For the second part of the goalie series, today, I want to tackle goalie rest! No, wait, come back, this will be different and interesting (or your money back) and at no time will I use a term with the initials L.M., I promise.
How many games should a starter get?
- One goalie can’t play 82 games.
- Finding a good tandem you can afford to play 41 each is nearly impossible.
- Somewhere in between 41 and 82 is the magic number of starts.
- Jordan Binnington played in 32 regular season games for the Blues, and won the Cup, playing another 26 in the playoffs, so that’s the precise number that is magic.
That’s the logic train that governs most good faith discussions on goalie rest. The bad faith ones begin and end with: Five games less than whatever Babcock does is right. I’ll go along with points one and two, raise an eyebrow at the idea that this question can be answered in a single number of games at point three, and laugh rudely at point four. Even if you remember to add in Binnington’s 16 AHL games played and get a number of 48 regular season games played, you still haven’t discovered an answer. Not even for Binnington himself.
Try this one:
- Goalies experience hockey in unique ways.
- They are susceptible to repetitive motion injuries of the knees, hips and connective tissues.
- Each body is unique, and unusual events can happen.
- It’s really hard to plan in advance the total workload a starting goalie can sustain to maintain peak performance for the playoffs and still get the team into the playoffs by maximizing the value of his starter’s contribution to wins.
That’s wishy-washy and unsatisfying. There is no villain. There is no single number answer that can be tweeted out with disapproval if the team violates this hard border. It can’t be talked about in the TV broadcast. It’s nothing like the pitch count limits for a baseball pitcher and basically is no fun.
I think that means I’m on to something.
But maybe it is like baseball pitchers, by which I mean just as poorly understood or managed. In major league baseball today, starting pitchers often have hard limits of pitches they’re allowed to throw before they’re pulled by the manager. But is that about limiting injury? Or is it about maximizing the value of the pitches in a situation where the batters gradually get better at hitting a pitcher on any given day as they see more pitches? Or is the goal the latter, and the former is a side effect? No one seems to know the answer.
There is an entirely different school of thought about pitching in Japan:
“Guys throw a lot here. Like, A LOT. Guys long toss pretty much everyday during batting practice. All bullpen pitchers throw bullpens scattered throughout the game, starting in the second inning. If you are a non set up guy or a young pitcher there really isn’t even an option.
“In spring training, pitchers routinely throw 100 pitch bullpens. In fact I always found it odd that the first question reporters would ask in spring training is for a pitch count on my bullpen. Definitely a change from MLB organizations where teams tend to stress ‘saving your bullets’ and quality over quantity.”
Meanwhile the number of days pitchers spend on the disabled list is rising and the number having surgery is also going up. It’s not clear why that’s happening, and assuming there is an actual rise in injury rates is likely unwise. Increases in diagnostic rates can have many causes.
Little League Baseball has hard pitch count rules for kids of various ages, and that definitely is an injury prevention scheme. They seem to believe that managers and parents need clear rules because just saying you should limit repetitive activities isn’t enough. We know this in the working world where, for example, grocery stores have shown reluctance to rotate their cashiers to other tasks to limit their injury risk until unions or health and safety laws force them to. The cost of replacing an injured worker alone isn’t enough to get managers to think beyond their short-term needs and their traditional practices.
What do we even know about goalie workload?
It’s likely naive to think that hockey is any different, and yet, unlike baseball, there are no rules or guidelines or even much study around the workload, the real workload, of goalies.
A less naive way to study workload effects than looking at minutes or games played is to look at shots faced.
re LTR. this summer at the Global Goaltending Retreat i shared some goalie workload insights. i haven't released the whole thing yet but playoff performance declined as a function of more shots against in the regular season (adjusted for rest). pic.twitter.com/1TqdecRUYY— Cole Anderson (@ice_cole_data) September 12, 2019
That kind of analysis indicates that workload of goalies impacts performance in the playoffs, although it’s scaled well below ability in terms of impact. There are very few goalies in the NHL so it’s hard to trust the results won’t be swamped with randomness, and the goalies who never make the playoffs aren’t measured. But beyond that, shots faced is not necessarily a complete indicator of a goalie’s workload.
Have you ever watched the goalie when he’s not making saves? Generally speaking, you don’t do that, you watch the puck and the players around it. But two years ago, I watched every minute of Frederik Andersen’s October and November games when the puck was in the Leafs zone. So, while he might have been standing and leaning on his stick while the puck was elsewhere, once it’s in his zone, he’s going up and down, in and out of VH, rVH and full butterfly all the time when not facing any shots at all. It isn’t the saving of the shot that’s the counter of his workload, it’s the body motion. So a deeper analysis might tighten up our understanding of how much work each goalie is really putting in, and how much is too much.
We aren’t there yet. This information does not exist, so teams have to use their own sports science department’s judgement about their starter, their backup and how to best use them. And you can’t just go get a tandem, a better backup, or even a just good enough backup as easily as is it to say it.
If you can figure out how to structure your team in terms of skill and cap space so that you can play with a league average or thereabouts goalie every night and still win, you can relax on the issue of number of games you start your pair of goalies. But for 15 or so teams in the NHL every season who have a starter who is significantly improving their win rate — and no one player has a bigger impact on winning than that sort of goalie — diluting his impact by sitting him for that average backup is a tough sell to any management team.
If a team has very good data on how much their starter’s performance degrades from too much work, and if they really understand the risk of injury associated with workload, they can calculate exactly how much it’s worth it to them and their playoff hopes to sit their starter. Assuming they also really understand their backup, his range of expected performance, and the chances of getting an average year out of him in 20 to 30 starts.
No team has that.
I think there’s a mental model at play with fans and media around workload and the cause and effect relationship with future performance that is unreasonably simplistic. It’s entirely possible this simple fantasy is shared by general managers and coaches too. But a starting goalie isn’t a video game character with a mana bar that drops every time he casts a spell. He’s not going to keel over powerless when it gets to zero. The effects of workload are real, but not an on/off switch, and you can’t point at a playoff loss or win and shout, “Ah ha!” Well, you can, but you’re being absurd, and perhaps this is the real goalie dilemma. The issue is much more complex than “go over x games and you’ll lose in the first round”, and there’s no indication that workload is as extreme a factor affecting performance as the certainty with which this gets discussed would lead you to believe. There’s no way to really tell what’s physical and what’s mental about the effect either.
We’re right back at that first logic train, stuck at item number 3. Somewhere between 41 and 82 is likely the perfect number of starts. And everyone is guessing at it. So, okay, say you take your starter’s historical performance, your backup’s and you calculate out what you can live with for your regular season program of rotation. You’ve done the best you can, and now you’re going to just play the games. If someone starts performing way off his average, you’ll rethink it. That’s all you can do. Make the best guess you can.
Other ways to reduce workload
What about the practices?
It’s not facing the shot, it’s the repetitive motion that’s the problem. And it doesn’t matter if that motion is in a game, in practice, or in the gym. If you want to cut out the workload, limiting practice time to the minimum required to maintain readiness to play is a way to do that. This is the MLB model for pitchers, as opposed to the Japanese model. And the difference between those approaches might be two different guesses about what works, or it might be driven by how many games a pitcher plays in each league.
However, It’s extremely difficult under the current rules for most NHL teams to sit their starter at practice and still have two goalies for their players to shoot at in drills. An NHL team is not allowed to bring over their handy AHL goalie from the next icesheet to be the practice goalie without calling him up to the NHL. If you don’t have the cap space for that, or the roster space, you aren’t allowed to do it. A workaround could be to keep putting someone from the AHL team on a PTO for practice days, or the NHL and the NHLPA could simply redo the CBA to allow those one-day practice recalls, but that would help the Leafs, the Sharks, the Jets and the Kings, and no one else. And at that, it only works if you’re both at home at the same time.
The Leafs have resorted to two other options. Last year they signed Joseph Woll in the spring when they had cap space and roster space and used him as the sacrificial lamb to let Auston Matthews whistle pucks past in practice. They also used goalies from Toronto university teams from time to time throughout the season.
The best concept for creating an ease to workload that doesn’t eat into a team’s chances of winning might be to allow the NHL team to hire on a practice goalie as a paid position outside of the salary cap. But who wants that job?
It could be that the opportunity to get paid to work with an NHL team’s coaches might be something a lot of young goalies would want, and that’s got to be the motivation for the university goalies. But any rule change has to work for every team, and the NCAA would prevent almost any sensible arrangement, so just relying on the local college isn’t a universal solution.
A different philosophy of the waiver and roster rules might allow teams to carry some extra goalies at the AHL level and to call one up for a longer period of practice days without causing the domino effect through the AHL and ECHL that leaves someone shorthanded. Now you don’t need your AHL team to be as local. But you do need the money to make this happen.
The NHLPA has two interests when negotiating the CBA that apply: They are there to ensure player health and safety is promoted and maintained, and they also just want more players in the NHL on NHL contracts. So allowing for an AHL-rostered player to come and go without waivers seems like something they would oppose, but creating a better work environment for the top 25% of the starters in the league is something they should be in favour of.
Mike Babcock told the press he wanted to get a third goalie in this season, and they had a mystery man at one practice who went unnamed. I’ll leave it as an exercise for the reader to discover who York’s backup is this year. He’s my suspect number one for who that was.
The NHL could stand to give this issue a long think. Because right now the deck is stacked against teams who actually do want to limit the starter’s starts, and the more parity in the league, the tougher it will be to convince any coach or GM to play the backup goalie more. So limiting practice work might have to be the first move in maximizing the health and effectiveness of the small pool of better than average NHL starters.
The best plan still is to have more than one potential backup on an NHL contract at all times. And this is something the Leafs have struggled to do for years, not because they haven’t tried, but because it’s hard to do even when you don’t skip drafting a goalie for years at a time like they did.
Like yesterday, the conclusion of this piece is exactly the same thing Micah McCurdy tweets from time to time: Recognize that your goalie performance might change, and if it drops, have a plan to deal with that. And not just for your win-loss record, but for the health of the starter.