Patrick Roy and the Incredible Improving Goalies

Why did save percentages go up so dramatically; 40 points between 1984–85 and 2014–15? There are many theories, some of which rise to the level of reasons: the butterfly is an obvious one. To watch highlights of ’70s and ’80s hockey, with stand-up goalies waving their gloves at pucks any modern midget tender would have stopped, is one time that the eye test tells you something very important. Pads, and sweaters, got bigger. Goalies got bigger: Pekka Rinne in 2017–18 was the first Vezina winner 6’4″ or taller since Ken Dryden, after whom only Marc-André Fleury and Igor Shesterkin have been shorter. Six of the first twelve Vezina winners under the modern voting rules were shorter than 6’0″ (Billy Smith, Pelle Lindbergh, John Vanbiesbrouck, Grant Fuhr, Ed Belfour twice); no goalie so short has won a Vezina since Tim Thomas in 2010–11. The neutral zone trap and other improved defensive strategies probably did no harm.

Any goalie could have adapted to this. Patrick Roy couldn’t get any taller, but as he pioneered the butterfly he also would have pioneered its improvements. Not only could he have put on larger equipment, boy did he: if you remember late-career Patrick Roy he looked like he was wearing a tarp and holding a salad bowl in his left hand. He couldn’t only glove Theoren Fleury’s best shot, he could glove Theoren Fleury. So now we come too far the other way; not only are we not surprised Patrick Roy’s save percentage improved with time, we’d be surprised if it didn’t.

In fact, Roy is in a very exclusive class. He was contemporaneous with great Vezina winners, Hockey Hall-of-Famers, some great goalies, and yet Roy stands out in the upper echelon. Below Dominik Hasek, who was better at playing goal than any other player was at anything, yet you have to say: Roy was not just the first of the goalie era, Hasek aside he may have been the best.

Here’s a look at the save percentages of some great goalies of Roy’s epoch, who had careers long enough to be worth charting (i.e. Pelle Lindbergh’s not there) and who won a Vezina Trophy in the modern, “you vote on it” era.

I’m sorry that this is a bit spaghettine; if you click on a name you’ll see that goalie’s line highlighted. At least spotting Patrick is easy; his line is the one that starts at 1.000. (As a 19-year-old, Roy played one regular season period and stopped both shots he faced, which would come in handy for chart-makers decades later.) The league average save percentage is the thick blue line, which you can see climbing up through Roy’s retirement at the end of 2002–03.

The Vezina winners of the clearly pre-Roy era, Billy Smith and Pete Peeters, look like what you’d expect. Peeters was a pretty good goalie but inconsistent, which makes it funny that he became such a well-known goalie coach; Smith was a plain good goalie who, in the timeline on this chart, got old. Both were old-style goalies, though said old styles were very different, with Peeters exemplifying the no-nonsense tendy who got the job done and kept people calm, while Smith found fighting the opposition somewhat more congenial than fighting his teammates but would have slashed the Zamboni if it was the only thing to hand. Peeters, though inconsistent, was never bad; his inconsistency was of the “will he be a Vezina candidate or will we get by with him?” variety. Smith, with a great team in front of him, was almost the axiomatic winner, but once his teams stopped winning it fell apart for Smith like a Jenga tower. Smith played relatively few games: he had one season of 58 games, two of 46, and played fewer than 40 games in more seasons (10) than he played 40 games or more (seven), an unusually light workload for a Hall of Fame goalie. Peeters was not really a workhorse but had a 63-game season (in which he led the league in many things and won the Vezina), a 51, a 49, and six seasons of fewer than 40 games out of 13 seasons overall. You liked having Pete Peeters around; you liked having the Stanley Cup parades that Billy Smith helped precipitate.

In every field except winning, Billy Smith’s heir was Ron Hextall, to a degree that was so obvious it became tiresome to repeat. Angry little guys who could use their sticks. At first it looked like Hextall was a winner too. All goalies, given enough opportunity, will have a fluke season or two either way. Hextall was lucky in that his fluke season was his first, and it was a monster. He won the Vezina and you can see why, especially once I tell you he also led the league in goalie games played. Great Vezina choice. Hextall was the perfect goalie for his Philadelphia Flyers team, a big battling bastard who took shots both at ankles and on the other net given any excuse, and the fit was so perfect that the surprise was not really that the Flyers made the Stanley Cup final, but that the Oilers were able to beat them.

That was it for glory; for the rest of his NHL career, Ron Hextall never made another final and as an individual slipped towards average. Not quite to average: he was still above the league mean a bit more often than not, but it had all gone a bit less right than anyone predicted. He held out before the 1989–90 season, when he returned he picked up niggling injuries (that dreadful .868 was in only eight games). If you acknowledge his rookie year as a fluke, because it was, it’s not quite fair to say Hextall “was never the same” after his holdout; he dipped a bit from seasons two and three but from time to time could reach comparable heights. Hextall’s stubbornness in contract negotiations was perhaps just a sign: that he would not adapt well to a league that was becoming goalie-dominated. His hamstring and his groin did the rest. Hextall had one great season and a few pretty good ones, i.e. he was a mean Pete Peeters, but as the league improved, Hextall did not.

The gentleman of our young-’80s-phenom goalie comparison was definitely Tom Barrasso. Like Hextall, Barrasso won the Vezina as a rookie; he wasn’t nearly as spectacular about it was also only 18 years old, and he was darned near as good the season after, fourth among starts in save percentage, first in goals-against average, first in shutouts. Barrasso, unlike Hextall, had a long career and got a second lease on stardom helping the Penguins to a couple Stanley Cups. He is in the Hockey Hall of Fame, elected in 2023, and God knows why. For the simple fact of the matter is that, as the league save percentage got better, Barrasso’s did not. His numbers look respectably consistent for a goalie, but everyone else was improving, and by the time of his Penguins triumphs Barrasso was a merely league-average goaltender most of the time. He came close to another Vezina in his two post-apex peak seasons, which show up on this chart like Everest and K2; they all came after, and before, such mediocrity that I think voters were glad to see the old guy stopping pucks again. Like many goalies of his era, Barrasso did not play heavy minutes, partially because he did not usually look better than some pretty average backups. In 1997–98, Barrasso’s last great year, he came third in the Vezina with a .922 save percentage, which was honestly really good, but his backup, Petr Skudra, was a .924. You remember how good Pete Skudra was, right? Those Penguins didn’t have Mario anymore but they did have a defense of Kevin Hatcher, Freddie Olausson, Brad Werenka, Jiri Slegr, Darius Kasparaitis, and Chris Tamer; you have to get to the number six before you find someone who wasn’t a really good own-zone player.

To avoid accusations of bias, we will look at Oilers legends Grant Fuhr and Andy Moog together. Fuhr has his number retired and is in the Hall of Fame; Moog has not and will not be. Moog’s greatest services to the Oilers were getting beat by Edmonton in the playoffs, in the 1990 Stanley Cup final as a Boston Bruin and the 1997 first round as a Dallas Star. I hold a soft spot for Andy because of that, but they are perhaps not demonstrations of excellence.

Moog’s reputation is as The Other Guy, while Fuhr was The Winner. Moog might play more in the regular season, it was Fuhr who got the starts in the playoffs, and he could handle monstrous minutes. Fuhr four seasons of over 60 games despite beginning his career in a platoon, he had three of over seventy, and he holds the NHL record, probably forever, for playing 79 games with the St. Louis Blues in 1995–96 as a 33-year-old. Fuhr wound up with 868 NHL games; Moog “only” got 713, and one season of 60 games or more. You therefore must give Fuhr massive points for being better than any goalie of his generation at being available, ready, and willing to play, or you would if he was any good, which he was not. Fuhr was below the NHL average save percentage in ten of his nineteen NHL seasons, and above in seven (two of them were ties).

“But he played for the Oilers, they played firewagon hockey!” Fuhr and Moog split the Oilers’ platoon for five seasons, 1982–83 to 1985–86; Moog had the better save percentage three times and they tied twice. The objection to that is that Glen Sather preferred Fuhr in the money games and sent Moog out against the Colorado Rockies, which is certainly true, but Moog just played too often for you to make too much of that. Besides, Fuhr was above the league average save percentage five times as an Oiler and twice everywhere else; those Oiler numbers were about as good as Fuhr got.

In that 79-game season, Fuhr had a save percentage of .903 against a league average of .898; his backups were garbage, thus 79 games, but he wasn’t exactly stealing points. The next year Fuhr played 73 games with a save percentage a bit below the league average. Like Tom Barrasso, Fuhr comes off as consistent because his numbers look consistent (in 1994-95, the big mid-career dip, he played only 17 games; in 1989-90, the other ugly-looking year, he played only 21). However, the league was getting better in goal faster than Fuhr was. By the end, as he dipped back below .900 in his mid-thirties, he was killing his teams, and of course his last season as a backup in Calgary was a travesty. Moog, by contrast, at least ticked up with the league save percentage a little bit; he was not very consistent, and didn’t play nearly as much, but was better-than-average much more often than not. He didn’t play for killer teams, apart from the dynasty Oilers; the Stars weren’t properly good until Moog’s last season with then, the Bruins were sometimes pretty good but usually sort of fine. Grant Fuhr is in the Hall of Fame and is famous for winning the big one; Moog is not and is famous for losing it. But let’s be honest, the ’80s Oilers were not really a difficult team to win with, were they? They would have been successful with even bit-better-than-average goaltending. As demonstrated by the fact that that’s what they had, and they were.

We are starting to see a rule; that top goalies of the Roy era did not improve much in relation to the league as time went on. However, one Vezina winner who is less remembered than some of those we’ve just named is John Vanbiesbrouck, and he breaks the pattern. Lightly, with a ballpeen hammer only, but definitely.

This type of layout flatters Beezer, because for a top goalie he didn’t play that much: in 20 NHL seasons he had only three of 60 games or more (Patrick Roy had 11). A typical Vanbiesbrouck season, especially once he became Beezer the Geezer, was, say, 55 or 60 games where you never felt like you could absolutely count on him. He was good enough on his day that you got over it, at least until you decided you couldn’t live with it anymore and traded him for, geez, Doug Lidster or the draft pick that became Jordin Tootoo or something else you’d massively regret. At a time when Ed Belfour or Curtis Joseph would be getting more money than Croesus to go lose with the Leafs, Vanbiesbrouck somehow managed to be underrated, always the odd man out. He’d get hurt or he’d hit a long cold streak and what he couldn’t do became so obvious that it overshadowed the many things he could.

More than most goalies, he never seemed to have the lasting confidence of his coach, and a bad run would have him out the door. Though he was repeatedly and sustainably excellence, nobody except the fans really seemed to believe you could “leave it to Beezer” for more than a season. His career playoff save percentage was a fine .911, but one bad series in 1991–92 meant that Mike Richter carried the mail to the Stanley Cup in 1992–93, which is the most Beezer thing to have happened in the Beezer’s career. To platoon with Richter on the championship Rangers was understandable to give up 25 or 30 games to the likes of Mark Fitzpatrick was just a weakness. Still, he was a darn good goalie a lot of the time, and you can see it.¹

We’ve surveyed some could-have-been greats and some supposedly-were greats. Generally, these goalies did not “improve,” though Vanbiesbrouck might have been an exception: if anything, they got a bit less great relative to the league as save percentages leapt up in the 1990s. But, and here is a big but, we have not yet discussed Patrick Roy and Ed Belfour.

The great goalies of the 1990s were Patrick Roy, Ed Belfour, and Dominik Hasek. Hasek was the greatest; he also didn’t become a starter until 1993. That was the year league save percentages made their first really big jump, ten points over the season before, and since Hasek had a .930 save percentage that year some of that really was “Dominik Hasek raised the level.” However, even handling 1,552 shots in a lights-out one-of-the-best-seasons-ever fashion doesn’t make that big a statistical impact on the whole league. Hasek came along at the right time to look even better than he was; he was still obviously extremely excellent.

Belfour and Roy both had some 1980s priors to be compared to. Roy became a starter in 1985–86 at age 20, though he was the senior man in a near-platoon with Brian Hayward for his first five seasons. Belfour played 23 games (badly) as a 23-year-old in 1988–89, then got in 74 games as a rookie in 1990-91 and won the Vezina Trophy during what was still the era of ’80s hockey, even though the calendar had changed. Roy was the chosen one since he was a teenager; Belfour exploded onto the scene like an atomic bomb, old for a rookie as excellent as he turned out to be. They were both, to put it gently, passionate men. They both won Stanley Cups in marvelous fashion and were undisputed “money” players. On the big chart at the top, they both stand out dramatically above the league average line, give-or-take interruptions by John Vanbiesbrouck.

As we’ve seen all goalies with careers long enough have some fluke seasons, for good and for ill. Belfour’s fluke seasons took him below the league average save percentage at their worst; way into Vezina territory at their best. He did not have a bad year between when he was a 23-year-old backup and when he was 36. Patrick Roy simply never had a bad year. His rookie season was ordinary, but he turned into Patrick Roy in the playoffs and won the Conn Smythe. His most disappointing season was probably 1995–96, the year of his famous feud with Mario Tremblay: Patrick beat the league average save percentage by a mere ten points and was only eleventh in the league. He was almost as “bad” in 1999–2000; ten points above average again and tenth in the league (though five ranked ahead of him played fewer than 40 games). Patrick played 60 games or more in a season 11 times; Belfour did it nine times, including three seventy-game seasons. In his 74-game season, 1990–91, Belfour led the league in save percentage and goals-against average to win one of the easiest Vezinas ever decided.

Patrick Roy’s 1989–90 season must be one of the best ever by a goalie not named Dominik Hasek relative to his peers. He finished with a save percentage 0.031 above the league average. His .912 was seven points better than second-place Mike Liut, who played only 37 games, and nine points better than the next-best starter Darren Puppa. Roy, Liut, Puppa, and Clint Malarchuk were the only goalies to even get above .899. Hockey Reference tracks an advanced stat called “Goals Saved Above Average”: Roy’s 47.4 was way ahead of Puppa’s second-place 35.7, which was in turn way above Jon Casey’s 25.8 in third. The post-’80s record is Curtis Joseph’s 57.4 in 1992–93; Roy is behind Joseph, Vanbiesbrouck, Linus Ullmark, Roberto Luongo, and like four Hasek seasons in the modern game. But most of those goalies saved a lot of goals above the expected average because their teams stank and they did a lot of work. The Beezer went 21-25-11 in his great season. Cujo was a game above .500. Luongo was eight games below. Ullmark played for a fantastic 2022–23 Boston team, but we have to say that season was one of the great flukes of all time by now, don’t we? Patrick Roy went 31-15-5 with the Canadiens; he didn’t keep out a million pucks because his team was so bad he faced ten million. Nor was he a fluke; he had two other seasons above 44, which for goalies as great as Luongo and Vanbiesbrouck was a once-in-a-lifetime achievement. Of course Hasek was even more unbelievable, but… yeah, he was.

And you can see, Patrick Roy improved. For Roy’s first five NHL seasons, league save percentages were improving but Roy’s were improving much faster.

We have not addressed Dominik Hasek, apart from as an example, because there’s really no point: he only became a starter in 1993–94, when the save percentage climb was already well underway. But just to give the Dominator his due, here they are during the period where their careers overlapped.

In his prime, the difference between Dominik Hasek and Patrick Roy was the difference between Patrick Roy and Bob Essensa. Dominik Hasek was very good.

1. Beezer played one game as an 18-year-old pre-rookie in 1981–82 where he stopped 29 of 30 shots; to keep the chart clear that game is omitted.