Let's celebrate Ben Bishop's career and take a VERY deep dive on how egregiously he was robbed of the Vezina in 2018-19

This post grew from what I originally intended to be a comment upon hearing of Big Ben's retirement. It ended up being quite long for a comment section, so I posted a raw, undeveloped draft on a small hockey forum. But it still didn't feel "finished", so for the last couple of days, I've set aside an hour or so to keep fleshing it out more: I liked a lot of the data points I'd found, and I wanted to do justice to Bishop's incredible tenure between the pipes in Big D. So this longform post is what I ended up with. I know not many folks will read the entirety—I ain't gonna pretend to be Jack Kerouac—or even see this post!

That's fine, because in hindsight, another big reason I wrote it was to formulate an argument for why Bishop deserved the Vezina over Vasilevskiy in 2018-19. To verbalize thoughts, whether orally or in writing, is the best way to think; as unintuitive as that sounds, few would dispute it. I'm happy with the case I end up making for Bishop here; whether or not it's convincing, it's certainly thorough! If you manage to read the whole thing, I salute and thank you. If you don't care to take the time, I encourage you at least to have a look at the numbers in the last section & the scatter plot that visualizes Bishop's inconceivable dominance and consistency. Honestly, I could've just posted that graph, and likely made just as good a point ;)

First of all, what a warrior he is for even trying to come back. Per Wyshynski's article on Big Ben's retirement, he missed very few practices with the team. Presumably, the doctors had told him the pain would likely be permanent, but when he realized his knee was as healthy as it was going to get, he said "what the heck, let's give her a try anyway." He asked for an AHL stint to see how manageable the pain would be during games; he allowed eight goals on 34 shots in his one start, so we can assume the pain was significant.

Would anyone have blamed him if he'd asked out after allowing a fifth, sixth, or seventh goal in the second period? As it happens, Coach Graham pulled Bishop after the seventh goal, with less than two minutes left in the second—but Bishop played the entirety of the third period. He knew he was getting lit up, and he could've hidden in the locker room for the third, icing his knee. But we can only assume he pushed through it, physically and mentally, because he needed to know how his knee would handle three periods.

In hindsight, the comeback was a long shot from the start. Most of us north of 30 have probably felt the pain of putting too much weight on one knee; if that even approximates the pain Bishop attempted to play through, he's one gritty fellow. Given that his job largely consisted of rapidly standing up from the butterfly, onto a skate blade a few millimeters wide, he probably made the right decision and certainly saved himself a lot of pain. That said, I'm glad he played the game down in Cedar Park. I can't help but be proud of him; had he retired without at least seeing how well the knee held up during game action—without putting himself through 60 minutes of physical agony—he would've likely asked himself "what if?" for the rest of his life.

But more to the point of this post, Wyshynski briefly summarizes Bishop's success. We all know he was elite when healthy, but I admit, I hadn't realized just how dominant Bishop was until Wyshynski put it thusly:

Bishop was one of the most successful NHL goaltenders of the past decade. Since 2013-14, he's tied for eighth in regular-season wins (204) with Carey Price of the Montreal Canadiens. That's despite not having played the past two seasons due to his injuries.

Bishop was a finalist for the Vezina Trophy in 2014, 2016 and 2019. He had an overall record of 222-128-36 with a 2.32 goals-against average and .921 save percentage in 413 games, playing 227 of them with the Tampa Bay Lightning and 143 of them with the Stars.

That career .921 save percentage of his? It's tied for the second-highest in NHL history, trailing only the two goalies who share the tie for highest at .922: Hasek and Dryden. (Tuukka Rask oscillates between .922 and .921, as a minor disclaimer.) The other aspect of Bishop's career that stood out to me was just how short it was. 11 seasons seems on the shorter end for a longtime starter, but certainly not anomalously short—but then you look at his 396 career starts and realize he only averaged 36 starts a year across those 11 seasons. Big Ben could've potentially made himself a Hall of Fame candidate, had his body cooperated.

That's not to say three Vezina nominations are anything to sneeze at, though. In fact, his final Vezina nomination, in 2018-19, is what I'm most eager to talk about. As Stars fans, we're used to our boys getting sold short in terms of postseason hardware.

Kaprizov over Robo? Well, maybe the voters didn't look at their games played totals or ice time. Heck, even if they knew Robo's production absolutely blew Kaprizov's out of the water on a per-minute basis, it's not the worst robbery. Kap had a healthy advantage in games played.

Ekblad over Klingberg? Eh, maybe a bit tougher to accept. Klingberg was better just about any way you slice it. But whatever, Ekblad had all the preseason hype and nobody outside Texas knew Klingberg's name before that season.

Brodeur over Turco? Yeah, I sure wouldn't mind a puff of whatever the voters were having then. This category is roughly where I'd rank Vasilevskiy winning the Vezina over Bishop: it was pure, unadulterated larceny, and not the good kind that we hear Razor mention every now and then.

A very brief skim-over of the not-so-fancy-stats I'll use:

  • Adjusted Goals Allowed Percentage (GA%-): Don't be intimidated; this is effectively just save percentage, re-packaged to tell us how much lower or higher than league-average a goalie's save percentage was in a season. The calculation is surprisingly simple and its only inputs are Sv% and league-average Sv% (lgSv%). It's normalized to 100 so that a 120 GA%- means a goalie allowed 20% more goals per shot against than the league average, and a 92 GA%- means a goalie allowed goals at a rate 92% of the league average, or 8% fewer goals per shot. See? Not so fancy.
  • Goals Saved Above Average (GSAA): Here, the inputs are the same as above—Sv% and lgSv%—plus a third: shots faced. GSAA tells us how many more/fewer goals a goalie prevented, given his Sv% and shots faced, than he would've if his Sv% had instead been league-average. Again, not so fancy; we're not getting into theory here, just plugging in lgSv% and solving for X. Do note that this is a cumulative stat (like Runs Batted In), not a rate stat (like Batting Average)it will favor goalies who start more games (though we can turn it into a rate stat if we divide it by games started, which would give us GSAA per start).
  • Quality Starts & Quality Start Percentage (QS & QS%): A quality start is simply a start in which a goalie posts a save percentage above the season lgSv% (or, if facing 20 or fewer shots, above .885, to account for the sensitivity of such a small sample). QS% is, you guessed it, a goalie's quality starts divided by his total starts; i.e., the percent of his total starts that were quality starts.
  • Finally, Really Bad Starts (RBS): The number of starts in which a goalie's Sv% was below .850. As with QS, we can theoretically derive RBS% using (RBS / Starts), though Hockey-Reference doesn't do so.

QS and RBS aren't particularly useful as predictive stats: given the volatile, variance-prone nature of Sv% outside of large multi-season samples, QS/RBS totals probably won't tell us much about how well or poorly a goalie will play in the future. But you can make a case for their value as descriptive stats that tell us how consistently a goalie has played, over a given data set of previous games or seasons. I.e., if a goalie finishes the season with a league-leading Sv% and QS%, and an extremely low RBS number relative to his total starts, it would mean he was performing at an elite level virtually every single night, and that he was having a freakishly, historically great season. You can probably see where I'm going with this, so keep that last thought in mind!

Few Have Defended Big D Quite Like Big B

Before we zoom in on the sheer dominance of Big Ben's 2018-19 season, his production before arriving in Dallas deserves a mention. You've probably heard folks dismiss Bishop's stats to defend Vasilevskiy's Vezina win, claiming that Bishop played in a much more "favorable system." Apparently, it's easy to overlook the fact that Bishop received two Vezina nominations playing for the Bolts, including three seasons in which the two were teammates.

Bishop recorded 41.6 total GSAA across his five seasons in Tampa, 40.1 of which came in his three seasons as their de-facto #1 starter. He was elite in two of these seasons, 2013-14 and 2015-16, with a Sv% of at least .924 and a QS% of at least .603 in both seasons. These years accounted for his first two Vezina nominations, and he finished as the runner-up for the trophy in 2015-16, leading the league in GAA and GSAA.

In the interim season, 2014-15? He was actually... just okay. Average to slightly above average, by just about every metric. Keep this in mind; I'll re-visit his one unremarkable season as TB's #1 in a bit. For now, though, it might be interesting to end our look at Bishop's Tampa tenure with a direct, apples-to-apples comparison between Big Ben and Vaz, in the three seasons they were teammates. Bishop was the clear #1 in 2014-15 and 2015-16; their splits in 2016-17 were more platoon-like due to Bishop's injuries in the 2016 playoffs and following season, which opened the door for Vasilevskiy. Each goalie's collective stats in those three seasons:

  • Bishop: 156 starts, .919 Sv%, 2.27 GAA, 95 GA%-, .571 QS%, 0.12 GSAA per start
  • Vasilevskiy: 81 GS, .915 Sv%, 2.60 GAA, 99 GA%-, .481 QS%, 0.03 GSAA per start

Of course, Vaz hadn't reached his prime yet, but let's be honest—neither had the roster in front of him. Another minor disclaimer: Bishop's numbers are technically undersold here, albeit to a meaninglessly small degree; the data includes his seven mediocre starts with the Kings after his trade, simply because Hockey-Reference wouldn't let me filter it out of that three-season span! In any case, next time someone says Bishop's 2018-19 numbers were inflated by Dallas's "system", kindly remind them that Bishop outperformed Vaz when they were teammates. I'll also have more on that thought in a bit.

Dallas is where Big Ben played his finest hockey. He only broke the 50-start barrier once in his three seasons as the Stars' 1A: 2017-18, when he started 51 games. In his second and third seasons, he started, respectively, 45 and 43 games. Keep in mind that he started, on average, 60+ games in his three full seasons as Tampa's #1 and accumulated 40.1 GSAA across those three seasons, two of which resulted in Vezina nominations.

Recall, also, that GSAA is a cumulative stat that favors goalies who start more games. Be amazed, then, that in his three seasons platooning as Dallas's 1A, Bishop compiled a remarkable 51.6 GSAA. A large chunk of that total came in his sublime 2018-19 campaign, in which he was robbed of the Vezina—and don't worry, I'm about to dive headfirst into making that case. The last thing I want to acknowledge beforehand is just how anemic Dallas's offense was in front of Bishop, just how impressive his win/loss record was in that context, and, most of all, just how badly the Stars needed the level of goaltending Bishop gave them every night.

Big Ben went 74-48-11 with Dallas, capturing 59.8% of available points in the standings in those games. For reference, a team posting a .598 points pace across 82 games would finish with 98 points and comfortably make the playoffs. That number becomes twice as impressive once we quantify just how mediocre-to-atrocious the team's offense performed in support of Big Ben. Here, by season, I present the Stars' goals per game in Bishop's starts, along with Bishop's win/loss record, and where the team's offense would've ranked league-wide with that goals-per-game output (couldn't link individual teams, but the curious reader can quickly select DAL from the drop-down menu):

  • 2017-18: 2.66 goals in support per game (would've ranked 26th of 31 teams, Bishop went 26-17-5)
  • 2018-19: 2.80 goals in support per game (would've ranked 20th of 31; Bishop went 27-15-2)
  • 2019-20: 2.35 goals in support per game (would've ranked 30th of 31; Bishop went 21-16-4)

Those rates sure don't look like the offense of a 45-29-8 team, do they? Yet that would be the record, give or take, for the 98-point team the Stars were with Big Ben between the pipes. We had the offensive production of a lottery team across Bishop's tenure, but his goaltending effectively made us a 100-point team, given a window of variance of a couple wins more or fewer. (Though of course we should credit the team's defense, too, and in particular give a solemn hat-tip to Dobby in 2018-19; 2.80 would've been a luxury for him. He got a meager 2.24 goals per game of support, and the team actually finished 29th in scoring overall.)

So, just how badly did the voters screw up with the Vezina in 2018-19?

If you've managed to read this far, congratulations (and thanks!)—we've come to the real heart and purpose of this post. Bishop finished 2nd league-wide in GAA in 2018-19 with a sparkling 1.98; Binnington led the league at 1.89—in only 30 starts, a smaller sample than Bishop's 45 starts by exactly one-third. Binnington also only faced a cushy 25.8 shots against per 60, compared to the 30.1 Bishop faced. The Blues also allowed the second-fewest high-danger scoring chances in the league, while Dallas allowed the 12th-fewest.

So we can probably get away with saying Bishop's 1.98 GAA was the most impressive that year, if only the second-best. But what categories did Bishop lead the league in? Well, he was first in each of the following:

  • Sv%: .934
  • GA%-: 73 (Bishop allowed nearly 30% fewer goals than a league-average goalie would've on the same number of shots.)
  • GSAA: 32.2 (remember, he put up 40.1 GSAA over 182 starts in his three seasons as Tampa's #1, a stretch in which he received two Vezina nominations. In 2018-19, Bishop saved 32.2 goals above average in 45 starts! Assuming a typical #1 starter starts 60 games in a year, it's comparable to a modern-day baseball player missing 40 games and still leading the majors in Runs Batted In.)
  • QS%: .689 (i.e., Bishop's Sv% was higher than league-average in nearly 70% of his starts! He "only" finished 10th in total Quality Starts, but his 31 Quality Starts was actually the sixth-highest total because of several ties ahead of him. Bishop was only 22nd in total starts; all of the goalies who had more quality starts than Big Ben had between 53 and 66 starts to Bishop's 45.)
  • RBS, Fewest (15+ Starts): 1. One single Really Bad Start in 45 starts, tied with Binnington and Stalock, who started only 30 and 16 games, respectively. Bishop's 45 starts were more than any goalie with one, two, three, or even four Really Bad Starts, and you have to get up to 4 RBS before any other goalies with 40+ starts even begin to show up. You can see in this scatter plot that no other goalie even approached Bishop's ability to avoid bad starts, in terms of total RBS per start:

2018-19 Goalies—Really Bad Starts & Total Starts

Of course, Ws are the sexiest stat of all, though if you've read this far you should be pretty skeptical about the value of Ws as an evaluation of a goalie's season. Nonetheless, we know Vaz won the Vezina because of his 39 wins—and no doubt, his season was exceptional, though his .925/2.40 line isn't quite in the same tier as Bishop's .934/1.98. But recall that Bishop received only 2.80 goals of support per game, and consider that, in support of Vaz, Tampa scored a whopping 3.69 per game (you'll have to select TBL from the dropdown menu). Nearly a full goal per game more than the goal support Bishop got.

The Bolts led the league in scoring at 3.89 goals per game overall, but even 3.69 would've comfortably led the league by about a fifth of a goal per game. To compare the "wiggle room" each goalie got from their team's offenses, the Stars scored only 0.82 more goals per game than the 1.98 Bishop allowed, while the Bolts scored 1.29 more goals per game than Vaz's 2.40 allowed. This difference of 0.47—nearly half a goal—lets us calculate how many fewer goals per game Bishop would've had to allow to match Vasilevskiy's winning percentage: Bishop would've needed a 1.52 GAA to match Vaz's W%, and given the number of shots he faced, a .949 Sv% to reach that 1.52 GAA!

Would Vaz have come anywhere close enough to those numbers in Dallas's "more favorable system" to preserve that 39-wins total, given the weak goal support Bishop got? Of course not. The absurdity of using a team stat like Ws to evaluate goalie play should, at this point, be screamingly obvious—we were all aware of the extreme degree to which a team's goal scoring determines their goalie's winning percentage, but now we've seen the magnitude of that effect laid out.

Still not convinced? Remember way back when I mentioned Bishop's two elite Vezina-nominated seasons with Tampa, with one average season in between them? In his two Vezina-nominated campaigns in Tampa, the Bolts ranked 9th and 12th in goals per game. The year he was entirely mediocre, Tampa led the league in scoring. Wanna guess which season it was where he set his career single-season high in wins?

Ben Bishop, 3 Seasons as TBYep.

As for that "more favorable system"? The Bolts allowed more scoring chances than Dallas, yes, but they were also leading 4-0 by the second intermission as often as not. The Stars, by contrast, were usually still scrapping it out in close games, so score effects explain much of the discrepancy in scoring chances. If you look at high-danger chances, Tampa allowed fewer per game than Dallas, score effects and all—but, while Bishop thus had to face more high-danger chances, he stopped them at a FAR higher rate than Vaz; the Vezina winner had a Sv% on high-danger chances of .747, while Big Ben denied them at an eye-watering .874 rate. We've already quantified just how favorable Tampa's 3.69 goals scored per game were for Vaz, and a couple thousand words ago, we directly compared Bishop and Vaz when they were teammates in the same system, behind the same roster.

So now that we've illustrated that Bishop's job wasn't significantly easier than Vaz's, it's probably safe to dispense with the more-favorable-system silliness. Vasilevskiy is great, and he was really great in 2018-19—but not as great as Bishop was. Not particularly close, really.

I'll wrap up this dissertation post by reiterating how freakishly consistent Bishop was in 2018-19, and more importantly, the surprising degree to which the Stars really did need him to be that consistent every night. The Stars had—in Bishop's starts, at least—the 20th-ranked offense; Vasilevskiy had the top offense. Dallas not only needed Bishop to perform at an elite level, they needed him to do it every game, with an absolute minimum number of poor performances. The Stars' 93 points that year were good for the 7th seed, ahead of the Avs at 90. The Coyotes were the first team out with 86 points—thus, that seven-point difference is the gap by which the Stars made the postseason.

That might sound like a big gap, and in some respects it is, but consider that Dallas would've missed the playoffs with four fewer wins. Imagine if Bishop had, say, three Really Bad Starts instead of one, and one fewer Quality Start in a close game. Had that been the case, those four likely extra losses would've cost Dallas their trip to the playoffs. Bishop's RBS total, relative to total starts, would've still been miniscule—and he still would've led the league in QS%, among goalies with 40+ starts.

His season still would've been elite and worthy of Vezina consideration—but we wouldn't have Klingberg's killshot against the Preds, nor would we have come within a millimeter of the Conference Finals on Benn's 2OT wraparound.

That's how historically great Bishop's 2018-19 season was, but more importantly, how much the team needed it.

In conclusion...

The Kings gave Big Ben to us for a 4th-rounder, and he ended up being one of the most pivotal Dallas players across this particular chapter of the franchise's history. Lehtonen was briefly great, borderline elite, but past his prime by the time he actually had a competitive roster around him. In Bishop, we had an elite goalie and a strong roster, for the first time since Turco in the early/mid aughts. For many young Stars fans, these past few seasons represented the first time they'd ever actually seen the Stars ice a strong roster in front of an elite goalie.

He was easily the most deserving goalie of the Vezina in 2019, and the voters were dead wrong, as they are wont to be. But in any case, Big Ben had a damn fine career, and we were lucky to have him for the brief time we did.

This is a user-created FanPost and does not necessarily reflect the views of SB Nation or Defending Big D. FanPosts are encouraged and may be featured on the front page depending on the quality of the post.