan to.

Read Article | 8 Comments

Phil Kessel scores, and there was never anything to worry about

Derek Zona
February 08 2013 10:48AM

In my perfectly timed article on Phil Kessel yesterday, I noted that his scoreless streak, which ended last night, was not out of the norm for a shooter like Kessel:

"That said, if he continues to get pucks on goal they will start finding the twine eventually."

Read Article | 2 Comments

Phil Kessel & Streaks

Derek Zona
February 07 2013 12:00PM

Phil Kessel is getting slammed by fans and the MSM and defended broadly in the Statosphere. Why? Kessel has played 10 games and hasn't tickled the twine. His first 42 shots on goal show nothing but a giant zero on the stat sheet. 

Should Leafs fans be worried?

Read Article | 1 Comments

Toronto Plays A Special Brand Of Bad On Special Teams

Derek Zona
February 06 2013 11:33AM

If Toronto's special teamers were racehorses, we would would call him Elmer.  If they were Tibetan, they'd be vulture food.  If they were seafarers, they would live in Davy Jones locker.  The Maple Leafs have been a pretty bad team for a very long time, but their special teams have been even worse.  In 2009-10, the Leafs hit an NHL post-lockout low with an STE of 88.6, but that's par for the course for these Leafs.

Read Article | 3 Comments

2012-13 Early Season Special Teams Efficiency (STE)

Derek Zona
February 05 2013 05:25PM

I developed Special Teams Efficiency (STE) as a way of comparing the gains and losses in special teams efficiency between teams.  It's a measure of efficiency (% of power plays converted + % of penalties killed) and nothing else.  Penalties drawn or taken don't figure into it directly.  I explained then:

If we had a simple metric to apply to special teams, we could get a sense of just how much of an impact special teams can have on a season.  Since the league average for power play success percentage plus penalty kill success percentage is 100, why not add the two and use it as a baseline for measuring special teams efficiency?  Though the tendency to regress toward the mean may be somewhat less strong (but then again, maybe not!), it may still be useful as an evaluation tool.  In this case, we'll call the combined number Special Teams Efficiency, or STE. Since the post-lockout orgy of power plays (2005-2007) the per team yearly average of special teams situations is 643.  If all teams were to draw the same amount of penalties, a team with a 100 STE would net zero special teams goals. It follows, then, that a team with a 105 STE would net 32 special teams goals more than average, and a team with a 95 STE would net 32 special teams goals less than average.

STE correlates with season-to-season performance much more strongly than I'd initially anticipated.  Below is a table containing the R^2 Correlation between team STE and total points:

 Though the season is still too young to gain anything from goals gained/lost, there are some interesting numbers at play.

Read Article | 0 Comments