Analyzing Goalie Changes in the NHL Playoffs

The 2016 Stanley Cup playoffs have certainly delivered upon the hype they receive from fans of the NHL each spring. Not only did the St. Louis Blues and San Jose Sharks finally overcome their first round playoff demons in advancing to the Western Conference Finals, but the Tampa Lightning have managed to return to the Eastern Finals despite missing key players throughout their lineup, including superstar forward Steven Stamkos.

And then there has been the goaltending. I quietly made the mistake of thinking that the biggest goaltending story of this postseason would be the implosion Dallas witnessed in their second-round defeat to the Blue. Both conference finals have clearly proven me wrong.

With each series just past their halfway points, no fewer than seven goalies have made starts for the four remaining teams combined. Saying that that number is an abnormality would be an understatement, as I truly can’t think of another season in which championship-caliber teams have had such little confidence in their starting goaltenders. Having a solid number one goaltender was once seen as a necessity for any team hoping to contend for a Stanley Cup, but the disparity this season has received widespread attention, and drawn tweets like the one below from Roberto Luongo.

But if you look at the bigger picture, these two playoff series are merely a microcosm of the league as a whole. Over the past five postseasons, teams have voluntarily switched starting goalies 19 times (not accounting for injury or illness), including six such changes in 2016 alone. This strongly suggests that the trend is a growing one around the NHL, but it doesn’t do much to show the effectiveness of such a dramatic lineup alteration.

Is Changing Goalies Effective?

To answer the question of whether changing goalies in a playoff series is an effective strategy, I looked at each of the last 19 times that such a swap has occurred. Including the 2016 postseason (and the recent swaps in St. Louis and Pittsburgh), the initial starting goalies for these teams had a 35-44 record at the time of the switch and averaged a .904 5v5 AdjSv% in their time on ice.1

Their replacements compiled a 28-31 record with an average 5v5 AdjSv% of .927, which initially seems to support the strategy of switching up goalies in a playoff series. However, there are several factors that these hard numbers do not account for when determining whether or not a team makes the right move in changing goaltenders.

Replacement goalies performed slightly better than initial starters after a change.

Replacement goalies performed slightly better than initial starters after a change.

For instance, the mere fact that a team was compelled to change goalies in the middle of a postseason series means that one of three things probably happened: either the initial starter played poorly, the team was in an early series hole, or both. With that in mind, it isn’t fair to say that just because the replacement starter posted better numbers than the original starter, that it was the right decision to take the first player out of the net. The .904 AdjSv% averaged by those initial starting goalies is evidence enough that the standard their replacement would need to exceed is fairly low. Those who read my article on the goaltender struggles in Dallas would know that a league average backup posted a .920 AdjSv% in 2015-16, so even having average goaltending would be an improvement over the initial starter’s performance in their allotted games.

Additionally, while the replacement goalies objectively posted better numbers than the initial starters did in each of their opportunities, we cannot say for certain that the initial starters would not have posted just as good, if not better, numbers than their replacements, had they been given the opportunity. After all, the playoffs are an incredibly small sample size to use when evaluating goalies, and it isn’t fair or accurate to say that a goalie will post poor numbers for an entire postseason just because they did so for the first few games of a series.

Finally, it is worth noting that six of the 19 teams who did pull their initial starter in favor of a new goalie, ended up going back to their initial starter after a few games. Notably, the 2014 Wild, 2015 Blackhawks, and 2016 Stars all made it past the first round of the playoffs while flipping between starting goalies – with the Blackhawks managing to win the Stanley Cup after returning Corey Crawford to their crease. Joel Quenneville made the decision to return to Crawford even though Scott Darling actually outperformed his counterpart in Chicago during the early stages of the team’s first round victory over Nashville, which serves as an interesting example for coaches thinking about making a goalie change.

Jun 15, 2015; Chicago, IL, USA; Chicago Blackhawks goalie Corey Crawford hoists the Stanley Cup after defeating the Tampa Bay Lightning in game six of the 2015 Stanley Cup Final at United Center. Mandatory Credit: Dennis Wierzbicki-USA TODAY Sports

Corey Crawford ended up with a Cup, even after being benched. Mandatory Credit: Dennis Wierzbicki-USA TODAY Sports

That change (and subsequent change back to Crawford) shows the value in playing the goalie that gives the team the best chance to win on a given night. The postseason is all about results, so it can make sense to give a quality backup a start if the team is in a hole and the starter’s performance has been subpar. But at the end of the day, the goalie with the best numbers over a larger sample size is going to be the goalie that is most likely to deliver a quality start when each game is looked at independently.

2016 Playoff Goalie Changes

Even the most recent example of a starting goalie change that enjoyed great success could have been predicted before the switch was made. The move I’m referring to is Pittsburgh’s decision to pull Marc-Andre Fleury in favor of Tomas Vokoun in 2013. The Penguins advanced to the Eastern Conference Finals after the switch, and the move might be a contributing factor to the recent surge in postseason goalie changes, but Vokoun’s success shouldn’t have been a revelation.

A comparison of Vokoun and Fleury from 2008-2013 via war-on-ice.com

A comparison of Vokoun and Fleury from 2008-2013 via war-on-ice.com

Not only did the veteran post a better AdjSv% than Fleury during the 2012-13 regular season, but his numbers had actually been superior to Fleury’s for years leading up to the change. It simply made no sense for Fleury to be starting initially, and Vokoun’s relative success after becoming the starter was something that easily could have been predicted.

Meanwhile, Brian Elliott’s superior numbers and playoff heroics make his replacement with Jake Allen appear all the more confusing. Fleury’s replacement of Matt Murray falls into a similar category as both teams and goalies were enjoying great success before each benching. Allen and Fleury could prove me wrong, but starting these two goalies seems like a colossal mistake for each team to make at this point in the postseason.

Overall, goalie changes are extremely tough to evaluate in the postseason mainly because it’s impossible to determine how one goalie would have played had they been in net for a given game. Chances are that coaches will have plenty more opportunities to examine this issue in the coming seasons too, as the increased popularity of the goaltender-by-committee approach has decreased the gap between starter and backup on many NHL rosters.

With all of this in mind, keep a close eye on the performances of Marc-Andre Fleury and Jake Allen in the coming days and judge for yourself if they really are the best options for their respective teams in net.


  1. numbers from war-on-ice.com 

About the author /


David is a sport management student at UMass Amherst who one day hopes to work in athlete representation. Keep up to date with his writing and other interesting sports news by following him on Twitter via @DavidTews13.