Could the Rangers be a statistical anomaly?

Kathy Willens/AP

What a strange ride this 2015-2016 season has been so far. After last night’s victory over the eminently talented St. Louis Blues, the Rangers sit just two points behind the mighty Montreal Canadiens for the best record in the entire NHL. However, it has not been a stream of endless adulation and bold championship predictions on Broadway. In fact, most pundits expect the Rangers to take a severe step back due to their unsustainable combination of low possession and high PDO.

Be sure to check out our “Metrics we use” tab for useful information on possession statistics, but PDO is much simpler. Basically it creates a “normalized” statistic by combining shooting percentage and save percentage and theorizing they will both regress to a league average of 100. The theory is that over time, unsustainably high or low shooting and save percentages will regress to the mean, and performance can be predicted to improve or decline as regression takes place. It’s hockey best attempt at quantifying “luck”.

The possession theory stands on a little more solid footing, in my opinion. If you possess the puck and drive shot attempts for, you will score more often than you are scored on. Makes perfect sense. The interesting (or infuriating, depending on where you stand on the issue) is that with these types of statistics in their relative infancy, unexpected statistic outliers and anomalies can be expected to emerge. I am wondering if the New York Rangers are one of those outliers. Let’s break it down a bit.

Before we get started, I just want to add one caveat about the defense. The defense has not been good. This is probably the best argument against the sustainability of the team save percentage, is eventually those goals are going to find the back of the net if the defense continues to perform at this level. While I don’t think I could put up much of a fight against the soundness of this theory, the save percentage and wins have already been banked under these conditions and the assumption is that the defense will not play this bad for the remainder of the season.

Let’s start with possession. Dave has already explored a few instances that have skewed the Rangers overall possession rating. Still, we can tell both from scoring chances statistics and simply watching the game that the Rangers have been on their heels for large stretches during this recent winning streak. The generally accepted possession theory states that if you control the puck and create more shot attempts than your opponent, you are more likely to win the game, or in the long run, be a generally successful club. The question for the Rangers in this equation: do they not possess the puck because they are unable to, or does the system they employ not prioritize possession? The theory is predicated on the fact that all systems are created equally. Dave is doing some very interesting research on in-zone coverages and systems changes on the fly and how they affect the metrics. As of right now, though, it is context neutral.

The Rangers employ a system that prioritizes neutral zone turnovers and defensive zone transition. This is the goal of the defensive zone overload. It is designed to outnumber the offensive players along the boards and move the puck out of the zone. When you calculate in the Rangers’ team speed, this system is effective at creating odd-man rushes. Usually, these types of chances result in less zone time and less shot attempts, but the disparity in bodies during these plays generally results in the better scoring chances and, at least in theory, a higher shooting percentage.

Which brings us to the PDO argument. I’m not a fan of PDO myself, I think it is too general and doesn’t take into account individual team strengths. Just look at the goaltending. Short of Montreal and maybe Boston, I think any team in the league would find Henrik Lundqvist to be an upgrade. This brings me back to my remark about the defense. Right now, Ranger goaltenders are working with a 96.2% even strength save percentage. This number, generally speaking, is not sustainable. However, both Ranger netminders have faced very difficult shots while playing behind a porous defense. Obviously, they have been riding an insane hot streak, but as the defense begins to play up to their talent level and the degree of difficulty on the shots decreases, wouldn’t they still forecast to have a well above-average even strength save percentage given the talent between the pipes? I don’t think a 94+% even strength save percentage is outlandish under the circumstances.

Let’s look at the other half of the PDO equation. The Blueshirts’ shooting percentage currently sits at 11.3%. The league average for the past several seasons has hovered right around 9%. Now, that is a pretty big spread and the 11.3% is probably not sustainable. However neither is their current point pace. As of today’s standings, the Rangers are on pace for 133.38 points. The highest point total for any President’s Trophy winner was 131 points back in 1996, prior to the implementation of the salary cap. Regression is evitable from the video game numbers of the first sixteen games. This does not mean that the Rangers are required to regress back to the league average PDO. We discussed the differing system the Rangers employ on the counter-attack, which would logically indicate less physical shots and a higher shooting percentage because of their shot type and location.

We have relied on positive possession as an indicator for long-term success during this stage of statistical development. This is a 10,000-foot view for trying to predict large samples for thirty franchises. I’m not saying the Rangers have broken the model or anything that extreme, but is it possible that all of these analysts predicting a precipitous fall for the Rangers might want to take a little bit closer look at the practical reasons why their metrics don’t line up. Dave tackled whether the Rangers were smoke and mirrors the other day, but I wanted to take a little deeper look at the factors affecting the theory on a conceptual level. Dave also raised a good point about the huge amount of time the Rangers have spent playing with the lead. Those score-effects have a real impact on possession statistics and high danger scoring chances.

It is entirely possible that there will be serious regression and the prevailing theories are correct. However, I feel that there is tangible evidence both from a personnel and systems standpoint that would indicate a deviation from traditionally accepted parameters for team performance. It is possible we are seeing an alternate concept of roster construction emerge from this group. Only time will tell.