Playing with Crosby and IPP

Scott Reynolds
November 20 2012 03:17PM


Photo by Michael Miller via Wikimedia Commons

Last month, I looked at something called individual point percentage (IPP) both for forwards and for defensemen. To recap the concept, individual point percentage is a calculation of the number of times an individual player gets a point (either a goal or an assist) relative to the number of total goals scored while he's on the ice. So, for example, if a player is on the ice for fifty goals-for during five-on-five play over the course of the season and he gets a point on forty of them, his individual point percentage at five-on-five would be 80%.

Most forwards end up at about 70% over the long haul, but there are some that buck the trend. Sidney Crosby led the league over the last four seasons with an IPP of 84%. One of the things discussed in the comments to those posts was what kind of impact playing with a guy like Crosby might have. Points are assigned on a zero-sum basis, so if he's getting more, who's getting less? 

As you might expect, the answer is everyone. If we look at the numbers of Crosby's most common linemates (both forwards and defensemen) over the last two seasons, we notice a distinct trend for them with Crosby on the ice:

Just one player has a better IPP with Crosby on the ice that with Crosby on the bench, and that's Brooks Orpik. We are, of course, looking at pretty small numbers here, so there's going to be some noise, but I think the trend is clear: players who are on the ice with Crosby are less involved in the offense and handle the puck less frequently. This makes good sense. Sidney Crosby is an elite player, and when he's on the ice, you want him handling the puck as much as possible, something that will turn his linemates into more complementary players.

When David Johnson looked specifically at what happens to Evgeni Malkin when he's on the ice with Crosby, he came to a similar conclusion noting that while the Penguins do very well with both players on the ice, it may not be in Pittsburgh's best interest to use them together except in high leverage situations, which is mostly what we've seen from Pittsburgh's coach, Dan Bylsma.

Thinking more broadly, this gives more weight to the idea that IPP is most helpful for telling us which players are driving the offense on their line, and not necessarily which players are the best at driving offense full-stop. Jamie Benn and Ales Hemsky are both excellent players, but I wouldn't put either of them among the top ten in the league at driving five-on-five offense even though both of them are among the top ten in IPP over the last four seasons. These results also help to provide context for a player like Joe Pavelski, who might be perfectly capable of driving offensive results, but doesn't do it as frequently when he's paired with Joe Thornton. In the end, I think that, when we have enough data, IPP tells us quite a lot about how individual players work together, and about which players are able to drive offense in specified circumstances, but is less effective at identifying the best players in the league in any kind of systematic fashion.

 

Previously by Scott Reynolds

 

889688bd2efc9140bec34eeba34eacde
Avatar
#1 Ben Wendorf
November 20 2012, 03:31PM
Trash it!
0
trashes
+1
0
props

The question you might dread: could this be one of those things where extreme examples like Crosby show up in the data, but the majority of the forward population will show minimal effect? That's essentially what I was finding with %AttSh, and I recall Gabe alluding to that when he looked at Mario Lemieux's effects on his linemates. I'm also well aware that that kind of research can be a royal pain in the ass, but I figured I'd throw it out there.

Avatar
#3 Derek Zona
November 20 2012, 09:16PM
Trash it!
0
trashes
+1
0
props

I think you understate Benn and Hemsky's value at 5 on 5. I know the number is overblown (probably) in each case, but there are very few players I'd rather have.

Comments are closed for this article.