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

Scott Reynolds
October 26 2012 07:14AM

Over the last few days I've looked at the 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) for defensemen during five-on-five play, starting with their performance in 2011-12, and then looking at their performance over the last five seasons. Defensemen, however, see a disproportionate amount of their offense generated on the power play, and so today I'll be looking at the individual point percentages for defensemen 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. Single-season data isn't usually going to be particularly significant, and so I've skipped straight to generating the data for the last five seasons combined. Even then, there are just 56 defensemen who were on the ice for at least 80 goals-for. Both the average and media individual point percentage for that group is 54.6% and the standard deviation is a very low 4.6%.

The first thing that jumps out to me is how much closer the defensemen are to the forwards at five-on-four. Forwards had an average IPP of 61.6% at five-on-four, for a difference of just seven percentage points. At five-on-five, the average for forwards was 69.5% compared to 30.2% for defensemen, a difference of thirty-nine percentage points. That's a massive swing, and I think, shows us just how different the game is when one team is playing up a man.

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

Not many players listed here, but one guy who would join them is five-on-five leader Erik Karlsson, who missed the cut-off by just seven goals. He had 51 points on 73 goals-for for an IPP of 69.9%. Even if we assume he gets no points on seven more goals (to bring him up to the cut-off), he'd still place third with an IPP of 63.8%.

The rest of the players on this list come as something of a surprise, at least to me. No one looks like he obviously doesn't belong, but this would not have been the group I picked for "most involved in their team's power play". I had figured on seeing most of the top goal-scorers from 2011-12, but of the top five, only Chara makes the list (Weber, Phaneuf, Kronwall are all on the outside looking in and Jason Garrison didn't have enough events to qualify).

Not much to choose from here. It's very rare that a player will be on the ice for more than 50 goals-for during five-on-four play in any given season, which means that this middle group would likely be separated by just four or five points.All of the shooters I mentioned previously make this list. Shea Weber is sitting right on the average, which I find quite interesting. Weber may only get points on 54.5% of goals, but he's no doubt involved in a lot more simply by being on the ice and stretching the opposition's box. It's always difficult to seperate individuals from the work of the team, but I think that's especially true at five-on-four when the strategy is at its most systemic.

Much like with the group at the top of the list, it would be difficult to predict that these seven men would be at the bottom. Boyle (puck-mover) and Salo (shooter) are especially surprising to me because both men seem very involved in the power play. Salo changed teams over the summer, but Boyle will presumably continue in a similar role with the Sharks, who are probably the best five-on-four team in the NHL. I'd like to watch him closely at five-on-four to better understand why he's, at least by this measure, less involved in the power play than most of his peers.

### Previously by Scott Reynolds

#1 Eric T.
October 26 2012, 08:06AM

I think I counted 56 guys on the list, so we're basically just looking at the top two PP guys on each team.

Some of your surprise at the rankings might be because we've effectively narrowed our focus to reasonably skilled PP guys. If we had 150 guys who met the threshold, we'd be looking at a wider range of talent and might be less surprised by who ends up at the top or bottom of the list.

I guess what I'm suggesting is that the noise in IPP might make it hard to distinguish the good from the great. The simple error on these numbers is in the 5-7% range, enough to easily move a guy from average to the top grouping.

#3 Eric T.
October 26 2012, 12:03PM

@Scott Reynolds

Ah, I didn't realize your 54.6% figure was just for those guys.

I guess calculating an overall average is a bit tricky. I was thinking you could count all 5v4 goals, count all 5v4 defensemen points, and do D pts / 5v4 goals / 2. But since teams don't always use two defensemen, it's not quite that simple -- you'd have to replace 2 with the proportion of 5v4 ice time that's allotted to defensemen. So the calculation would go as follows:

Step 1, Calculate 5v4 D usage: For each player, multiply GP * 5v4 TOI/60 to get their 5v4 TOI. Sum that up for all players, then sum for defensemen. Divide the D number by the total number and multiply by 5 to get the approximate average number of defensemen on the ice. (Not quite perfect because of empty net, but close enough.)

Step 2, Calculate 5v4 D points: For each defenseman, multiply GP * 5v4 TOI/60 * 5v4 Pts/60 / 60 to get their 5v4 Pts. Sum that up for all defensemen.

Step 3, Calculate 5v4 goals: For each player, multiply GP * 5v4 TOI/60 * 5v4 G/60 / 60 to get their 5v4 goals. Sum that up for all players.

Step 4, Calculate overall D IPP: (D Pts) / (Total Goals) / (Avg # of D on ice).

This may be more work than it's worth, since you already showed by lowering the threshold that we are indeed excluding the less-involved players. But if you already have everything in a spreadsheet and ready to go, that's what I'd do.