Last year, after the Boston Bruins lost to the Florida Panthers, 7–5, in Game Six of the first round of the Stanley Cup playoffs, a reporter asked the Bruins goalie Linus Ullmark how his body felt. “I feel good,” he replied. “Thanks.” This was a lie, and everyone knew it. It wasn’t just that the Bruins, who were coming off a historic regular season—the fastest team ever to record fifty wins—had let a 3–1 series lead over the lowly eighth-seeded Panthers slip away, and were now facing an elimination game. In Game Five, Ullmark had turned over the puck behind the net, leading to the Panthers’ game winner. He spent much of Game Six flailing, losing track of loose rebounds, pulled out of position. It was clear that he was hurt or drained, or both. That game was the first time all year that Ullmark, who’d been the league’s best goalie during the regular season, had given up six goals. It was also the first time that year that he had played in six straight games.
During the regular season, Boston, in a departure from tradition, had alternated starting goaltenders, playing Ullmark one night and Jeremy Swayman, who’d also been consistently terrific, the next. By the time the postseason arrived, Ullmark hadn’t played two straight games in months. But the overwhelming belief was that you can’t platoon goalies during the playoffs. At the start of the series against Florida, a reporter for The Athletic asked the former Bruins goalie Andrew Raycoft, now an analyst for NESN, whether the team should continue alternating netminders each game during the playoffs. “No,” he replied. “No, no, no, no, no. No rotation. This is Linus’s show. You just roll with the guy. Rotations don’t work.” He’d done his research, he added, and was convinced that it was the consensus view. But, ultimately, starting only Ullmark didn’t work, either. “I think Linus is fine,” the Bruins coach, Jim Montgomery, told the press after Game Six. “I looked him in the eyes a few times on the way back to the bench, and I liked how his eyes were looking. He looked intense, and he looked keen.” It later emerged that he was banged up; Montgomery replaced him with Swayman for Game Seven. Swayman, who, aside from a few minutes on the ice in Game Four, was coming off seventeen days’ rest, now had to save the season. The Bruins had their chances, and the game went to overtime—Swayman made two spectacular stops in the final minutes. But he was helpless against the game winner. The Bruins lost, in one of the biggest series upsets in N.H.L. history.
This season, except for a stretch in which Ullmark was injured, in January, the Bruins went with the tandem again, alternating every game. The last time a Bruins goalie started two straight games was in mid-February. Swayman had the more impressive season, but only marginally; the duo’s save percentages were almost identical. Swayman started the team’s first playoff game, against the Toronto Maple Leafs, and put in a sterling performance. But, this time, Montgomery stuck with his platoon: Ullmark started Game Two, and played very well in a close loss, and Swayman was back in the net for Game Three, which Boston won. Game Four is on Saturday night. Having a high-quality tandem is a strength for any contender, given the possibility of injury and fatigue. A number of teams use more than one goalie, but there’s usually a clear hierarchy. No team has treated two goalies as No. 1 starters and won the Cup since the Bruins did it back in 1972.
It’s a truism in hockey that, during the playoffs, you have to ride the goalie with the hot hand, that the guy who’s doing well today is the guy who’s more likely to do well tomorrow. If the backup has the hot hand, he’s vaulted into that spot—until he falters. It’s why Mike Vernon, the Detroit Red Wings’ backup, was asked to lead Detroit to the Cup in 1997, instead of the team’s usual starter, Chris Osgood; it’s why Washington Capitals fans revere the name Braden Holtby, the goalie who got the Caps their only title. Last year, Adin Hill, who began the playoffs as the backup for the Vegas Golden Knights, got his first start halfway through the second round, and proceeded to win eleven of his fourteen starts as the Golden Knights won it all. Playoff lore is full of similar stories. Even with established starters, the great hope is that they’ll be hotter than the goalie across the ice.
If you ask hockey lifers, they will explain that goaltending is a position that relies on runs of confidence. “He trusts himself,” Raycroft said of Ullmark last year, after he played spectacularly in the first game against Florida. “He trusts his size in those plays,” Raycroft added, referring to some of Ullmark’s difficult saves. “That comes from confidence. He’s not trying too hard. You start trying too hard on plays like that, you open up. You cause holes through you. Instead, he’s not getting out too far. He’s got perfect depth, whether that’s middle of the crease or all the way back on the line, depending on what’s going on in front of him. He trusts that depth. He trusts his size.” Trust, confidence—these are old-fashioned sports virtues, ones that can sound out of place in a world where more and more coaching and commentary concerns precise statistics and meticulous measurements, a mathematically informed understanding of how games are won and lost. The idea that there’s something going on that can’t be reduced to numbers seems almost quaint. But what if it’s right?
In the nineteen-eighties, a group of researchers, including the influential psychologist Amos Tversky, set out to study the “hot hand” phenomenon—the belief, widely held, that people have runs of success in which one positive result makes the next positive result more probable. They focussed on shooting in basketball. It was conventional wisdom that a player who sinks a few baskets in a row is hot, and more likely to sink the next one. But the researchers found no convincing statistical evidence that this was true.
Their finding was later challenged; the current consensus among researchers is that, when it comes to shooting a basketball, there is a mild positive effect created by hitting a shot. In baseball, too, there’s some evidence of streaky runs—hitters who string together good at-bats and gain a degree of momentum, and pitchers whose fastballs start zipping faster. Still, there is a broader finding in keeping with the original study by Tversky et al.: we tend to overstate the power of these phenomena, and there is a psychological tendency to see patterns in randomness. This is true not only in sports but also in gambling, among other activities.
So what about goaltending? In 2021, four researchers at the University of Alberta looked at the effect of shot performance on the probability of saving the next shot—the hot-hand idea. They found that the effect was actually negative. But hot-hand studies have been wrong before, and I was struck that that was the only one of its kind I could find. Hockey was slower to embrace analytics than some other sports, and it still has its share of romantics. Every team now uses analytics, though, and there’s little doubt that the Bruins have a wealth of data, proprietary and otherwise, which they use to advise the coaching staff. It’s harder to know just how meaningful that advice is. Analytics has shown that the intense focus on goaltending is justified: good goalies are important, and some goalies are simply great. What’s more, the effect that a goalie has on the game is amplified during the playoffs. But it remains difficult to predict which goalies are going to be good from season to season or during the postseason. Twelve different goalies have been finalists for the Vezina Trophy, awarded to the regular season’s top goaltender, in the past five seasons. Only two of them have been on the list more than once.
I asked a couple of thoughtful and hockey-mad friends for their opinions. Is the hot hand a myth? They scoffed at the question. “An athlete gets into a flow,” one of them told me, in an e-mail. “You can deny this but you’d be wrong!” Of the anti-hot-hand apostles, he added, “The only people who believe this stuff, these stats guys, must be people who don’t play anything. Do the stats guys honestly try to argue that a basketball shooter heaves up the same ball, night after night, and some nights it goes in, and some nights it doesn’t?” Our other friend allowed that this is, in fact, what some stats guys have honestly argued, about basketball—and, he went on, they weren’t entirely wrong. But hockey, this other friend maintained, was different. It was a game of confidence, just as Raycroft had said. It seemed silly to pretend otherwise. There’s a hockeyism attributed to the former commentator and player Jim McKenny: “Half the game is mental, half the game is being mental.” I was starting to see why it would drive a person crazy.
Hockey is dynamic and mostly continuous. There are usually twelve people on the ice at a time, and ten of them are constantly moving. Skaters move faster than people in sneakers, and the puck moves fast, too. When skaters tire, they leave the ice and are replaced without the clock stopping; the game never slows.
And then there’s the goaltender, who is either distant from the action or right at the heart of it—very lonely or in hell. His job is to block a small disk of frozen rubber which may be travelling at a hundred miles an hour, to track its irregular bounces through collapsing gaps between flailing sticks and limbs, to withstand the force of large bodies crashing into his, to cover for a defenseman’s lapses, and to make no mistakes at all. “You’re afraid of getting hurt, but you’re even more afraid of being humiliated,” a goaltender told Sports Illustrated years ago, for a story about an evergreen topic: goalies’ bizarre preparatory rituals (vomiting before games, talking to goalposts, wearing thirteen pairs of socks). The legendary goaltender Ken Dryden once described the position as largely “aphysical.” It requires, Dryden wrote, a “certain character of mind.” He added, “Because the demands on a goalie are mostly mental, it means that for a goalie the biggest enemy is himself.”
The Bruins aren’t the only team to use tandem goaltenders these days, though they’ve been at the vanguard in doing so. The game has become more physical and faster, the shots harder, the contracts bigger, the schedule more compressed. Sixteen years ago, six goalies had at least seventy starts; in the past five years, there hasn’t been a single one. Goalies tend to be a little more interchangeable than they used to be—they play a more homogeneous style, with an emphasis on using oversized pads to block the puck instead of catching it or covering it up. And some teams have discovered the obvious: there are mental as well as physical benefits to regular rest.
Many activities benefit from a degree of learned automatism, but goaltending is at the extreme, Joe Bertagna told me. Bertagna is a former goalie who served as the goaltending coach for the Bruins and for Team U.S.A. before he became the commissioner of the Eastern College Athletic Conference and the Hockey East Association, two of the premier conferences for college hockey. A goalie might face thirty or more shots on goal in a game—and that’s just the start of his troubles. “Handling pucks, tying up pucks, quick decisions of whether to stay up or down, to be aggressive”—these are instinctive decisions, Bertagna explained. If a goalie overthinks his actions—if he thinks consciously at all—he’s toast. A goaltender has to have “an unrealistic level of confidence,” Bertagna said. “If he had the same level of confidence in society as during a game, he would be a very offensive individual.” But a goalie can’t imagine himself as too heroic, either, for that can also be fatal: his job is not to win the game; his job is to not lose it.
That kind of mental disposition, that focus, can come and go. Players describe stretches when the game seems to slow down for them, when their reflexes quicken and their sense of timing becomes perfect. Hence, perhaps, the hot hand. It has been, at least so far, as good an explanation as anything else. “Goalies are voodoo,” my friend who scorned the stats guys said. He was echoing, it turns out, a lot of frustrated stats guys. “It is like watching an origami master in action, constructing not a paper crane, but a perfect wall,” Dryden once wrote of a goalie at work in the net. He was offering a critique of the trends in the game, which made for a less entertaining, less flowing style, but his description evokes another truth. Goaltending is an art, and not one that can be easily measured. Sometimes you have to rely on trust. ♦
An earlier version of this article inaccurately attributed a prize to Amos Tversky.
