TOI & Points Per Minute?

[member 51464]

Hi all,
I was wondering if there was any easy way to figure out stats of points per minute or say "points per 60 minutes" played or some such thing.

Only reason I wondered was just as another way to compare scorers. Word on the street is that 1st line (and presumably 2nd line) forwards in the 80's played a decent bit more per game, which definitely adds up in the aggregate over the course of the season. So I was wondering how much the gap is in points per minute/game/whatever between players then and now.

Also, what year did the NHL start tracking TOI? If it was fairly recent, then I suppose my idea is pointless.

Click to read more...

 

[seventieslord]

Hi all,
I was wondering if there was any easy way to figure out stats of points per minute or say "points per 60 minutes" played or some such thing.

Only reason I wondered was just as another way to compare scorers. Word on the street is that 1st line (and presumably 2nd line) forwards in the 80's played a decent bit more per game, which definitely adds up in the aggregate over the course of the season. So I was wondering how much the gap is in points per minute/game/whatever between players then and now.

Also, what year did the NHL start tracking TOI? If it was fairly recent, then I suppose my idea is pointless.

I don't think that all 1st line forwards generally played more, it was more that the superstars did.

the NHL only began tracking in the 1999 season, though the 1998 season was somehow tracked too, I'm not sure how public those results are.

A sheet was made up by a clever stats guy a few years back, that used the GF/GA figures in all situations to approximate everyone's situational icetime going back to 1968. It was compared to actual results and had a correlation of 90%, so it's very reasonable.

I'm not sure how far you'll get with this though, because the biggest determinant of scoring opportunity, by far, is PP time, and not ice time in general. That will skew everything, unfortunately, if you're going to just focus on ice time in general.

 

[Derick*]

Yeah you need to split it up into PP, ES, and PK.

 

[leoleo3535]

1997 the NHL started tracking ice time.

 

[reckoning]

97-98 was the first year it was an official stat.

I ran the numbers for that season to see who the leaders would be in even-strength points per 60mins of even-strength play (minimum 25 games):

1. Pierre Turgeon 4.20
2. Peter Forsberg 4.17
3. Jaromir Jagr 4.16
4. Scott Fraser 4.13
5. Mike Modano 4.07
6. Josef Stumpel 4.02
7. Jason Allison 3.84
8. Brett Hull 3.78
9. Pavol Demitra 3.64
10. Wayne Gretzky 3.63

Therein lies the flaw in this stat. Of those players, only Forsberg and Modano would be on the ice taking defensive zone draws to protect a lead late in the game. When you're a player with offensive talent who is never counted on for any checking responsibilty, then you're going to do well in this category because most of your minutes are offence-related. It's no surprise Turgeon leads here. I'd bet he and Craig Janney would've ruled at this stat in their prime.

Both Theoren Fleury and Rod Brind'amour finished among the top 15 scorers that season, but didn't finish very high in this stat because of all their defensive minutes. But I'd still take them over Allison or Demitra.

 

[member 51464]

97-98 was the first year it was an official stat.

I ran the numbers for that season to see who the leaders would be in even-strength points per 60mins of even-strength play (minimum 25 games):

1. Pierre Turgeon 4.20
2. Peter Forsberg 4.17
3. Jaromir Jagr 4.16
4. Scott Fraser 4.13
5. Mike Modano 4.07
6. Josef Stumpel 4.02
7. Jason Allison 3.84
8. Brett Hull 3.78
9. Pavol Demitra 3.64
10. Wayne Gretzky 3.63

Therein lies the flaw in this stat. Of those players, only Forsberg and Modano would be on the ice taking defensive zone draws to protect a lead late in the game. When you're a player with offensive talent who is never counted on for any checking responsibilty, then you're going to do well in this category because most of your minutes are offence-related. It's no surprise Turgeon leads here. I'd bet he and Craig Janney would've ruled at this stat in their prime.

Both Theoren Fleury and Rod Brind'amour finished among the top 15 scorers that season, but didn't finish very high in this stat because of all their defensive minutes. But I'd still take them over Allison or Demitra.

Thanks for breaking down the explanation, that is a good way to look at it.

I was actually only thinking of it in terms of the high-end scorers though. Like top 3 or top 5 across generations. Like how the Top few compared to some of the great scoring seasons in the 80's or whatever.

 

[Uncle Rotter]

Word on the street is that 1st line (and presumably 2nd line) forwards in the 80's played a decent bit more per game, which definitely adds up in the aggregate over the course of the season.

The early to mid 80s, while very high scoring, saw fewer power plays. From 1981-82 to 1984-85 the league average per team was 322 per year. From 1985-86 to 1988-89 it was 388 per year. The more PP there are, the more ice time for the top players. Also during the early 80s, the better you were, the fewer PP you received. The Islanders & the Oilers routinely finished near the bottom in terms of PP awarded.

 

[ssh]

the NHL only began tracking in the 1999 season, though the 1998 season was somehow tracked too, I'm not sure how public those results are.

A sheet was made up by a clever stats guy a few years back, that used the GF/GA figures in all situations to approximate everyone's situational icetime going back to 1968. It was compared to actual results and had a correlation of 90%, so it's very reasonable.

The correlation for total average ice time is around 96%, situational ice times around 92-94%. For season totals they are 99% and 96-99%, respectively. A lot of the errors come from players who’ve only played a few games.

That’s not to say that the model is somehow foolproof but it’s reasonably accurate at least for seasons 1998-2004.

 

[seventieslord]

The correlation for total average ice time is around 96%, situational ice times around 92-94%. For season totals they are 99% and 96-99%, respectively. A lot of the errors come from players who’ve only played a few games.

That’s not to say that the model is somehow foolproof but it’s reasonably accurate at least for seasons 1998-2004.

wait, I'm confused. 1998-2004 have actual situational icetime tracked. And to my knowledge, that sheet uses the actual icetime where available and not calculated ones.

 

[ssh]

wait, I'm confused. 1998-2004 have actual situational icetime tracked. And to my knowledge, that sheet uses the actual icetime where available and not calculated ones.

The "model" was tweaked to match the icetime data available at the time (1998-2004) and it seems to do so reasonably well. Without accurate data from earlier seasons it's difficult to say how it performs. It may be that situations where it has problems are more frequent in the pre-1998 NHL.

And yes, official icetime data was used instead of calculated data as it's the best we've got.
There might be typos or other errors though.