# Individual Point Percentage on the Power Play (2008-2012)

Scott Reynolds
October 16 2012 07:59AM

Photo: Beanhugger/Wikimedia/CC BY 3.0

Over the few days I've been talking about individual point percentage (i.e. the number of times an individual player gets either a goal or an assist compared to the number of total goals-for scored while he's on the ice) during five on five play. Of course, there are also a significant number of goals scored on the power play, and so today I'll be looking at the individual point percentages for forwards at five-on-four.

Five-on-four play is always a little bit tougher to analyze because we don't have nearly as much data. As such, I've decided to skip over presenting the data from the 2011-12 season (though I do have it, so if you'd like a copy, just let me know) and jump straight to presenting the data for the last five seasons combined.

When I calculated the IPP during five-on-four play for all of the forwards in the NHL over the last five seasons, there were just 136 who were on the ice for at least 80 goals-for, which is where I decided to set the cut-off. The average individual point percentage in that group is 61.6% and the median is 62.6%, and the standard deviation is 8.1%. You'll notice that those first two numbers are significantly lower than what we saw for five-on-five play (70.0% average and 69.5% median). This makes complete sense: defensemen play a much bigger offensive role when teams are on the power play, which leaves fewer points available for the forwards. You'll also notice that the standard deviation is higher, which also comes as no surprise (we're dealing with a smaller sample, both in terms of total qualifying players and in the number of goals).

Before going any further, let's take a look at the chart  (the raw data comes from Gabriel Desjardins' behindthenet.ca):

This list of the top players really drives home just how much more involved defensemen are during five-on-four play (we'll be taking a look at the IPPs for defensemen next week).

The other thing that jumps out for me is the presence of one San Jose Shark (Ryane Clowe), two Anaheim Ducks (Teemu Selanne, Bobby Ryan), and one Detroit Red Wing (Pavel Datsyuk). In other words, of the top twenty-one players, four come from the league's three best power plays. Of course, there are also some really poor power plays represented on this list. No one has been worse on the power play over these last several seasons than the Edmonton Oilers (dead last in shots per minute for four out of five seasons and 25th in the other), and yet there sits Ales Hemsky in 8th overall.

So why is that? I think it's because this chart tells us as much about the system as it does about any one individual. For the Oilers, everything ran through Hemsky (until the arrival of Ryan Nugent-Hopkins). Having a quarterback like that can work well, but it can lead to a power play that's way too stagnant.

Toward the top of this list of the murky middle, we see a lot of primary puck distributors (Joe Thornton, Jason Spezza, Henrik Sedin) and shooters (Jarome Iginla, Daniel Sedin, Steven Stamkos), and towards the bottom we see guys who spent signifcant time on the point (Mikael Samuelsson, Jarret Stoll) or parked in front of the net (Ryan Smyth, David Backes). The last part of this chart is filled with more examples of the latter:

That's a whole lot of net presence. It's no secret that these players don't get credited in the boxscore for setting screens, and it's also no secret that teams think that this tactic has real value (every team is doing it). Points simply aren't measuring the value that these players bring when their team is up a man. That makes it tough to know whether or not these players are more effective than their peers. Tomas Holmstrom and Ryan Smyth are often given plenty of credit for what they do in front of the net, but are they really any better than (say) Todd Bertuzzi or Dustin Penner? And if so, how much? The way we're counting data right now makes that next to impossible to answer.

#1 Eric T.
October 16 2012, 09:43AM

I wrote elsewhere (http://www.broadstreethockey.com/2012/8/10/3233574/claude-giroux-goals-assists-points-goals-n-more-goals) about how IPP was at the root of Giroux's enormous power play assist total last year.

He assisted on 53% of the goals his teammates scored with him on the ice at 5v4 in 2010-11. He assisted on 88% of them in 2011-12. The second-highest figure in 2011-12 was 82%, and no player was over 70% for both '10-11 and '11-12.

So it seems natural to project his 5v4 IPP falling quite a bit, and the power play assist total coming down with it.

#3 Eric T.
October 16 2012, 10:59AM

@Scott Reynolds

Yeah, I'm just breaking IPP out into two separate components: what fraction of the time did he record an assist when a teammate scored, and what fraction of his personal shots went in?

His 2010-11 IPP was fueled by his 24.2% shooting percentage. His 2011-12 IPP was fueled by him getting assists on 89.3% of his teammates' goals. Neither of those figures is likely to be sustained.

His long-run rate might be above the mean, but it's close enough to the mean that he could easily be below the mean in any given season; all it'd take is a season where he combines his 2010-11 assist rate with his 2011-12 shooting percentage.

#4 Derek Zona
October 18 2012, 11:20AM

So you're saying Hemsky is either terrible and dragging down the Oilers, or the Oilers are terrible and Hemsky is downright amazing.