Photo Credit: © John E. Sokolowski-USA TODAY Sports

Doerrie: An Inside Look at How NHL Teams Decide Which Goalie Plays, and When

Much has been made about the Toronto Maple Leafs goaltending situation thus far. It seems as if we can’t go a single back-to-back without falling into an argument on Twitter — the kind which universally take the form of backhanded slaps directed at Mike Babcock.

So, here’s the sitch: having worked in an NHL office, I’m going to give you a behind the scenes look into what gets looked at when deciding goalie usage — including a study I conducted, who has a say, who makes the final call, and the arguments for each.

When you have a clear number one in Frederik Andersen, you know he’s getting at least 50 of his team’s games — and one of every back to back situation. That part is easy. The debate arises when you have games that aren’t part of a back-to-back situation. In that case, when do you give your workhorse the night off?

When facing a back-to-back, though, the dilemma gets a whole lot tougher. Which game does the starter get? OR, to take it a step further, does he play both?

Advertisement - Continue Reading Below

Here are some facts: Andersen hasn’t played both games of a back to back since the 2016-17 season, where he played three sets of back-to-backs, and lost the second game two of three times (66.7%). He played 66 games that year, another 66 in 2017-18 and then 60 games last season (he was hurt and on pace for ~66), and all anyone could talk about the past few summers was the Andersen was playing too much. He hasn’t played a single back to back in two seasons and it is widely accepted that he plays too much.

To this, I would say: an ideal number of games for a bonafide starter is between 55 & 60. This gives him enough time to rest, but allows him to stay sharp and work to peaking in the playoffs. Andersen had his best statistical season when he played 43 games in 2015-16 (Anaheim), winning the Jennings trophy. It also shouldn’t come as a surprise that it was his best playoff performance, where he posted a .947 SV%, compared to the .915, .896 and .922 he’s posted in his Toronto playoff campaigns.

The Process

For teams that employ extensive information and research departments (the Leafs have the biggest one in the NHL, by the way), the process of looking into a back-to-back begins a few days prior to the game.

The department will first pull information on their opponents, how each goalie has faired against each opponent over their career, how the goalie plays in the first game of a back to back vs. the second game, and then peruse through NHL trends on goalies playing certain games (or both). The different trends would include the workload (whether or not they’re comparable NHL wide), the SV% when playing one or both games, the numbers of shots the team gives up during the second game, if the particular goalie’s play quality drops after facing a certain number of shots (fatigue), and various other things.

Advertisement - Continue Reading Below

The department will evaluate the numbers and make a statistically sound recommendation. However, it doesn’t stop there.

Once the study is complete, it’s usually then sent to the head coach, goalie coach, and management. That group will weigh the presented recommendation and its evidence, they will also weigh other factors such as the psychological impact of a goalie playing one game or both, the way the team plays in front of each goalie, the magnitude of the game, and situational implications.

There will be numerous phone calls between the parties involved, an intense sharing of thoughts, the goalie coach discussing the looming back to back with the goalies, and then will all culminate in each individual conducting their own introspective analysis.

Workload Within Games

In the three seasons Andersen spent in Anaheim, the average number of shots he faced was around 26 per game. In his three-and-a-quarter seasons in Toronto, Andersen has faced 34 shots per game. To illustrate how intense his workload has become, in the 66 games Andersen played in 17-18, he faced double the number of shots as he did in 43 games as a Duck.

This, of course, is taken into consideration when looking at a goalie’s workload and evaluating his usage moving forward. It isn’t how many games he plays, rather, it’s how many shots he faces, how many scoring chances he faces, etc.

Against San Jose, the Leafs only gave up 17 shots, which is a very light workload. However, in 8 of their 13 games this season, that final total has topped 30, surrendering over 38 (!) shots on five occasions. Needless to say, that’s a decidedly heavy workload to put on a goaltender.

Advertisement - Continue Reading Below

On the second half of back-to-backs this season, the Leafs have given up 39, 42 and 38 shots (and 38 vs CBJ in Game 1 of the 2nd set). Therefore, it is reasonable to predict that they will give up more than 35 in their average game two of a back to back.

Facing more than 55 shots in 24 hours is not recommended — it’s a quick recipe to fatigue your goalie, in fact. That’s without even mentioning what happened against Montreal, a game in which the Leafs gave up five breakaways and a fair share two-on-ones, which are more difficult than the in zone scoring chances.

The way the Leafs currently play — which is without much regard for the shot clock or the quality of scoring chances allowed — it is simply unreasonable to expect Andersen can (or should) play two nights in a row.

Day of the Week/Broadcasting

I hate to tell you this, but “it’s Saturday night against the Montreal Canadiens,” weighs less than zero to front offices having this argument.

TV has a lot of pull, sure, but I can unequivocally tell you that broadcasting doesn’t get mentioned in the discussion, never mind actually factor into the final decision. The opponent does, of course, as does the venue (home v. away).

In both my experience and in speaking to others, it would take an extremely compelling argument for the starter to not play at home on the front end of a back-to-back. Regardless of opponent strength, there is an advantage to playing at home, and you want to maximize that as a hockey team. Playing the starter, unless he’s historically awful against the opponent (which does happen), is usually the decision.

Advertisement - Continue Reading Below

I know Leafs fans grumble about their starter not playing on HNIC, but if the team started capitulating to the broadcast, things would go downhill — fast. One sentiment I absolutely agree with is: If you start basing your decisions on what the fans want, you’ll end up joining them (aka you’ll get fired).

More heavily weighted is the opponent, and no, “they are our rivals” doesn’t factor in, either. To be honest, it may come up once or twice, but it doesn’t end up influencing the decision to a meaningful extent.

For example, if the starter is always brutal in your arch rival’s barn, and the backup has a pretty good record against them, you’d probably lean to the backup. In the same way that players have their kryptonite; so do goalies. It would be unwise to put your goalie in a spot where, historically speaking, they haven’t played well, when you could otherwise play them against an opponent where they have had success.

Further to that, if the back-up has success against a particular team, use that confidence to your advantage.

Opponent Factors

Here’s what has bothered me the most this season: “Hutchinson should play Team X because they are weaker and Freddy should play Team Y because they are better.”

Advertisement - Continue Reading Below

Do you watch the NHL? Do you see that Ottawa beats Tampa, LA beats Calgary, and that anyone can beat anyone on any given night? This isn’t the Wild West of the NBA, where the result is more skill-based than randomized. Look at the betting lines. They are closest in hockey (no more than -200), whereas Golden State (NBA) is -800 on any given night. Not to mention, if I look at prediction models, all I tend to see are people yelling about how they aren’t always right. It’s hockey! A puck taking a weird bounce could decide the Stanley Cup.

On any given night, any team can win. That’s what parity is.

This notion that your backup should play a team on one night simply because a team higher in the standings is on the docket the next one is crazy to me. If you get caught thinking that as a staff, your players eventually start to think that the game isn’t as tough and they play that way. I’ve seen it. It’s very apparent when this happens.

Playing more nonchalant than the Leafs already are is a pretty difficult thing to do, so sending a message that one team is weaker than another, and potentially altering the subconscious of the players, could lead to you getting embarrassed in a big way. I’m sure that would go swimmingly in Leafs Nation. You’re all rational people, right?

Going into a back to back, it’s more about how your team is playing, how your goalies are playing, and what their track record against the opponents are. The NHL is a league with too much parity for its teams to be thinking “this is a weaker opponent.” That’s precisely how you get run out of a building.

Advertisement - Continue Reading Below

Playing One or Both

In ‘The Process’ part of this article, I mentioned the magnitude of the game and its situational implications. This is where playing one or both enters into the equation.

From the study I conducted, not once did I recommend that a goalie play both games. Not once. However, the game’s magnitude will sometimes override that recommendation. It only happens, realistically, in the later stages of the season when you’re battling for a playoff spot. For instance, if you’re playing a team that you’re chasing in the standings in March, or one that is just behind you in the standings and closing in fast, then your starter is probably playing that game.

At that point, the recommendation would be to ride your number one option, full stop. If a team finds itself in this situation, it becomes more likely that the starter will play both games.

When the Devils were chasing a playoff spot in 2017/18, Kinkaid had the hot hand. The Devils, then, had him play two sets of back-to-backs, four games in total, all of which he won. And considering how New Jersey barely squeaked into the playoffs, this was an instance where the situational implications at hand overrode the ‘only play one game’ rule of thumb.

On average, the starting goalie plays the front end of the back-to-back more than 80% of the time. On the occasions where the starter didn’t, it was a situation wherein the goalies were splitting games.

Advertisement - Continue Reading Below

This has been the NHL-wide trend for years. The thinking here is, “if the goalie doesn’t have a heavy workload in the first game, we’ll talk to him, and maybe have him play game two as well.” To this, I will say: If the goalie throws up a 30 save shutout, you probably let him have the next game if he wants it. However, anything close or over 35 shots, or if the team gave up more than 20 scoring chances, should be a no-go the next night unless the playoffs are definitively on the line.

The Decision

Considering it’s October, the playoffs are absolutely not on the line for the Leafs — or any team, really. Playing a goalie two straight and risking injury, however, could very quickly put the Leafs in a position where that changes come February.

Had Andersen played against Montreal, for example, I’d argue that the Leafs still would have surrendered three or four goals, most likely ending in a loss. Right now, the Leafs have not shown an ability to consistently play defensively structured hockey, and even less so in the second game of a back-to-back.

Knowing this, it undoubtedly factors into the decision on who to play on the first night.

Sticking to the status quo means the starter plays the first game and the backup always plays the second game. Unless the numbers, situation or conversation with the goalies prove to be so overwhelmingly in favour of a different decision, coaches will stick to the status quo.

Advertisement - Continue Reading Below

Ultimately, after the introspection occurs, the GM will relay his thoughts on the matter to the coaches, and the goalie coach will then give his own opinion to the head coach, and then the relay the gist of that conversation to the goalies.

In my experience, the opinion of the goalie coach always carried the most weight. Goalies understand goalies. Therefore, the goalie coach has a unique relationship with their team’s netminders and will be able to pick up on signs of confidence or fatigue in either player’s mental makeup.

The end of the process is where things tend to differ depending on who happens to be the head coach. If the goalie coach is trusted and very well respected (Roland Melanson/Francois Allaire), then the head coach usually cedes the decision to him, albeit after a discussion wherein the head coach outlines his thinking. The head coach is involved, of course, but he should trust his goalie coach to evaluate the necessary research in order to make a sound decision.

In other organizations, however, the goalie coach relays his opinion and thought process, and the head coach makes the final call.

Regardless of whoever ultimately has the final say, something significant would have to happen in order for the plan to change both during or after game one. For example, if the starter plays game one and gets shelled for four goals in the first 30 mins, you should probably consider giving him the hook and letting him reset for the next night. If the starter throws up a shutout, however, there’s a discussion about whether to run him out again — it would depend on fatigue level and the workload from game one.

Keep in mind, the starter plays the first game unless there is significant evidence to suggest otherwise. The Leafs playing Andersen on night one is the sound decision, because as it stands, there isn’t overwhelming evidence to suggest he shouldn’t.