Orr's TOI

[BobbyAwe]

There's been threads, (or at least questions that came up), about the TOI of past greats, before the stat was kept. This because of many comments from "old timer's" who said that so and so was "always on the ice". "Always" is admittedly hyperbole, but how much TOI some past greats may have logged in an earlier era has been the source of a lot of speculation, and I think that no player's TOI is more wondered about than Bobby Orr's.

Well, I came across something that I think is pretty significant (?) I watched an entire Bruin's game on YouTube a while ago, from 1974. One of the commentators mentioned, between the 2nd and 3rd periods, that Orr had logged 23 minutes after two. That figures to 34 and 1/2 minutes for the game assuming he was on ice at the same rate in the 3rd period.

Now that's just one game, of course, so there may have been games where he logged even more minutes, and/or maybe he was on ice for MORE than 11 and 1/2 minutes in the 3rd of that particular game, but I think we have a realistic estimate there in general, of what they meant when they said he was "always" on the ice.

It's a safe assumption he logged over 34 minutes in that game, and it WASN'T the 7th game of a playoff, nor do I think we can assume that this game was the most TOI he ever had, so is it possible that he could have been close to 40 minutes in whatever game WAS the most TOI he ever had?

Maybe this can help us differentiate between legend and fact regarding the subject, or at least speculate a little more realistically?

Click to read more...

 

[Midnight Judges]

Maybe you could link the game and the time when the commentator said it? That would allow for some additional context.

 

[BobbyAwe]

I think it's Bruins vs. Blackhawks...I'll try to find the time on the comment.

 

[Hockey Outsider]

Here's an old post (from @reckoning in 2011) that talks about Orr's ice time - Bobby Orr's icetime during his prime

There's little doubt that Orr sometimes played 40 minutes in a game. But it's virtually impossible that he averaged that much ice time - as that would imply that the Bruins scored fewer goals per minute at ES when he's on the ice, compared to when he's not - which simply doesn't make sense.

It sounds like 35 minutes per game (which, obviously, is still a ton of ice time) is near the upper limit of what's plausible. (EDIT - I'm re-reading that thread. Lots of good discussions there. It looks like Orr's estimated ice time, per the ice time formula developed by Iain Fyffe, is around 29-30 minutes, on average).

 

[ted2019]

Here's an old post (from @reckoning in 2011) that talks about Orr's ice time - Bobby Orr's icetime during his prime

There's little doubt that Orr sometimes played 40 minutes in a game. But it's virtually impossible that he averaged that much ice time - as that would imply that the Bruins scored fewer goals per minute at ES when he's on the ice, compared to when he's not - which simply doesn't make sense.

It sounds like 35 minutes per game (which, obviously, is still a ton of ice time) is near the upper limit of what's plausible. (EDIT - I'm re-reading that thread. Lots of good discussions there. It looks like Orr's estimated ice time, per the ice time formula developed by Iain Fyffe, is around 29-30 minutes, on average).

The 29-30 minutes per game seem a lot more plausible. That is what some of the very top #1 defensemen average now.

 

[Dennis Bonvie]

The 29-30 minutes per game seem a lot more plausible. That is what some of the very top #1 defensemen average now.

No one averaged over 26 minutes per game last season.

 

[reckoning]

Once I tried tallying Orr's icetime in Game 6 of the '74 Final against Philadelphia, since the game is on the Flyers DVD.
The Bruins went most of the game with two pairings: Al Sims-Bobby Orr, and Carol Vadnais-Dallas Smith. There was a 5th defencemen, Daryl Edestrand, who was used very sparingly (only about 5 min in the game), and a 6th defenceman dressed in the lineup, Al Simmons, who didn't get any shifts.

1st Period: I have Orr's TOI at 10:08
2nd Period: Unfortunately a big chunk of this period is missing from the DVD. The first five and a half minutes are absent, as well as over four minutes later in the period. But in the 10:14 of play that is there, Orr is on the ice for all but 34 seconds. Simply incredible. I could've used a prorated estimation for the missing time, but at the start of the third period the announcers mention that he played 14 and a half minutes in the second period, so that's probably more accurate. Therefore, let's give him 14:30 for the second period.
3rd Period: I timed Orr at 8:40. There are 45 seconds missing from the game, and it's a safe bet that Orr played them because the Vadnais-Smith pairing was on for the shifts before and after it. So that gives Orr 9:25 for the period.

The three periods added together gives him 34:03 for the game. It probably could have been higher, but he took a penalty late in the third when the Bruins likely would have had him on the ice for the last minutes of the game, seeing as how they were on the brink of elimination

 

[Dennis Bonvie]

Once I tried tallying Orr's icetime in Game 6 of the '74 Final against Philadelphia, since the game is on the Flyers DVD.
The Bruins went most of the game with two pairings: Al Sims-Bobby Orr, and Carol Vadnais-Dallas Smith. There was a 5th defencemen, Daryl Edestrand, who was used very sparingly (only about 5 min in the game), and a 6th defenceman dressed in the lineup, Al Simmons, who didn't get any shifts.

1st Period: I have Orr's TOI at 10:08
2nd Period: Unfortunately a big chunk of this period is missing from the DVD. The first five and a half minutes are absent, as well as over four minutes later in the period. But in the 10:14 of play that is there, Orr is on the ice for all but 34 seconds. Simply incredible. I could've used a prorated estimation for the missing time, but at the start of the third period the announcers mention that he played 14 and a half minutes in the second period, so that's probably more accurate. Therefore, let's give him 14:30 for the second period.
3rd Period: I timed Orr at 8:40. There are 45 seconds missing from the game, and it's a safe bet that Orr played them because the Vadnais-Smith pairing was on for the shifts before and after it. So that gives Orr 9:25 for the period.

The three periods added together gives him 34:03 for the game. It probably could have been higher, but he took a penalty late in the third when the Bruins likely would have had him on the ice for the last minutes of the game, seeing as how they were on the brink of elimination

Damn that Art Skov!

As Johnny Most would have said, "Orr was called for a penalty, he didn't commit one".

 

[seventieslord]

This is why I don't believe that the 35-40 minute guess can be accurate (pasted from another thread)

Start by using the TOI spreadsheet and multiplying every Boston defenseman's ESTOT/GP by their GP. This gives us a number representing their approximate total ESTOI. Then, using overpass' spreadsheet, plug in the ESGF and ESGA for every Boston defenseman for every season. This is the easiest way to get this because overpass' numbers have approximate pluses and minuses from SHG removed so these are much more accurate than just subtracting PPGA from TGA and PPGF from TGF.

So from there, you have total ESTOI and total ESGF and ESGA. Now set up a formula where you calculate every player's per minute rates (or per-60 rates). As a side note, I realize that this is sort of folding the formula for TOI calculation back onto itself, but:

1. The point is not to come to real conclusions about each player's GF and GA per 60, only hypotheticals
2. Since every player has a unique ratio, their GF and GA rates would be unique even if their per-minute totals might work out fairly closely

Here's what I get for every defenseman who played at least two 40+ game seasons in that time:

AWREY, DON 3.61 2.79 6.4
DOAK, GARY 2.25 2.01 4.26
EDESTRAND, DARRYL 3.22 2.3 5.52
GREEN, TED 3.54 2.96 6.5
ORR, BOBBY 4.71 2.17 6.88
SIMS, AL 4 2.32 6.32
SMITH, DALLAS 3.81 2.66 6.47
SMITH, RICK 3.3 1.97 5.27
VADNAIS, CAROL 3.88 3.12 7

So, Already at about 30.9 minutes a game, Orr is not the "highest event" player, that distinction goes to Vadnais (which may have been true, but was probably not). If we start adding minutes to Orr, though, what do you think happens to these numbers? Remember, we can't add any more time to his PP or PK - those are very limited time situations and we have a clear answer for what percentage of the time he spent there. So if we want to get him to 35.0 per game, we have to add 4.1 minutes of even strengt time to his averages every season. If I do that, his ESGF/60 goes to 3.99 and his ESGA/60 goes to 1.83, with a total of 5.82. This would make Orr a "low event player" - Lower than Don Awrey, Ted Green, Al Sims, and Dallas Smith, and much lower than Carol Vadnais.

Keep in mind too, that if we take 2,296 (4.1 X 560) even strength minutes and give them to Orr over 8 seasons, that comes off the ES totals of all the other defensemen, on a roughly proportional basis. The very rough average for ESTOI/GP for a non-orr, 40+ game defenseman in this time is 18.9 and there are 3.5 of these per season. We're taking away about 6.3% of their even strength time. As we do that, their per minute rates for goals go up about 6.6% each. So in a situation where Orr is playing 35 minutes, we have numbers like the following:

AWREY, DON 3.85 2.97 6.82
DOAK, GARY 2.40 2.14 4.54
EDESTRAND, DARRYL 3.43 2.45 5.88
GREEN, TED 3.77 3.16 6.93
ORR, BOBBY 3.99 1.83 5.82
SIMS, AL 4.26 2.47 6.74
SMITH, DALLAS 4.06 2.84 6.90
SMITH, RICK 3.52 2.10 5.62
VADNAIS, CAROL 4.14 3.33 7.46

Sure, I'd like to believe that GA/60 number for Orr, but there's no way that GF/60 number is correct. He's not on the ice for fewer goals than Sims, Smith and Vadnais, and he's sure as hell not a lower event player than everyone except Smith (Barely) and Doak.

I know we'll never have a 100% conclusive answer on this, but what looks like it's closer to the truth?

Scenario 1: Orr is the highest-event Boston defenseman along with Carol Vadnais, only with a much better ratio. He sees, by far, the most goals go in for the Bruins while on the ice, and nearly the lowest amount against, aside from conservative, low-event players Doak and Smith. He averages 30.9 minutes which is a higher amount than anyone else has been recorded or estimated to have played over an 8-season span in the past 52 seasons.

Scenario 2: Orr is the 3rd-lowest event Boston defenseman, barely ahead of Rick Smith in this regard. He sees Sims, Smith and Vadnais amass more goals-for than he gets, and he is far and away the best at stopping pucks from going in at his own end. He plays 35 minutes a game over 8 seasons, a good 5.7 (19%) more than anyone else (Ray Bourque, 29.3) has been recorded or estimated to have averaged over an 8-season span in the past 52 seasons.

This is why I can't get on board with Orr playing 35 (or more) minutes. It makes him a low-event player and we know for a fact he wasn't.

 

[Iron Mike Sharpe]

In Orr's era, NHL teams had 17 rather than 18 skaters on their roster, so oftentimes teams dressed only 5 D and rolled two pairs.

 

[ted2019]

No one averaged over 26 minutes per game last season.

I stand corrected. The last player to approach 29 minutes per was Erik Karlsson in 2015-16 with 28:58. Ryan Suter in 2014-15 had 29:04TOI and in 2013-14, Suter had 29:25 TOI.

 

[Hockey Outsider]

Now that we have GF and GA data for 1960 to 1967, it should be possible to estimate the ice time for those seasons. Has anyone attempted to do that already?

 

[seventieslord]

Now that we have GF and GA data for 1960 to 1967, it should be possible to estimate the ice time for those seasons. Has anyone attempted to do that already?

I've done really crude estimates but not with the precision that Iain Fyffe had (taking unique team penalty situations into account, etc)

 

[Michael Farkas]

I have done a few Bobby Orr games manually...

Boston at Toronto - Regular Season (Jan. 6, 1968)
Bobby Orr 32:36

##

Boston at Toronto - Regular Season (Mar 14, 1970)
Bobby Orr (shift chart)
1st Per:
- 1:34 (0:00 to 1:34)
- 1:00 (4:04 to 5:04)
- 0:30 (all SH)
- 0:34* (6:54 to 7:29, shift contained a stoppage = *)
- 2:50* (shift contained a stoppage 2:06 into it)
- 1:16
- 1:48 (15:21 to 17:09, 1:30 was SH)
- 0:58 (0:58 to 0:00)
Tot: 8 shifts, 10:30 (8:30 EV, 0:00 PP, 2:00 SH)
2nd Per:
- 1:16 (0:00 to 1:16)
- 3:45 (2:17 to 6:02, including 2:00 PP)
- 2:38 (7:10 to 9:48 1:53 of PP)
- 2:20 (11:05 to 13:25)
- 1:17 (14:48 to 16:05)
- 1:15* (17:43 to 18:58)
Tot: 6 shifts, 12:31 (8:22 EV, 3:53 PP, 0:00 SH)
3rd Per:
- 1:18 (0:00 to 1:18)
- 1:12*
- 1:23 (0:31 was SH, 4:47 to 6:10)
- 1:52
- 1:42*
- 1:36** (14:12 to 15:48)
- 2:15**
- 0:22
Tot: 8 shifts, 11:40 (11:09 EV, 0:00 PP, 0:31 SH)
Game Tot: 22 shifts, 34:41 (1:35 avg)

##

Boston @ Philadelphia - May 19, 1974 (Stanley Cup Final - Game 6)
Bobby Orr (shift chart)
1st Per:
- 2:17* (2:00 PP, 0:32 to 2:49)
- 1:53
- 1:08
- 2:03* (2:00 PP, 10:18 to 12:21)
- 0:24 (all SH, 13:58 to 14:22)
- 1:42*
- 1:24** (18:36 to 20:00)
Tot: 7 shifts, 10:51 (7:27 EV, 4:00 PP, 0:24 SH)
2nd Per:
[Played about 3:15 of the first 5:36...there isn't video for this time...but the broadcast said two distinct times that he played 14 to 15 minutes of the 2nd period before settling on 14:30]
- 5:17 (5:36 to 10:53, at least 1:05 was SH, at least 0:55 was PP)
- 3:34 (2:00 PP, 15:02 to 17:36)
- 2:14 (2:00 SH, 17:46 to 20:00)
Tot: 4+ shifts, 14:30 (at least 2:55 of PP, at least 3:05 of SH, likely about 7 or so minutes at EV)
3rd Per:
- 1:27**
- 1:12
- 1:11* (7:02 to 8:12)
- 3:04** (2:00 PP, 10:22 to 13:26)
- 0:52
- 0:48 (16:50 to 17:38)
- 0:22 (19:38 to 20:00)
Tot: 7 shifts, 8:56 (6:56 EV, 2:00 PP, 0:00 SH)
Game Tot: 18+ shifts, 34:17 (1:54 avg)

 

[Big Phil]

I'll try to do math in the middle of the night here. But let's take a year where I would assume if anything he gets as much ice time as ever. 1971. The Bruins were running and gunning more than ever that year, both Orr and Esposito had record seasons and the Bruins scored 399 goals, only bested by Gretzky's Oilers. So it is easy to assume he is getting loads of ice time that year.

Of the 399 goals he is on the ice for 258 of them. Of the 207 goals against he is on the ice for 85 of them.

So..................he is on the ice for 65% of the goals the Bruins scored and on the ice for 41% of the goals against.

65% of a regular NHL game is almost literally 40 minutes or two periods. If Orr is to play regularly at 40 minutes - which we know is impossible - that means the Bruins scored the same goals per minute when he wasn't on there, which wouldn't have happened. Sure Esposito and others were still there, but this is Orr we are talking about, there is no way it is the same. So let's take off 40 minutes per game off the table right now.

35 minutes per game means...............58% of the game. To me that is still too unrealistic on average. It still means he was "only" on the ice for 65% of the goals. For a player to be as great as him there has to be a bigger separation. 30 minutes per game sounds about right. That's 50% of the game. So the Bruins scored 65% of the goals while he was on the ice and the other 50% of the time got the other 35% of their goals. I can get behind that for sure. You still had Esposito, Bucyk, Hodge, Stanfield, McKenzie, etc. scoring goals for them and they still did it without Orr on the ice for sure, but I can see that sort of discrepancy with Orr off the ice.

I don't know if this makes sense, nor is it scientific by any stretch but it is a start. I think regularly playing 30 minutes a game is likely what he did.

Here is another good example. Ice time started getting tracked in 1998-'99. If you take the best defensemen in the NHL year after year you can see what they played per game. Here is the best each of them did. I could add the playoffs but overtime games sort of throw that off a bit:
Nick Lidstrom - 29:20 in 2003.
Chris Pronger - 30:36 - 1999
Scott Niedermayer - 27:31 - 2007
Ray Bourque - 29:31 - 1999
Brian Leetch - 29:52 - 1999
Zdeno Chara - 27:11 - 2006
Drew Doughty - 29:00 - 2015
Al MacInnis - 29:07 - 1999

So we can sort of judge it from that. These are the best from any of them. Here is another good thing to judge. Ray Bourque in 1998 logged two games with heavy minutes in the playoffs. This was Games 2 and 3 vs. Washington. 42:20 and 45:40 respectively. Both went into double overtime. Those games went 81 and 86 minutes respectively. That's slightly over half the game each time. This is a playoff game. So come to think of it, is it crazy that a prime Orr logs 34 minutes in a particular playoff game? Or regular season? I am sure he did it from time to time, especially considering as good as Bourque is there is no defenseman more than Orr you want on the ice that often. I think back in the 1980s and early 1990s you might have seen Bourque average 30 minutes a game, just guessing. So yeah, Orr would have too. But I am thinking early 30s at best, not even mid 30s and definitely not 40.

The thing with star players is that we always assume they are on the ice more often because we are looking for them and because they make plays a lot better and are more noticeable out there.

 

[seventieslord]

I'll try to do math in the middle of the night here. But let's take a year where I would assume if anything he gets as much ice time as ever. 1971. The Bruins were running and gunning more than ever that year, both Orr and Esposito had record seasons and the Bruins scored 399 goals, only bested by Gretzky's Oilers. So it is easy to assume he is getting loads of ice time that year.

Of the 399 goals he is on the ice for 258 of them. Of the 207 goals against he is on the ice for 85 of them.

So..................he is on the ice for 65% of the goals the Bruins scored and on the ice for 41% of the goals against.

I agree with your conclusion but not necessarily how you got there. Orr played for all intents and purposes 100% of the PP ice time, and what appears to be about 70% of penalty kill time. Goals for and against happen at about twice the usual rate on special teams. So if you're just doing a rough calculation that treats all time and goals the same, a player who plays an inordinate amount of time on special teams (as Orr did) will get inflated - as you saw in the numbers you calculated.

Earlier in this thread I show the math for why 35 minutes doesn't make sense and why 30-31 does.

 

[plusandminus]

I have done a few Bobby Orr games manually...

Boston at Toronto - Regular Season (Jan. 6, 1968)
Bobby Orr 32:36

##

Boston at Toronto - Regular Season (Mar 14, 1970)
Bobby Orr (shift chart)
1st Per:
- 1:34 (0:00 to 1:34)
- 1:00 (4:04 to 5:04)
- 0:30 (all SH)
- 0:34* (6:54 to 7:29, shift contained a stoppage = *)
- 2:50* (shift contained a stoppage 2:06 into it)
- 1:16
- 1:48 (15:21 to 17:09, 1:30 was SH)
- 0:58 (0:58 to 0:00)
Tot: 8 shifts, 10:30 (8:30 EV, 0:00 PP, 2:00 SH)
2nd Per:
- 1:16 (0:00 to 1:16)
- 3:45 (2:17 to 6:02, including 2:00 PP)
- 2:38 (7:10 to 9:48 1:53 of PP)
- 2:20 (11:05 to 13:25)
- 1:17 (14:48 to 16:05)
- 1:15* (17:43 to 18:58)
Tot: 6 shifts, 12:31 (8:22 EV, 3:53 PP, 0:00 SH)
3rd Per:
- 1:18 (0:00 to 1:18)
- 1:12*
- 1:23 (0:31 was SH, 4:47 to 6:10)
- 1:52
- 1:42*
- 1:36** (14:12 to 15:48)
- 2:15**
- 0:22
Tot: 8 shifts, 11:40 (11:09 EV, 0:00 PP, 0:31 SH)
Game Tot: 22 shifts, 34:41 (1:35 avg)

##

Boston @ Philadelphia - May 19, 1974 (Stanley Cup Final - Game 6)
Bobby Orr (shift chart)
1st Per:
- 2:17* (2:00 PP, 0:32 to 2:49)
- 1:53
- 1:08
- 2:03* (2:00 PP, 10:18 to 12:21)
- 0:24 (all SH, 13:58 to 14:22)
- 1:42*
- 1:24** (18:36 to 20:00)
Tot: 7 shifts, 10:51 (7:27 EV, 4:00 PP, 0:24 SH)
2nd Per:
[Played about 3:15 of the first 5:36...there isn't video for this time...but the broadcast said two distinct times that he played 14 to 15 minutes of the 2nd period before settling on 14:30]
- 5:17 (5:36 to 10:53, at least 1:05 was SH, at least 0:55 was PP)
- 3:34 (2:00 PP, 15:02 to 17:36)
- 2:14 (2:00 SH, 17:46 to 20:00)
Tot: 4+ shifts, 14:30 (at least 2:55 of PP, at least 3:05 of SH, likely about 7 or so minutes at EV)
3rd Per:
- 1:27**
- 1:12
- 1:11* (7:02 to 8:12)
- 3:04** (2:00 PP, 10:22 to 13:26)
- 0:52
- 0:48 (16:50 to 17:38)
- 0:22 (19:38 to 20:00)
Tot: 7 shifts, 8:56 (6:56 EV, 2:00 PP, 0:00 SH)
Game Tot: 18+ shifts, 34:17 (1:54 avg)

To me, the quoted post is the probably most interesting post in this thread because it actually lists the factual ice times.

(When I was new on this board, around 2011, I also came up with about 34 minutes.)

When people try to use GF and GA shares to estimate TOI on ES, PP and SH, they tend to neglect that different players may have much higher or lower GF or GA per 60 minutes. To stereotype, think Coffey and Housley vs defensive minded players. (The example does not necessarily hold up in reality.) But regardless of player, these things vary a lot. It varies far more than some might be able to comprehend. And we should not assume we know how, because we don't know.

 

[seventieslord]

To me, the quoted post is the probably most interesting post in this thread because it actually lists the factual ice times.

(When I was new on this board, around 2011, I also came up with about 34 minutes.)

When people try to use GF and GA shares to estimate TOI on ES, PP and SH, they tend to neglect that different players may have much higher or lower GF or GA per 60 minutes. To stereotype, think Coffey and Housley vs defensive minded players. (The example does not necessarily hold up in reality.) But regardless of player, these things vary a lot. It varies far more than some might be able to comprehend. And we should not assume we know how, because we don't know.

Which is exactly why Orr could not have averaged 34-35 minutes per game in his prime. Read my post above.

 

[Filthy Dangles]

This is off topic to the TOI thing but i'll use this thread since it's here rather than starting a new one since it's about Orr. For some reason, this clip of Lanny McDonald talking about him and a particular play has always stuck with me in terms of his greatness.

Talking about being dumbfounded and shell shocked after trying to flush Bobby from behind the net twice unssuccesfully while Toronto was on the man advantage....



Seems like a longshot, but was wondering if there's any clip of this play being mentioned or another one like it?

 

[BraveCanadian]

I have done a few Bobby Orr games manually...

Very interesting to see, thanks.

 

[Big Phil]

I agree with your conclusion but not necessarily how you got there. Orr played for all intents and purposes 100% of the PP ice time, and what appears to be about 70% of penalty kill time. Goals for and against happen at about twice the usual rate on special teams. So if you're just doing a rough calculation that treats all time and goals the same, a player who plays an inordinate amount of time on special teams (as Orr did) will get inflated - as you saw in the numbers you calculated.

Earlier in this thread I show the math for why 35 minutes doesn't make sense and why 30-31 does.

I agree with it being 30-31 on average throughout his career. That sounds about right.

 

[tarheelhockey]

I agree with it being 30-31 on average throughout his career. That sounds about right.

Another way to think about it (a mathematically less rigorous one):

Your numbers and Mike’s numbers both suggest that Orr skated close to 35 minutes on a busy night. But not every night is a busy night. Sometimes you have a blowout game and the benchwarmers get an extra shift. Sometimes you spend 10 minutes in the box. Sometimes you twist an ankle and miss the rest of a period.

If 35 is a busy night, then an average around 30, maybe 32 seems about right. That seems to be where all the estimates are pointing, even if we use several different paths to get there.

 

[Big Phil]

Another way to think about it (a mathematically less rigorous one):

Your numbers and Mike’s numbers both suggest that Orr skated close to 35 minutes on a busy night. But not every night is a busy night. Sometimes you have a blowout game and the benchwarmers get an extra shift. Sometimes you spend 10 minutes in the box. Sometimes you twist an ankle and miss the rest of a period.

If 35 is a busy night, then an average around 30, maybe 32 seems about right. That seems to be where all the estimates are pointing, even if we use several different paths to get there.

I agree. Regardless of era it just simply isn't feasible for even the greatest defenseman in NHL history to be on the ice that much. Even Orr needed rest, you need a breather. Plus, the other thing not mentioned, even though Orr was partnered with some of these guys it isn't as if the Bruins had nothing but schlubs as their remaining defensemen. Ted Green, Don Awrey, Dallas Smith, Carol Vadnais, Rick Smith, etc. These weren't all necessarily the same season, but either way even without Orr it is at the very least a serviceable defense corps that would get their own regular ice time when Orr isn't on.

It is a bit like when you hear old stories about Phil Esposito and how you had to practically use a hook to get him off the ice. Well, that would be true to an extent, I know Esposito liked the long shift, but even when you break down his numbers by using something similar like the goals for stat that he was on the ice for then you can clearly see he wasn't getting 30 minutes a game by any stretch of the imagination.

I remember Steve Stamkos back in his prime and for whatever reason I was curious at how much ice time he got since he along with Crosby, Malkin and Ovechkin was the best player in the NHL. Stamkos never got more than 22:01 in a season at his best. Crosby was slightly below 22:00. Ovechkin 23:06. Malkin 22:31. That's it. How is that possible, I thought? It just seems like these guys are on the ice more than that. Well, they stand out, plus they get power play time. So it always seems like they are on there more often. The first season ice time was an official stat was when Jagr racked up an almost unheard of 25:51 as a forward. This was his Hart year in 1999 when the Pens were literally being carried on his back and he was getting a ton of ice time. So that made sense, because a guy with no other stars on the team in his prime is the perfect storm for high ice time. But even then, still less than 26 minutes. Mario has his best recorded time with 24:22 in 2001. I would assume his best time would be around 1989, when the team wasn't so stacked and he was at his best and playing in all situations. So he is probably being relied upon a lot more at that time than in 1993. So if I were to take a guess, I would say Mario in 1989 was on the ice 27-28 minutes a game at the absolute most. The reason being, he was on the ice for 254 goals for, while the Pens scored 347. That's being on for a lot of goals and I don't think it is crazy to think the rest of that Pens team only scored 93 goals for the remaining 32 minutes per game. Just a hunch.

But the point is nowhere can I imagine a star forward playing 30 minutes a game regularly. Just like it is hard to see Orr at 35-40 minutes a game.

 

[Hockey Outsider]

There are always questions about how accurate the ice time prediction formula is. I stumbled upon an old email (from January 2005) that I sent to Iain Fyffe, where I used his formula to estimate ice time for (at the time) current players:

"I was doing some work on the "Estimating Ice Time" essay you have on your site. You tested your results on a sample of 93 players. I was curious to see how well it would hold up over a larger sample.

I compared estimated total ice time to actual total ice time for every player I had data for from 1999-2003. There were some players that didn't match up (most likely due to retired players getting removed from the NHLPA's database and/or spelling errors), but overall I compared 2,712 players. Your system passed with flying colors.

The correlation coefficient (R) between actual and estimated ice time is: 0.987. Very impressive.

The standard error is 86.8 while the average is 1,026.7. So the standard error as a percentage of the mean is 8.4%. Again very impressive.

I made a slight change in the method used to normalize ice time (I'm still fiddling with it and trying to find the optimal levels-- I'll let you know once I'm done). Other than that I used the exact system that you did.

So, in case you were ever wondering, your ice time estimate system does an excellent job of holding up over an even greater sample size."

Three caveats. One, the data that I used for this analysis is long gone, so I can't show the step-by-step results in case anybody asks. Two, I mentioned a "slight change" in the approach to ice time normalization, but I don't know exactly what the change was. Three, it's unclear from my email whether I was looking at total ice time (I suspect that was the case) or getting into situational (ES/PP/SH) splits.

Still, I think this demonstrates conclusively that the ice time estimation formula (developed by Iain, and modified by others) is, over large sample sizes, highly accurate.

 

[MadLuke]

If I understand this correctly:

1942-43Because of wartime restrictions on train scheduling, regular-season overtime was discontinued on November 21, 1942. Player limit reduced from 15 to 14. Minimum of 12 men in uniform abolished.

1983-84Five-minute sudden-death overtime to be played in regular-season games that are tied at the end of regulation time.

Was there no regular season over time during all of Orr career, I think that would play some role in average toi and can make comparing to the stars D that played in era with a lot of 5v5 overtime game going all the way through era a little bit misleading. Same for before/after TV ads break change.

That said if there is 12 tie in a season + 10 games that ended during an OT, that could still add less than 1 minute a game.

Has for events, in 2014-2015 Karlsson was on the ice for 131 goals of the 238 sens goal (55%) and 115 of the 215 against them (53%), while playing 27:15 minute a night (82 games), I imagine the sens played just a little bit above 61 minutes a night.

I am not sure if it is a good comparison of those deep Bruins team, but can give an idea of the possible, it is quite possible for a big offensive D to be involved in more than 55% of the GF while playing less around 45% of the time, the special team making that easier if you play in a era they are common.