Follow Up on Special Teams On-Ice & On-Ice Impact by Pending 2010 UFA Centers

Over the past few days, I've focused my posts on unrestricted free agent centers as the New Jersey Devils have a need at the position.  By analyzing who's there by a number of means, we can have a better idea of who to hope Lou signs on July 1 (or thereabouts). 

On Tuesday, I put up this post looking at the on-ice and on-ice impact numbers from Behind the Net for 4-on-5 situations among the UFA center pool.  I received this important comment by ILWT user triumph44:

i (and i’m just aping gabe desjardins and the contrarian goaltender on this issue) hesitate to take very much out of penalty killing stats in a single season because the sample sizes are so tiny. for instance, eric belanger played 130 minutes on the PK this season. and his team allowed 34.5 shots/60 when he was on the ice, so by my math, that’s 75 shots. but he allowed 10.26 goals/60 – by my math, that’s 22.2 goals. rounding off to 22 – that’s a shooting percentage of 29.3 while on ice against eric belanger. while i’d believe, without any real evidence to suggest it, that penalty killers have some control over shooting percentage against, that’s quite a ridiculous number and i’d be shocked if belanger managed to repeat it next year, no matter where he ends up.

The emphasis is mine because triumph44 makes a good point.  And I believe it applies for the power play as well.  Some level of time on ice per 60 is required to show that the player actually got some minutes when he was on either special team.  However, whether they maintained it for the whole season or not is a whole different story. Even then, the number of special team opportunities vary from game to game, which hurts the overall sample size. 

Incidentally, I think the even strength numbers have meaning. While rare, a game can have no calls at all; and even in games where there's lots of penalties, the majority of the game will likely be played 5-on-5.   The sample size is much larger for even strength as opposed to special teams. Additional data wouldn't hurt at all or be a waste of time - it can be used to strengthen or weaken a case for a player. 

That said, to follow-up on my agreement with triumph44's point, I went back to Behind the Net and pulled the 5-on-4 and 4-on-5 numbers from the 2007-08 and 2008-09 seasons for each of the players in both posts.  For the purposes of expediency, I did not rank the numbers; I'm presenting them as-is.  Lastly, set your viewing to wide to ensure that you see the entire table before continuing after the jump.  There, we can determine, for example, if 2009-10 performance was exceptionally good compared to their history or more in line with what they accomplished.

5-on-4 Situations 2007-08 & 2008-09 On-Ice and On-Ice impact

Remember:  The impact numbers are calculated by off-ice stat minus on-ice stat.  Since these are offensive numbers, negative values are desirable in those columns.

Given that my metric in the original post was a minimum TOI/60 of 1, Glen Metropolit wouldn't be here if I applied it. That said, the player who really sticks out here is Olli Jokinen.  He managed to have a positive impact on goals for per 60 (GF/60) and shots for per 60 (SF/60) in both seasons.  No other UFA center from the original post managed that, though Matt Cullen came close.   Jokinen's numbers are even more impressive given the large amount of TOI/60 he logged on 5-on-4 situations.  Oddly enough, Cullen somehow saw his TOI/60 drop like a rock from 2007-08 to 2008-09.  4.83 TOI/60 was one of the highest in the league, so it's a really noticeable drop. Did he somehow start off so poorly and was unable to get his position back on the Carolina power play in 08-09?  What happened, Canes fans?

One item of interest is to see how a player did when they ended the next season a totally different team. Brendan Morrison was night-and-day here. Though playing so few games in 07-08 may have prevented the numbers from falling to where his true impact may have lied?  Kyle Wellwood saw his impact increase after his time in Toronto.  Vinny Prospal was quite good on both teams - his negative impact wasn't all that large to be honest.

Which leads me to this point, don't just look at the impact numbers, look at the on-ice GF/60 and SF/60 stats too. Tomas Plekanec witnessed the most PPGs on-ice in 2007-08 and his presence led to a decrease in GF/60. Yet, that decrease wasn't even 0.1; his on-ice GF/60 was quite high.   It also shows that perhaps these aren't sustainable numbers for some players.  Plekanec saw less TOI/60 on 5-on-4 situations in the following season; actually had a big positive impact on SF/60, yet the goals weren't being scored so often - hence the vastly lower GF/60.  Perhaps the real impressive UFA center here is Saku Koivu, as he maintained his numbers across two seasons in Anaheim.

For convenience's sake, here's the 2009-10 5-on-4 table that you can use for comparison purposes.  Let me know in the comments what you observe and think as a result.  One thing I noticed, Saku Koivu looks more impressive as someone who can help on the power play.

4-on-5 Situations 2007-08 & 2008-09 On-Ice  & On-Ice Impact

Remember: Now the focus is on defensive numbers, and so the impact numbers should be positive - since we want the players to reduce shots against per 60 (SA/60) and goals against per 60 (GA/60).

First, I did not make a mistake.  Rickard Wallin was in the original post, yet he is not here because he was not in the NHL in 2007-08 or 2008-09.  Wallin supporters, I'm sorry.

Second, Richard Park wasn't as awful in terms of GA/60 or total on-ice GA as he was in 2009-10.  Let's hope for his sake and who he signs with that this past season is not truly indicative of what Park can do in a 4-on-5 situation.

Third, the UFA centers who played significant time on penalty killing situations like John Madden, Rob Niedermayer, Manny Malhotra, and Jeff Halpern, most of them had a negative impact on SA/60 and/or GA/60.  Players like Cullen (who wouldn't have met the criteria in 07-08), Plekanec, and Koivu had great impact but didn't play nearly as much.  The exception was Scott Nichol and that only held for 07-08; a little more responsibility in the following season saw his numbers shoot up across the board. Does more ice time lead to worse numbers in general? 

Again, I think the on-ice stats definitely have to be taken into context and perhaps consistent results are better than just great impacts.  In that way, that Niedermayer, Madden, and Malhotra (among others) got and keep getting significant spots on PK units make some more sense.   No, the GA/60 and/or SA/60 may get worse; but you can understand what you're getting and play them accordingly.  Malhotra was the most consistent between 2007-08 and 2008-09.

For convenience's sake, here's the 2009-10 chart for reference and comparison.  Malhotra got less time in San Jose and achieved worse numbers.  Before you say, "so much for consistency," he may have been a totally different role and so the numbers were affected accordingly. Speaking of Sharks, Nichol definitely bounced back after 2009-10.  Let me know what else you discover in the comments.

Your Turn

Thanks for reading, especially to triumph44 for raising the point.  There's more data. We can make more comparisons.  Yet, I now wonder whether this is really the best way about judging special teams performance.  Perhaps I'm overthinking this and the idea is to understand what the player does at even strength and their talents and then decide whether they are suitable for special teams.  This way, a checker like Madden or Malhotra will get PK time and a scoring talent like Prospal or Jokinen will get PP time regardless of whether they truly make their teams better in terms of shooting and scoring rates.

Please let me know what you think of all this, what you observed, and other insights you may have in the comments.