Andrew MacDonald's 2015-16 NHL season appeared over before it began.
With the Philadelphia Flyers facing a logjam of eight defensemen on the roster at the conclusion of training camp, general manager Ron Hextall made the tough choice and sent MacDonald and his six-year, $30 million contract down to the AHL. The reasoning was obvious -- a spot had to be cleared, and Hextall rightfully believed that no team would pluck MacDonald off the waivers that were necessary for re-assignment to the Lehigh Valley Phantoms. He could not say the same with certainty about other options like Luke Schenn and Radko Gudas.
So MacDonald accepted his new role as first-pair AHL defenseman, apparently serving as a mentor of sorts to the young blueliners in Lehigh Valley, and waited for his chance to prove himself again at the NHL level. He was briefly called up in December after a Luke Schenn injury combined with a Gudas suspension forced the Flyers' hand, but the stint lasted just one game before he was returned to the Phantoms.
Even after Schenn was traded in early January, MacDonald stayed in the minors, a victim of the emergence of Shayne Gostisbehere in Philadelphia. It looked like the 29-year old defenseman was destined to spend the entire year with Lehigh Valley.
Then, Michael Del Zotto suffered a season-ending wrist injury, and the Andrew MacDonald comeback tour began.
|Contract Status||Signed through 2019-20 for $5,000,000 per year|
2015-16 Regular Season Numbers:
|Total||5-on-5||Power Play||Penalty Kill||Other|
|Corsi For %||Corsi Rel %||Goals For %||PDO|
|Points Per 60 Minutes||Penalty Differential||Scoring Chances Per 60||Shots On Goal Per 60||Shot Attempts Per 60||Offensive Zone Starts||Defensive Zone Starts|
|Controlled Entry Percentage||Entries Per 60||On-Ice Entries % For||Neutral Zone Score|
Timing and MacDonald
Let's start this exercise with two undeniably true statements:
1. Andrew MacDonald was called up on February 13, 2016 to replace Michael Del Zotto, and started all but one game for the Flyers following his return.
2. The Flyers played their best hockey of the season starting in mid-February and lasting through the conclusion of the regular season.
Considering the fact that both of these statements are true, one could come to the understandable conclusion that adding MacDonald to the Philadelphia lineup was a major contributing factor to the team's late-season surge. After all, the timing was impeccable. The Flyers posted a 17-7-4 record following the MacDonald call up, after starting the year an underwhelming 24-20-10.
But it wasn't just wins and losses. The Flyers went from being a poor team by 5-on-5 advanced stats prior to the MacDonald recall (48.6% score-adjusted Corsi) to a legitimately strong one from February 14th through the end of the season (53.1%). Philadelphia wasn't just getting lucky -- they were legitimately playing great even strength hockey.
It's easy to see how this lays the foundation for a narrative celebrating the reborn MacDonald, who used his time in the AHL to regain confidence, address weaknesses in his game, and become a key part of a revitalized Flyers team. The trouble is, a deep dive into the numbers doesn't support that conclusion at all.
Let's begin by checking the performance of each member of the Flyers' regular defense during that time period from an on-ice shot attempt standpoint. If MacDonald was a primary driver of the team's improvement at 5-on-5, it should show up here.
|Defenseman||Score-Adjusted Corsi after 2/14|
Now we have a clearer idea of what really happened from a puck possession standpoint. Manning took a gigantic leap forward in efficiency, Gudas continued to post strong numbers and Streit moved into respectable territory after struggling with injuries previously. MacDonald, on the other hand, posted the worst score-adjusted Corsi of any Flyers defenseman after being recalled from the minors.
Now, a 48.9% score-adjusted Corsi is far from terrible in a vacuum. It's close enough to 50% (break-even) that MacDonald was not a major liability in terms of driving play in the grand scheme of things, even if he likely contributed to Shayne Gostisbehere's advanced statistical decline. But the Flyers undeniably did much worse in the shot attempts battle with MacDonald on the ice versus the time he spent on the bench.
When MacDonald took a seat, Philadelphia drove play to the tune of a 53.4% score-adjusted Corsi, giving the defenseman a negative-4.5% Corsi relative to his teammates. The Flyers may not have been terrible from a possession standpoint with MacDonald on the ice, but they were significantly better when he was off it.
So where did the perception that MacDonald was a major contributor to the team's success originate? Was it just as simple as timing and the existence of an attractive narrative?
On-ice goals drove positive perception
This chart sheds some light on why so many were impressed by MacDonald's play during the stretch run.
|Defenseman||On-Ice 5v5 Goals For||On-Ice 5v5 Goals Against||5v5 Goals For Percentage|
The Flyers scored a lot more goals than they allowed with Andrew MacDonald on the ice during the stretch run towards the postseason. In fact, he posted the best 5v5 Goals For Percentage of any Flyers defenseman during the time period. Possession stalwarts Gudas and Manning surprisingly brought up the rear.
Now, some will say that this chart simply showcases the limitations of Corsi. MacDonald helped to create goals and prevent the other team from scoring, they'd likely say. Who cares if his Corsi was bad?
The prevailing research holds that Corsi and similar metrics are better predictors of future Goals For percentage than actual Goals For percentage, which makes intuitive sense because goals happen infrequently and therefore make for smaller, more unreliable samples than those that include all shot attempts. In addition, just by the nature of goals, not every player on the ice is responsible for their creation or prevention on a given play. The same thing could be said about shot attempts, of course, but the larger sample size increases the likelihood that one player had an impact on the creation/prevention of at least some of them.
But rather than reference dry research, let's just think back to the last time that on-ice goals were used to judge the effectiveness of Andrew MacDonald.
The year was 2014. The Flyers had traded two draft picks to acquire MacDonald, and the early reviews were positive. One experienced hockey writer even wrote, "The strength of the Flyer D... is a third pair that Andrew MacDonald has made as good as the first pair." That's not meant to call out Jay Greenberg -- a legendary writer in Philadelphia -- and more to note the prevailing sentiment surrounding MacDonald at the time.
Flyers GM Paul Holmgren bought in, giving MacDonald a six-year, $30 million contract that has quickly turned into an albatross on the team's salary cap situation. But what drove positive opinions of the defenseman? One theory is that the team did a stellar job of goal prevention with MacDonald on the ice during that final month of the 2013-14 regular season. The Flyers gave up just nine goals in over 342 minutes of 5v5 ice time with MacDonald playing, feeding into his preexisting reputation as a solid puck mover strong in his own zone.
The problem was that MacDonald's goal prevention was built on a 94.8% save percentage at 5-on-5 while he was on the ice, an obviously unsustainable mark since the best full-season 5v5 save percentage delivered by one goalie over the past five NHL seasons was 94.4%. Given a larger sample, it was inevitable that MacDonald's on-ice save percentage would drop to more reasonable levels, and that's exactly what happened the following season. The goals against increased, MacDonald's past fans began to turn on him, and the Flyers implicitly admitted that the contract was a mistake by sending the defenseman down to the AHL to start the 2015-16 season.
The lesson here? Don't fall into the same trap this year as the Flyers did in 2014 in their evaluation of Andrew MacDonald. The underlying numbers (particularly relative to his teammates) imply that he's basically the same guy who couldn't make the Philadelphia roster out of training camp.
Understanding MacDonald's skillset and weaknesses
Shot attempt and goal metrics present a high-level view of the performance of the team with a certain player on the ice. But they do little to describe the direct impact of that player on the direction of the game. That's where manually-tracked individual metrics like zone entries and exits can help to paint a more accurate picture of a player's strengths and weaknesses.
The most widely-discussed weakness in MacDonald's game has been his supposed passive play in the neutral zone. At Broad Street Hockey, we've evaluated this aspect of his skillset on a number of occasions, and always came to the same conclusion -- Andrew MacDonald gives up too much space while defending the middle of the ice, allowing opponents to get in on the attack with possession of the puck far too often.
That didn't change in 2015-16. On the Flyers, only Nick Schultz allowed a higher percentage of controlled entries when specifically targeted by an oncoming opponent.
|Defenseman||5v5 Controlled Entry Percentage Allowed|
|Michael Del Zotto||53.18%|
But even defenders of MacDonald usually acknowledge that his gap control isn't exactly stellar. They instead point to MacDonald's skills with the puck as his biggest asset, specifically his ability to generate efficient defensive zone exits.
However, that claim doesn't really hold up under closer inspection. Last season, MacDonald sat right around Luke Schenn in terms of zone exit efficiency, and yet again that proved to be the case in 2015-16.
|Defenseman||Percent of Controlled Exits out of all Successful Exits||Percent of Controlled Exits out of All Exit Attempts|
|Michael Del Zotto||61.50%||48.63%|
Luke Schenn's puck skills were always a bit underrated, but he's rightfully not making anyone's list of above-average NHL puck-moving defensemen. That's Andrew MacDonald's closest comparable, as the idea that MacDonald is in the class of true puck-movers like Streit and Del Zotto just isn't backed up by the facts. He's not terrible with the puck on his stick, but he's still far from elite.
Looking at this data, it's really not a shock that Andrew MacDonald grades out poorly by advanced metrics like Corsi. His passive play in the neutral zone allows opponents to generate controlled offensive zone entries rather than dump-ins, and entries with possession have been proven to create more shots than uncontrolled ones. In addition, MacDonald's inability to be much more than average in transitioning the puck from defense to offense (via zone exits) limits the amount of shots his team can create with him on the ice.
We're left with a poor neutral zone defender and an average puck-mover. That doesn't seem like a recipe for an effective defenseman.
In 2015-16, Andrew MacDonald battled through his most difficult professional hockey season yet. Sent down to the AHL at the start of the season, he was forced to accept a minor league lifestyle after years of earning his place at the highest level of the sport. In addition, he was expected to function as a mentor to young defensemen like Shayne Gostisbehere, Robert Hagg and Samuel Morin, while trying to hide his understandable disappointment and frustration regarding his new assignment.
But MacDonald kept his head down, did his job, and finally got the call in February to replace an injured Michael Del Zotto on the Flyers. He was further rewarded by the team's surprise run to the playoffs, where MacDonald had a chance to face off against Alexander Ovechkin and the powerhouse Washington Capitals. It seemed a just reward for the hours of thankless work down in the minors.
Some theorized that it was the addition of MacDonald to the Flyers' lineup that helped them to start their playoff push. But MacDonald remained the same poor play driver relative to his teammates as he had been over the previous four seasons. The 29-year old defenseman did post strong on-ice goal metrics, but those are unlikely to be sustainable moving forward if his overall on-ice attempt differentials stay poor.
The idea that MacDonald showed real signs of improvement over the final two months of the 2015-16 is based on the same questionable evidence that very well may have led to the ill-advised six-year, $30 million extension in the first place. The truth is, Andrew MacDonald is the same player today as he was the day that the Flyers traded for him back in 2014 -- a defenseman with measurable weaknesses lacking sufficient plus skills to overcome those flaws.
Player Card (via hockeyviz.com):