Are Neutral Zone Regroups Beneficial To Power Play Units?

I haven't made a secret of the fact that I am in favor of more structure when it comes to the power play. One of my early claims based on watching units in action regarded the neutral zone regroup, when the puck barely squeaks out of the offensive zone and teams immediately try and re-enter. The benefits of such attempts are obvious: They save valuable seconds, they can force tired penalty killers to remain on the ice, and sometimes they can even catch opponents napping on the break. From watching, however, it seemed as though these entry attempts were less successful than full regroups because of the lack of structure involved whenever players try and rush something. My feeling was that we were seeing a lot of this....

...when we should have been seeing more of this....

And the early data backed up what my eyes were seeing. In the first chunk of the season, neutral zone regroups were leading to far more failed entries and dump-ins than full regroups, and on aggregate they appeared to be the inferior play. But of course, it was early, which is why I waited until this week to write up the numbers on this critical tactical question. So now that I have tracked more than 300 team games-worth of power plays, what do the numbers suggest about the neutral zone regroup?

First, in order to avoid subjectivity, I had to carefully define it. A neutral zone regroup, I decided, would be a regroup where either a) the puck didn't enter the power play team's defensive zone before an entry attempt took place, or b) it left the zone briefly, but no players other than the puck carrier complete a curl to generate speed and head up ice. I also decided to create a third category, which is pressured defensive zone breakouts. If the penalty killing team pressures the regroup either following a shorthanded chance or just with a super aggressive forecheck, to the extent that the player breaking out of the zone is trailed by an opponent, I termed that a "pressured" breakout rather than a full regroup. These pressured breakouts tend to be more improvised but have less penalty killers to get through so should theoretically be more successful. My early season results didn't account for pressured regroups, which may have skewed the results in favor of full regroups to an extent.

So how should we measure effectiveness of a regroup? The first possibility is seconds in the offensive zone. A failed entry leads to zero seconds of zone time, and while what happens once setup can be independent of the entry, generally a good entry will lead to more seconds in the zone than a poor one, even if the latter technically crosses the blue line. So divided into controlled entry attempts and dump-ins -- and based on data from the six teams I've been tracking -- I looked at how neutral zone regroups, full regroups and pressured regroups compare in terms of average offensive zone seconds off of each attempted entry

Interestingly, dump-ins from neutral zone regroups lead to more zone time. Maybe this is a result of puck pursuers already being at the blue line ready to chase, but the differential is pretty small, so it could be simple variance. The sample size for pressured dump-ins is practically zero, and there's a reason for that. If one manages to evade forecheckers and create a rush up ice, dumping the puck is a waste. There seems to be almost no difference in success of controlled entries in terms of o-zone seconds. Zone time measurements aren't great though, because spending time fighting for pucks along the wall, while advantageous at even strength because you're in essence preventing goals against while trying to generate offense, on the power play every second counts. So let's look at shot attempts per entry attempt instead, which should paint a slightly more accurate picture. Note that all shot numbers on these graphs are shot attempts.

Here we see the bigger difference in effectiveness between controlled entries and dump-ins that one would expect, and note that full regroups do tend to lead to more shots per attempt than neutral zone regroups. The difference is notable, but modest, at 0.81 compared to 0.68. One factor that this graph doesn't account for, though, is the fact that full regroups and obviously pressured regroups lead to a higher percentage of controlled entries compared to dump-ins. Many teams when the puck squeaks out of the offensive zone will either immediately fire it back in or will look to find an open man and if nobody emerges will resort to a dump-in. On full regroups, teams attempt controlled entries 73 percent of the time, compared to 67 percent of the time for neutral zone regroups (all of this data is 5-on-4). So the next step is to combine controlled entries and dump-ins to see overall shots per entry attempt.

Even when accounting for the dump-ins, the shot difference between full and neutral zone regroups isn't too large, now (coincidentally) 0.73 to 0.67. So when a team attempts a neutral zone regroup, the expectation is slightly less shot attempts -- in other words, slightly less likely to complete a successful entry. But teams don't attempt neutral zone regroups on the power play because they think they're the most successful, they do it to save time. So the critical next question is, how much time do they save?

Turns out the answer, over the decently large sample I now have, is "a lot". The average neutral zone regroup try takes 5.27 seconds, whereas the average full regroup costs 14.25 seconds. These are attempt times; the number of seconds it takes to get the puck back into the offensive zone from the neutral zone is closer to double digit seconds, but trying to account for that here would mean double-counting from the shot success-based evaluation above. So it appears as though by and large neutral zone regroups are worthwhile. Just how many extra shot attempts do they lead to on average? Well it's difficult comparing a shot-based metric with a second-based metric, but from the teams I've tracked this year, we know that teams get 0.05 shot attempts per second in the offensive zone at 5-on-4. So saving approximately nine seconds on a regroup leads to 0.45 more shot attempts per entry attempt for the neutral zone regroup. It's not a perfect science, but that would mean that instead of a 0.73 to 0.67 shot attempt average in favor of full regroups, the pendulum swings in the other direction to the tune of a 1.12 to 0.73 edge. So a neutral zone regroup try leads to on average 0.39 more shot attempts than a full regroup.

Now there are a few necessary caveats. I don't believe that shot attempts are a great way to measure power play success, as evidenced by the low correlation between shot attempt numbers and power play success the last few years. Unlike at even strength, where a shot that is blocked is at least evidence of driving play in the right direction and limiting chances against, on the power play getting a point shot blocked isn't a very difficult task and can do more harm than good.

I also think that structured regroups have advantages that aren't presented here. For example, a well-designed structured regroup leads to players entering the zone in positions where it takes very few seconds to get into the desired formation. I don't have enough data yet to confirm it, but my belief is that neutral zone regroups can mean teams taking longer to get into formation once in the zone, which could swing the equation back in the other direction.

I also have numbers on a team-by-team level, to get an idea of if some teams do better or worse using neutral zone regroups. 

The teams in this chart are ordered from biggest to smallest difference between shots off full regroups and shots off neutral zone regroups. So while the Capitals and Canadiens seem to be justified trying to rapidly return to the offensive zone, teams like the Flyers and Leafs might be more justified in completely regrouping. The Lightning, as you can see, have struggled with any kind of entry all season, so a little more structure couldn't hurt, something they may well be putting more emphasis on going forward.

Above are the average regroup times for each of the teams, sorted from smallest to largest gap between full and neutral zone times. The Islanders take a long time to attempt neutral zone regroups, while the Capitals are very efficient with them, clocking in at a mere 4.5 seconds. That's what players like Evgeny Kuznetsov, Alex Ovechkin, Nicklas Backstrom -- guys who are incredible at stickhandling in tight spaces -- can give you. The Leafs might be a good example of a team that takes its time to get a preferred, fairly structured breakout started, using up more seconds to ensure a successful entry and thus more shot attempts. Despite the lack of success for the unit as a whole, it may not be a bad model.

So what's the takeaway? Contrary to my early belief, neutral zone regroups appear to be best practice in a lot of cases. Certainly in situations where there is an immediate easy route back into the zone -- whether a clear skating path or passing lane wide open -- it makes sense to take that option. If the other team is going for a line change, pushing the pace and hoping for a quality rush chance is also probably worthwhile. Beyond that, I'm still somewhat skeptical than an improvised congested neutral zone regroup without an option easily available is preferable to a well-crafted set play (and the obvious compromise is a well-designed crossing pattern regroup that I've seen some teams occasionally use). With additional practice time devoted to the power play, I think full regroup rates could be boosted to the point where only with obvious openings are neutral zone regroups justified. That said, until that day comes, it seems teams are doing the right thing and saving those nine seconds. That extra shot generated every three neutral zone regroups can make all the difference.