As you can see in the graph above, there are two key points in team we have to realize before asking these questions; those points involve NHL rule changes that had an obvious effect. In 1950-51, NHL teams were first required to have “emergency” goaltenders available to play* ; by 1965-66, teams had to dress two goaltenders for each game. The graph above suggests that the 1950-51 change didn’t have a major impact…and that might be true but for this graph:
It seems that the 1950 rules change was reflective of a general interest among teams to reduce fatigue and wear on their goaltenders, as well as provide a ready contingency plan should their main goaltender go to seed. No more Sugar Jim Henrys, or Glenn Halls, goaltenders who had such extraordinary resiliency were no longer expected.
At a theoretical level, this shouldn’t necessarily result in better goaltending. Backup goaltenders, as a population, are expected to be league-average, but typically their error bar (low NHL sample) or their age regression will factor in and should push them down about 0.005 … in addition to the usually 0.005 lesser they are likely to be than their starters. On the other hand, we do know that goaltenders today (and likely back then) play worse in back-to-back games, so the difference there (minus 0.01) should be made negligible if the replacement was done on basis of fatigue. But they weren’t, so (at least theoretically) there should be a slight dip. There doesn’t appear to be, though, not until the 1980s. And notice in the top graph its relationship to replacing the goaltenders…the peak of replacement was reactive to the highest-scoring era. As goaltending settled into the low-scoring of the last two decades, so has the goaltender replacement rate similarly settled.
It deserves some more finite analysis, but it doesn’t particularly look like much changed due to switching goaltenders, league-wide.
* As reader @NHLDrafter points out, the emergency goaltender was rarely dressed, and provided by the home team. Incredibly, away teams did, on rare occasions, use these opposing goaltenders, and the results were very interesting. There were four players I could find that definitely played emergency goaltender for the opposing team:
- Ross Wilson – Also known as “Lefty,” the Detroit Red Wings trainer is maybe better known by his handiwork in designing one of Terry Sawchuk’s first goalie masks. In 65 minutes of emergency work for the opposing Maple Leafs and Bruins, he gave up just one goal on 32 shots. Hockey historian Joe Pelletier has a piece on him here.
- Claude Pronovost – The Montreal Canadiens goaltender played 3 career games, one against his team. The Boston Bruins’ Terry Sawchuk was injured and the backup goaltender forgot his skates, so Pronovost stepped in and did nothing more than record a 31-save shutout against his teammates.
- Len Broderick – The Maple Leafs’ emergency goaltender, at the time playing in Junior A in Toronto, relieved some guy named “Jacques Plante.” Plante had suffered an asthma attack before the game; Broderick did his best imitation and stopped 20 of 22 shots to backstop the opposing Montreal Canadiens to a 6-2 victory. “You’ll never work in this town again,” sneered Conn Smythe (I think), and Broderick never played another game in the NHL.
- Don Aiken - He came in for Jacques Plante when Doug Harvey and Vic Stasiuk knocked Plante out cold in the second period of a Habs-Bruins tilt. The two collided into Plante, whose head fell back, hit the crossbar, then fell forward onto the ice. Aiken came in and 2:30 later he’d given up his first NHL goal. Three minutes after that, his second – and 30 seconds later, his third. He waited 1:30 to give up his fourth, and all told he gave up 5 goals on 9 shots in the 2nd before settling down (or score effects) allowed him to stop all 8 shots in the third period. The Bruins won 7-3.