Last project, I was low on Bill Durnan relative to the rest of the voting pool. I spent a bit of time this week looking at him again, and while I think there have been some good arguments made for him and it's clear that his contemporaries rated him pretty highly, I'm not quite fully convinced he deserves to go on the list just yet.
1. The Problem with Bill Durnan's Save Stats
The main reason that I was hesitant about Durnan last time was that there are unofficial save percentage numbers from the 1948-49 and 1949-50 season that don't really make it seem like Durnan stands out all that much. Here is the data I have, with the missing numbers estimated using the recorded SA rates and the goalie's actual GA numbers:
Year | Goalie | Team | Official MIN | Official GA | Record MIN | Recorded SA | Recorded GA | Est Missing SA | Est SA | Est SV | SV% | SA/60 |
1949 | Rayner | NYR | 3480 | 168 | 3120 | 1661 | 143 | 192 | 1853 | 1685 | 0.909 | 31.9 |
1949 | Lumley | DET | 3600 | 145 | 2720 | 1155 | 90 | 374 | 1529 | 1384 | 0.905 | 25.5 |
1949 | Durnan | MTL | 3600 | 126 | 3540 | 1300 | 122 | 22 | 1322 | 1196 | 0.905 | 22.0 |
1949 | Brimsek | BOS | 3240 | 147 | 2640 | 1255 | 121 | 282 | 1537 | 1390 | 0.904 | 28.5 |
1949 | Broda | TOR | 3600 | 161 | 3120 | 1426 | 134 | 219 | 1645 | 1484 | 0.902 | 27.4 |
1949 | Henry | CHI | 3600 | 211 | 2740 | 1393 | 164 | 437 | 1830 | 1619 | 0.885 | 30.5 |
Year | Goalie | Team | Official MIN | Official GA | Record MIN | Recorded SA | Recorded GA | Est Missing SA | Est SA | Est SV | SV% | SA/60 |
1950 | Lumley | DET | 3780 | 148 | 3220 | 1398 | 118 | 243 | 1641 | 1493 | 0.910 | 26.0 |
1950 | Durnan | MTL | 3840 | 141 | 3600 | 1400 | 132 | 93 | 1493 | 1352 | 0.906 | 23.3 |
1950 | Rayner | NYR | 4140 | 181 | 3740 | 1728 | 156 | 185 | 1913 | 1732 | 0.905 | 27.7 |
1950 | Broda | TOR | 4040 | 167 | 4040 | 1632 | 167 | 0 | 1632 | 1465 | 0.898 | 24.2 |
1950 | Brimsek | CHI | 4200 | 244 | 3600 | 1802 | 201 | 300 | 2102 | 1858 | 0.884 | 30.0 |
1950 | Gelineau | BOS | 4020 | 220 | 3360 | 1485 | 182 | 292 | 1777 | 1557 | 0.876 | 26.5 |
Just eyeballing those numbers, while taking into account team quality and coaching, it sure looks to me like Chuck Rayner was the most effective goalie in the league over those two seasons, despite Durnan walking off with the First Team All-Stars.
And if we're not entirely sure whether he was the best in those years, then that's a pretty big issue for the perception of his career as a whole. Durnan didn't have a great year in 1947-48 (it looks like he pretty much collapsed together with the rest of his team when Toe Blake got injured and they missed the playoffs), and that only leaves two non-wartime years left where he might have been the best goalie in the NHL.
Even in 1945-46 and 1946-47 we have a bit of a problem.
@overpass posted a bunch of playoff shot and save data from Durnan's entire playoff career, and from that info we can estimate that he faced a very low 23.2 SA/60 during his playoff career (very much in line with the regular season data above). As a result, we can strongly infer that the Habs were likely very good at preventing shots for Durnan's entire career (other than possibly 1947-48, particularly in the second half), as teams tend to have playoff shot prevention that correlates pretty closely with their regular season shot prevention (especially teams that played multiple postseason series).
If we use that info in an attempt to fill in the gaps and estimate Durnan's results, I honestly doubt that he won a single save percentage title in his whole non-wartime career, other than possibly 1946-47 (where I'd estimate from the playoff results and surrounding seasons that it was likely neck-and-neck between him and Rayner). For a guy who was viewed as the best goalie in the world playing in a six team league, that's not particularly great.
Of course that doesn't tell the whole story, because we need to consider the quality of those shots. There have absolutely been some team situations where a team was preventing shots at the cost of shot quality, giving up fewer shots but with a higher average degree of danger on the ones that did get through. From my coaching stats, here are a few bench bosses who consistently fit this pattern (remember less than 1 is better than average and greater than 1 is worse than average):
Coach | GP | Shot Impact | Save % Impact |
Gerry Cheevers | 376 | 0.84 | 1.02 |
Don Cherry | 480 | 0.87 | 1.07 |
Jim Schoenfeld | 580 | 0.91 | 1.03 |
Bill Peters | 438 | 0.91 | 1.09 |
Terry Crisp | 631 | 0.92 | 1.04 |
Joel Quenneville | 1768 | 0.93 | 1.02 |
Andy Murray | 738 | 0.95 | 1.04 |
Brian Sutter | 1028 | 0.96 | 1.03 |
The problem for Bill Durnan is that his coach was Dick Irvin, who had a positive effect on both metrics throughout his coaching career in the save percentage era:
Coach | GP | Shot Impact | Save % Impact |
Dick Irvin | 550 | 0.91 | 0.92 |
That doesn't mean it's impossible that Irvin had a different system in his earlier career or that Durnan played under different conditions than Irvin's later goalies, but I'm not convinced that's the case. It looks to me like the Habs were great at preventing shots, and a lot of people attributed too much of that defensive impact to Durnan. And as for the "Canadiens stopped winning Vezinas when Durnan left" argument, the reconstructed stats suggest that the Canadiens' shot prevention dried up a lot more in the years following than their save percentage did (in 1950-51 and 1951-52 combined, my data has McNeil at a .906 save percentage on 26.5 SA/60).
Overall, Durnan has some senior years which should count on his record, and we can't completely exclude his performance during the wartime years because it certainly wasn't his fault that there was a global conflict, but I still don't think his actual resume nearly matches up to how highly he's viewed historically.
2. Was Bill Durnan Less Clutch in the Playoffs than Turk Broda?
Having knocked Durnan a bit above, I'm going to defend him in this part. No, I don't think he was less clutch than Turk Broda, I think the difference in team success was mainly down to their respective teammates.
First, here's the breakdown by period:
Bill Durnan, Playoff Career:
Period | MIN | GF | GA | GF/60 | GA/60 |
1 | 900 | 42 | 29 | 2.80 | 1.93 |
2 | 900 | 45 | 40 | 3.00 | 2.67 |
3 | 900 | 59 | 26 | 3.93 | 1.73 |
4 | 170 | 6 | 4 | 2.12 | 1.41 |
Turk Broda, Playoff Career:
Period | MIN | GF | GA | GF/60 | GA/60 |
1 | 2004 | 78 | 71 | 2.34 | 2.13 |
2 | 2034 | 88 | 50 | 2.60 | 1.47 |
3 | 2006 | 84 | 81 | 2.51 | 2.42 |
4 | 343 | 12 | 9 | 2.10 | 1.58 |
Broda has his worst GAA in the third period, while Durnan has the better GAA in OT. The Leafs killed the opposition in the second period, while the Habs look to be really clutch offensively in the third. However that last part is a bit misleading, as we'll soon see.
Let's look in more depth at those third period results:
Bill Durnan, Third Period and Overtime by Score, Playoff Career:
Score | MIN | GF | GA | GF/60 | GA/60 |
Leading 2-4 | 275 | 23 | 7 | 5.03 | 1.53 |
Up 1 | 79 | 6 | 3 | 4.57 | 2.28 |
Tied | 275 | 8 | 9 | 1.75 | 1.97 |
Down 1 | 151 | 9 | 6 | 3.59 | 2.39 |
Down 2-4 | 143 | 5 | 2 | 2.10 | 0.84 |
5+ goal diff (either team) | 46 | 10 | 0 | 12.94 | 0.00 |
Turk Broda, Third Period and Overtime by Score, Playoff Career:
Score | MIN | GF | GA | GF/60 | GA/60 |
Leading 2-4 | 546 | 27 | 14 | 2.97 | 1.54 |
Up 1 | 317 | 16 | 14 | 3.03 | 2.65 |
Tied | 811 | 28 | 21 | 2.07 | 1.55 |
Down 1 | 265 | 10 | 13 | 2.27 | 2.95 |
Down 2-4 | 327 | 11 | 20 | 2.02 | 3.67 |
5+ goal diff (either team) | 83 | 4 | 8 | 2.90 | 5.81 |
In a way Broda was at least a bit clutch, if you define clutch as being better in important moments than in less important moments, since he didn't seem to care at all during garbage time given that 5.81 GAA when one team was leading by 5 or more. Durnan, on the other hand, kept a clean sheet in those situations, although both of these are very small samples.
Broda's goals against results were great in tie games and when his team had built a multigoal lead. However, when the game was separated by a goal late and one of the teams had to push for an equalizer, creating an unbalanced game state, then his stats became quite poor relative to Durnan's. I think the Leafs really protected him with their conservative system in tie games and when they were in front and had the game well in hand, but despite that defensive support Broda was still worse than Durnan at holding 1 goal leads and he was especially unreliable at making saves when his team was pushing to equalize when trailing (I suspect this was one of the only times that the Hap Day Leafs were likely to be giving up significant rush chances against).
As for Durnan, unfortunately that seemingly clutch third-period offence was mainly just coming with his team already in the lead. With the game tied, the Canadiens were not clutch at all offensively. It is possible that they were also very defensive in that game state, like I suspect the Leafs were, and in that case it would be advantage Broda. But Durnan wins basically every situation other than game tied, which makes me suspect that probably wasn't the case.
Oh, and if you're into teams scoring first/giving up the first goal, I didn't break them down exactly but I do have first period with the score tied for both guys:
First Period, Score Tied:
| MIN | GF | GA | GF/60 | GA/60 |
Durnan | 545 | 24 | 14 | 2.64 | 1.54 |
Broda | 1183 | 47 | 43 | 2.38 | 2.18 |
Again, advantage Durnan.
In short, if I'm a Durnan skeptic, then I'm a straight up non-believer when it comes to Turk Broda. All of the points I mentioned in part 1 about reconstructed save percentages apply just as much to him as they do to Durnan, and I'm even more confident on the coaching effects in Toronto. It's great for Broda that he played behind a great defensive team in the playoffs and he gets the winning cred that goes with that, but I'm far from convinced he was actually a better goalie than the competition here. It is entirely possible that he ends up ranked last on my list for the second straight week.