Throughout the NHL playoffs, I had been collecting my own data for every team, analyzing their power plays and the defending penalty kills. And while I gathered much more data on different events, one of the many things I tracked was whether or not the power play unit deployed three forwards or four forwards. Matt Cane has already done a lot of work over at Hockey Graphs looking at the differences and possible benefits of using four forwards instead of three on the power play and he came to the conclusion that four forward units generate more offense than three forward units. I looked to confirm his results with my own data and ended up finding some interesting results.
Before I get into the data and results, I would like to put in a bit of a disclaimer. While I have tracked about 1,000 entries, which amounts to a little over 40 games in the playoffs, sample size is a bit of a concern. And after discussing a bit with Ryan Stimson, there is the possibility that the three forward unit performed so well because all the entries were taken from the playoffs. As he says, there “might be a bias based on how good [the playoff team’s] dman are.” I hope to continue tracking these events and eventually add regular season data too.
But with that, I still think the data holds value and merits discussion, especially considering that (to my knowledge) I am the only person to track zone entry success based on formation. If interested, I have a much more statistical analysis here if you’d like to see the math behind my research but this article is going to focus much more on the results and possible conclusions of the differences between four forward and three forward units.
I started the playoff data project to try to get a better look at the NHL teams playing in the playoffs and how they were performing. I had always been fascinated with special teams and, since there’s been little research into the subject, I decided to go about collecting my own data. For almost every playoff game, I analyzed every power play and tracked a variety of different statistics. Besides a lot of other things, I looked at whether the attacking team deployed four forwards or three on their power play and the ensuing results. To determine which unit performed better, I looked at their success at entering the zone and how many shots, shot attempts, scoring chances, and goals they generated.
Overall, the four forward unit did perform slightly better than the three forward unit in a few metrics such as shots, shot attempts, and scoring chances. However, these differences were not statistically significant in my test that I conducted. I originally started tracking this to try to confirm the idea that four forward power play units are better than three forward units. One would think that swapping a defenseman for a forward should lead to more skill and offense on the ice and, while I still believe that, now I am a bit more skeptical of the magnitude of the impact. I think a main reason for this lack of significant difference in my study could potentially be sample size issues but also the possibility that the effects of adding an extra forward aren’t as impactful as initially thought. Hopefully as I continue collecting more data I can revisit this study and analyze the results, but from what I have, it seems that a four forward power play unit is only slightly more effective than a three forward unit, though I would still recommend a four forward power play unit as opposed to a three power play unit.
But besides shots, shot attempts, and scoring chances, there was one other metric tracked that was unique to my study and deserves analysis: zone entry success. Because zone entries are not recorded by the NHL, Matt Cane was not able to observe the differences between four forward and three forward units at entering the zone in his research. In my data, I found that three forward units actually performed slightly better than four forward units at successfully entering the zone on the power play. Three forward power play units were successful at entering the zone 91% of the time while four forward units were only successful 88% of the time. It’s far from a huge difference but its notable that the four forward units didn’t come anywhere close to outperforming three forward units in successfully entering the zone.
Originally I thought that, since four forward units would probably be more likely to have more skilled players, they would have greater success at entering the zone than three forward units. However, since the initial data proved otherwise, I decided to take a deeper look at some of the differences between the units to see why this might be.
My hypothesis is that four forward power play units generally try more difficult and skilled entries into the offensive zone which result in a slightly lower success percentage while three forward units are more conservative and utilize easier entry strategies. Teams might load up their skilled players on four forward units and this can possibly leave the less talented players to be assigned to the second unit (which is sometimes the three forward unit if a team employs both a four forward unit and a three forward unit). With the most talented players potentially on a four forward unit, it’s likely that they’ll try higher-risk or more complex plays and have the “big play” mentality to enter the zone (these two entry attempts by Alex Ovechkin come to mind).
This potentially shows itself in the passing data I collected before each entry, as I recorded the number of passes completed before every entry attempt. The four forward power play units attempted to enter the zone more often than three forward units with both zero passes before the attempt and 3+ passes before the entry. 21% of all four forward entries were comprised of attempts where the puck carrier single-handedly tried to force the puck into the zone by himself without passing to another player, where it was only 17% for three forward units. The four forward units also had a slightly higher percentage of entry attempts formed by connecting three or more passes before the actual entry.
But the most telling thing from the data was how much likelier the three forward units were to dump the puck into the zone than the four forward units. While the four forward units only dumped the puck in 14.5% of the time, the three forward units dumped the puck into the zone 32.7% of the time. As we all know, it is much easier and simpler to dump the puck in rather than to attempt to carry the puck in with possession, and this helps explain why the three forward units were able to successfully enter the zone more often than the four forward units. While overall the three forward units had a higher success rate, the quality of those entries was significantly lower, as the four forward units were trying more complex and harder entries compared to the three forward units.
Take the Ottawa Senators for example. When they deployed four forwards, they only entered the zone successfully 74.3% of the time. However, when they used three forwards, that number shot up to 95.5%. Why? Well a major reason for this was because the three forward unit was much more likely to simply dump the puck and get in the zone, dumping the puck 40.1% of the time compared to 21.7% of the time for the four forward unit. The Senators four forward unit also averaged less passes before entries and had more entry attempts without passing the puck once before entering the zone, with 29% of their entries were comprised of attempts where the puck carrier single-handedly tried to force the puck into the zone by himself. In comparison, only 18.2% of the three forward unit’s entries were like this; they were more willing to pass the puck to try to enter the zone instead of force the puck in with individual skill.
It would make sense that Ottawa’s four forward unit would try to execute these more skilled plays, especially when considering how involved Erik Karlsson was in the four forward unit compared to the three forward unit. Out of all the four forward unit’s entries, Karlsson was the one entering the zone 30% of the time and was the player either entering the zone or the player who passed it to the player entering the zone 50% of the time. With a guy as talented as Karlsson directly involved in half of your entry attempts, you can be sure that he’s not just looking for the easy, simple way to get into the zone. However, for the Senators three forward units, Karlsson did not enter the zone once. Instead, guys like Clark MacArthur and Zack Smith were tasked with entering the zone.
So overall, what can I say about four forward power plays units? Well, like past research has shown, there is some increased offensive output from four forward units in comparison to three forward units. How significant the difference is is up for debate, but there is a difference, and this aligns with what you would initially expect. Forwards are more offensively skilled than defensemen so replacing a defenseman with a forward should lead to more offensive.
However, what was found in my data was that four forward units were not more successful at entering the offensive zone than three forward units. In fact, the opposite was true, with three forward units being slightly more successful. But digging deeper into the data, it seemed that the quality of these entries was significantly less than the four forward unit’s entries. The three forward units may have been able to successfully enter the zone more often but they accomplished this by dumping the puck into the zone significantly more often than the four forward units and generally executing safer, simpler plays. The added forward in the four forward unit may allow for the team to be more creative and attempt more complicated entries than they would with only three forwards on the ice.
I plan on continuing to gather data so I’ll probably revisit this topic again at some point in the future with more data, but for now, use four forwards on the power play.