Accuracy of using a players powerplay goal on ice totals to determine powerplay ice time

[TheStatican]

This thread extends the concepts previously explored here but with a specific and more in-depth focus on powerplay time exclusively. It is reasonable to assume that a player's powerplay goals on ice closely correlates with their actual powerplay Time On Ice (TOI). However the crucial question lies in understanding the degree of correlation, identifying the margin of error and determining whether these numbers can be reliably used to estimate a player's actual powerplay PP TOI in seasons predating the tracking of TOI (pre-1997-98). To address these questions, I have compiled a list comprising 100 full or near full-length player-seasons, ensuring a relatively large sample size for data verification. This list includes 52 of the top 100 and 89 of the top 200 seasons based on powerplay ice time per game(for forwards) since tracking began. Ice time figures were sourced from either the NHL.com website or Quanthockey. Strangely, some totals differ between the sites by a single second(rounding variations?)

The majority of player seasons in the sample were 82-game samples (67 out of 100), with only six seasons falling below 70 games. These exceptions were included due to their significance, such as having a notably high percentage of powerplay goals on ice for or featuring a substantial powerplay TOI. Analyzing seasons with missed games is challenging and very time consuming as first I needed to find out exactly which games the players missed and then I had to removed the team totals(PP goals & PP minutes) from those games from the teams full season totals, all of which was done using manually inputted dates on NHL.com. However, players with missed games could offer additional insights into how a team's powerplay performs with and without these players, a potential avenue for future exploration.

Originally, I organized the list by descending order of PP TOI. However I later reordered it based on the highest to lowest percentage of a team's powerplay goals a player was on the ice for. In my opinion this ordering makes more sense, since we don't actually know the PP TOI in seasons prior to tracking - we can ascertain the % of the teams total powerplay goals they were on ice for.

Definitions:
"PP" - where used is short for "powerplay".
"Pts" - Speaks for itself, I've listed the players point totals for that season just to help quickly identify the season in question
"Rk" - Is the all time rank of the season by PP TOI(for forwards). Any season without a Rk number is still a high powerplay TOI season but not quite within the top 200 all-time. These seasons were included due to their significance i.e. an Art Ross winning season or other very high scoring year.
"Ratio" - Is the ratio of a players estimated PP ice time which is calculated by dividing the number of PP goals the player was on the ice for verses the teams total powerplay goals and using the team totals only in the games they played in.
"Est. TOI" - Is how much ice time a player would have had if their PP ice time was proportional to the percentage of the teams PP goals they were on the ice for.
"PP+/-" - Is the number of PP goals more or less than expected that the player was on the ice for, based on their actual PP ice time.
"Accuracy" - Is the accuracy of the players ratio to the average ratio of all player-seasons above 60% team powerplay goals on ice for, which was 1 to 1.17

Rk​

Name​

Season​

Pts​

Team PPG's​

Player PPG's​

Team PP TOI​

Est. TOI​

Actual TOI​

% team PP goals​

% team PP mins​

Ratio​

Accuracy(in relation to 1.17)​

​

Draisaitl​

22-23​

128​

87​

87​

04:56​

04:56​

03:58​

100.0%​

80.4%​

1.24​

6.3%​

​

McDavid​

20-21​

105​

48​

47​

04:50​

04:43​

04:11​

97.9%​

86.6%​

1.13​

-3.3%​

​

McDavid​

22-23​

153​

89​

87​

04:56​

04:49​

03:55​

97.8%​

79.4%​

1.23​

5.2%​

51​

Malkin​

08-09​

113​

62​

60​

07:04​

06:50​

05:33​

96.8%​

78.5%​

1.23​

5.3%​

​

Draisaitl​

20-21​

84​

48​

46​

04:50​

04:37​

04:13​

95.8%​

87.2%​

1.10​

-6.1%​

148​

Ovechkin​

13-14​

79​

66​

63​

05:19​

05:04​

05:02​

95.5%​

94.7%​

1.01​

-13.8%​

76​

Ovechkin​

08-09​

110​

84​

80​

06:06​

05:48​

05:24​

95.2%​

88.5%​

1.08​

-8.0%​

28​

Kariya​

98-99​

101​

83​

79​

07:32​

07:10​

05:44​

95.2%​

76.1%​

1.25​

6.9%​

​

Draisaitl​

19-20​

110​

59​

56​

04:25​

04:11​

03:50​

94.9%​

86.8%​

1.09​

-6.5%​

34​

Kovalchuk​

07-08​

87​

50​

47​

06:30​

06:06​

05:43​

94.0%​

87.9%​

1.07​

-8.6%​

141​

Ovechkin​

09-10​

109​

63​

59​

05:49​

05:26​

05:03​

93.7%​

86.8%​

1.08​

-7.8%​

2​

Kovalchuk​

03-04​

87​

59​

55​

08:23​

07:48​

06:51​

93.2%​

81.7%​

1.14​

-2.5%​

189​

Ovechkin​

19-20​

67​

41​

38​

05:23​

04:59​

04:53​

92.7%​

90.7%​

1.02​

-12.7%​

49​

Selänne​

98-99​

107​

76​

70​

07:32​

06:56​

05:33​

92.1%​

73.7%​

1.25​

6.9%​

6​

Weight​

00-01​

90​

59​

54​

08:04​

07:22​

06:38​

91.5%​

82.2%​

1.11​

-4.9%​

​

Malkin​

11-12​

109​

53​

48​

05:27​

04:56​

04:21​

90.6%​

79.8%​

1.13​

-3.0%​

36​

Jágr​

99-00​

96​

42​

38​

07:30​

06:47​

05:42​

90.5%​

76.0%​

1.19​

1.8%​

1​

Kovalchuk​

05-06​

98​

91​

82​

10:06​

09:06​

08:10​

90.1%​

80.9%​

1.11​

-4.8%​

26​

Lemieux​

02-03​

91​

58​

52​

07:10​

06:25​

05:47​

89.7%​

80.7%​

1.11​

-5.0%​

77​

Savard​

06-07​

96​

71​

63​

08:10​

07:14​

05:24​

88.7%​

66.1%​

1.34​

14.7%​

194​

Bure​

00-01​

92​

46​

40​

07:18​

06:20​

04:52​

87.0%​

66.7%​

1.30​

11.5%​

32​

Jágr​

00-01​

121​

75​

65​

07:13​

06:15​

05:43​

86.7%​

79.2%​

1.09​

-6.5%​

43​

Ovechkin​

07-08​

112​

65​

56​

06:44​

05:48​

05:36​

86.2%​

83.2%​

1.04​

-11.5%​

70​

Kariya​

01-02​

57​

43​

37​

08:01​

06:53​

05:27​

86.0%​

68.0%​

1.27​

8.2%​

132​

Robitaille​

00-01​

88​

71​

61​

07:26​

06:23​

05:04​

85.9%​

68.2%​

1.26​

7.7%​

22​

Jágr​

98-99​

127​

64​

55​

07:34​

06:30​

05:50​

85.9%​

77.1%​

1.11​

-4.7%​

4​

Ovechkin​

05-06​

106​

70​

60​

09:45​

08:21​

06:43​

85.7%​

68.9%​

1.24​

6.3%​

157​

Crosby​

09-10​

109​

56​

48​

06:29​

05:33​

05:00​

85.7%​

77.1%​

1.11​

-5.0%​

19​

Fleury​

00-01​

74​

48​

41​

07:05​

06:03​

05:59​

85.4%​

84.5%​

1.01​

-13.6%​

9​

Kovalchuk​

06-07​

76​

67​

57​

08:20​

07:05​

06:31​

85.1%​

78.2%​

1.09​

-7.0%​

96​

Hossa​

06-07​

100​

67​

57​

08:20​

07:05​

05:17​

85.1%​

63.4%​

1.34​

14.7%​

16​

Whitney​

02-03​

76​

71​

60​

08:23​

07:05​

06:07​

84.5%​

73.0%​

1.16​

-1.0%​

53​

Kovalev​

00-01​

95​

76​

64​

07:10​

06:02​

05:33​

84.2%​

77.4%​

1.09​

-7.1%​

128​

Sakic​

06-07​

100​

79​

66​

07:40​

06:24​

05:05​

83.5%​

66.3%​

1.26​

7.7%​

​

Bure​

97-98​

90​

48​

40​

06:59​

05:49​

04:30​

83.3%​

64.4%​

1.29​

10.5%​

151​

Malkin​

07-08​

106​

77​

64​

07:39​

06:21​

05:01​

83.1%​

65.6%​

1.27​

8.3%​

103​

Sakic​

98-99​

96​

65​

54​

08:26​

07:00​

05:13​

83.1%​

61.9%​

1.34​

14.8%​

50​

Straka​

00-01​

95​

76​

63​

07:10​

05:56​

05:33​

82.9%​

77.4%​

1.07​

-8.5%​

​

Kucherov​

22-23​

113​

71​

58​

05:22​

04:23​

04:00​

81.7%​

74.5%​

1.10​

-6.3%​

60​

Smyth​

00-01​

70​

59​

48​

08:04​

06:33​

05:31​

81.4%​

68.4%​

1.19​

1.7%​

62​

Allison​

00-01​

95​

64​

52​

07:47​

06:19​

05:30​

81.3%​

70.7%​

1.15​

-1.7%​

52​

Recchi​

99-00​

91​

69​

56​

06:55​

05:36​

05:33​

81.2%​

80.2%​

1.01​

-13.6%​

​

Kucherov​

18-19​

128​

74​

60​

04:57​

04:00​

03:43​

81.1%​

75.1%​

1.08​

-7.7%​

177​

Yashin​

98-99​

94​

59​

47​

08:16​

06:35​

04:56​

79.7%​

59.7%​

1.33​

14.1%​

14​

Jágr​

05-06​

123​

83​

66​

08:28​

06:43​

06:12​

79.5%​

73.2%​

1.09​

-7.2%​

176​

Jokinen​

05-06​

89​

63​

50​

07:58​

06:19​

04:56​

79.4%​

61.9%​

1.28​

9.5%​

23​

Crosby​

06-07​

120​

92​

73​

09:16​

07:21​

05:50​

79.3%​

62.9%​

1.26​

7.7%​

134​

Richards​

06-07​

70​

69​

54​

07:50​

06:07​

05:04​

78.3%​

64.7%​

1.21​

3.4%​

64​

Bertuzzi​

02-03​

97​

87​

68​

08:24​

06:33​

05:29​

78.2%​

65.3%​

1.20​

2.3%​

125​

Francis​

97-98​

87​

64​

50​

08:14​

06:25​

05:06​

78.1%​

61.9%​

1.26​

7.8%​

102​

Kovalev​

99-00​

66​

54​

42​

07:19​

05:41​

05:13​

77.8%​

71.3%​

1.09​

-6.8%​

57​

Ovechkin​

06-07​

92​

67​

52​

08:21​

06:28​

05:31​

77.6%​

66.1%​

1.17​

0.4%​

​

Kane​

15-16​

106​

57​

44​

04:54​

03:46​

03:07​

77.2%​

63.6%​

1.21​

3.7%​

92​

Fleury​

98-99​

93​

48​

37​

07:05​

05:27​

05:18​

77.1%​

74.8%​

1.03​

-11.9%​

75​

Näslund​

02-03​

104​

87​

67​

08:24​

06:28​

05:25​

77.0%​

64.5%​

1.19​

2.1%​

183​

Iginla​

08-09​

89​

61​

47​

07:17​

05:36​

04:54​

77.0%​

67.3%​

1.15​

-2.1%​

121​

Iginla​

01-02​

96​

55​

42​

07:43​

05:53​

05:08​

76.4%​

66.5%​

1.15​

-1.9%​

​

Thornton​

07-08​

96​

70​

53​

07:28​

05:39​

04:47​

75.7%​

64.1%​

1.18​

1.0%​

17​

Kariya​

05-06​

85​

94​

71​

10:28​

07:54​

06:06​

75.5%​

58.3%​

1.30​

10.8%​

180​

Kovalchuk​

02-03​

67​

64​

48​

07:39​

05:44​

04:55​

75.0%​

64.3%​

1.17​

-0.3%​

​

McDavid​

16-17​

100​

56​

42​

04:58​

03:43​

03:01​

75.0%​

60.7%​

1.23​

5.5%​

35​

Jágr​

06-07​

96​

75​

56​

08:09​

06:05​

05:43​

74.7%​

70.1%​

1.06​

-9.0%​

203​

Fleury​

97-98​

78​

43​

32​

07:24​

05:30​

04:51​

74.4%​

65.5%​

1.14​

-3.0%​

7​

Richards​

05-06​

91​

81​

60​

10:01​

07:25​

06:34​

74.1%​

65.6%​

1.13​

-3.4%​

18​

Hossa​

05-06​

92​

100​

74​

10:17​

07:36​

06:06​

74.0%​

59.3%​

1.25​

6.6%​

30​

Heatley​

05-06​

103​

102​

75​

09:24​

06:54​

05:44​

73.5%​

61.0%​

1.21​

3.0%​

84​

Staal​

06-07​

70​

67​

49​

09:08​

06:40​

05:21​

73.1%​

58.6%​

1.25​

6.7%​

154​

Brind'Amour​

06-07​

82​

63​

46​

09:08​

06:40​

05:00​

73.0%​

54.7%​

1.33​

14.0%​

38​

Alfredsson​

05-06​

103​

99​

72​

09:25​

06:50​

05:41​

72.7%​

60.4%​

1.21​

3.0%​

48​

Marleau​

05-06​

86​

91​

66​

10:09​

07:21​

05:34​

72.5%​

54.8%​

1.32​

13.0%​

201​

Forsberg​

98-99​

97​

69​

50​

07:45​

05:36​

04:51​

72.5%​

62.6%​

1.16​

-1.0%​

82​

Thornton​

05-06​

125​

94​

68​

09:32​

06:53​

05:22​

72.3%​

56.3%​

1.29​

9.8%​

12​

Savard​

05-06​

97​

100​

71​

10:17​

07:18​

06:22​

71.0%​

61.9%​

1.15​

-2.0%​

160​

Heatley​

06-07​

105​

72​

51​

08:14​

05:49​

04:59​

70.8%​

60.5%​

1.17​

0.0%​

152​

Thornton​

06-07​

114​

92​

65​

08:15​

05:49​

05:01​

70.7%​

60.8%​

1.16​

-0.7%​

91​

V.Bure​

99-00​

75​

59​

41​

06:48​

04:43​

05:18​

69.5%​

77.9%​

0.89​

-23.8%​

117​

Cheechoo​

05-06​

93​

91​

63​

10:09​

07:01​

05:08​

69.2%​

50.6%​

1.37​

17.0%​

79​

Iginla​

05-06​

67​

87​

60​

09:30​

06:33​

05:24​

69.0%​

56.8%​

1.21​

3.7%​

178​

C.Lemieux​

98-99​

51​

71​

49​

07:43​

05:19​

04:56​

69.0%​

63.9%​

1.08​

-7.7%​

159​

Morrison​

02-03​

71​

87​

58​

08:24​

05:36​

04:59​

66.7%​

59.3%​

1.12​

-4.0%​

136​

Gomez​

05-06​

84​

78​

52​

09:11​

06:07​

05:04​

66.7%​

55.2%​

1.21​

3.3%​

108​

Recchi​

06-07​

68​

94​

62​

09:16​

06:06​

05:11​

66.0%​

55.9%​

1.18​

0.8%​

73​

Francis​

02-03​

57​

58​

38​

08:49​

05:46​

05:25​

65.5%​

61.4%​

1.07​

-8.9%​

158​

Gionta​

05-06​

89​

78​

51​

09:11​

06:00​

05:00​

65.4%​

54.4%​

1.20​

2.6%​

40​

Crosby​

05-06​

102​

93​

60​

09:50​

06:20​

05:40​

64.5%​

57.6%​

1.12​

-4.3%​

80​

Brind'Amour​

05-06​

70​

93​

59​

10:47​

06:50​

05:24​

63.4%​

50.1%​

1.27​

8.3%​

63​

Staal​

05-06​

100​

95​

60​

10:47​

06:48​

05:30​

63.2%​

51.0%​

1.24​

5.8%​

147​

Straka​

05-06​

76​

83​

52​

08:28​

05:18​

05:02​

62.7%​

59.4%​

1.05​

-9.9%​

153​

Rolston​

05-06​

79​

77​

48​

09:05​

05:39​

05:01​

62.3%​

55.2%​

1.13​

-3.5%​

170​

Satan​

01-02​

73​

50​

31​

07:44​

04:47​

04:58​

62.0%​

64.2%​

0.97​

-17.5%​

156​

Stoll​

05-06​

68​

88​

54​

09:29​

05:49​

05:00​

61.4%​

52.7%​

1.16​

-0.5%​

191​

Sanderson​

02-03​

67​

71​

43​

08:23​

05:04​

04:53​

60.6%​

58.3%​

1.04​

-11.1%​

65​

Recchi​

05-06​

64​

89​

53​

10:15​

06:06​

05:28​

59.6%​

53.3%​

1.12​

-4.6%​

44​

Näslund​

05-06​

79​

96​

54​

10:16​

05:46​

05:36​

56.3%​

54.5%​

1.03​

-11.9%​

47​

Bertuzzi​

05-06​

71​

96​

53​

10:16​

05:40​

05:34​

55.2%​

54.2%​

1.02​

-13.0%​

69​

Doan​

05-06​

66​

96​

53​

10:50​

05:58​

05:27​

55.2%​

50.3%​

1.10​

-6.2%​

173​

O'Neill​

02-03​

61​

58​

32​

08:49​

04:51​

04:57​

55.2%​

56.1%​

0.98​

-16.0%​

164​

Kariya​

06-07​

76​

71​

39​

08:32​

04:41​

04:58​

54.9%​

58.2%​

0.94​

-19.3%​

85​

Sakic​

05-06​

87​

89​

47​

09:31​

05:01​

05:20​

52.8%​

56.0%​

0.94​

-19.5%​

181​

Morrison​

05-06​

56​

96​

44​

10:16​

04:42​

04:55​

45.8%​

47.9%​

0.96​

-18.2%​

Conclusions​

As previously mentioned, the mean of the entire dataset yielded a ratio of 1 to 1.15 1 to 1.16 with a range spanning from 0.89 to 1.37 It should be unsurprising that star players exhibit a higher incidence of being on the ice for powerplay goals relative to their powerplay ice time. This observation aligns with the fact that during the powerplay minutes they don't participate in lower-skilled players are utilized as substitutes, with the result being lower powerplay scoring rates. Additionally, if the teams main powerplay unit fails to score during their time on the ice, they are likely to leave the ice towards the end of the powerplay once possession is lost. This in turn places the new players joining the ice for the remainder of the powerplay in a disadvantageous position for scoring i.e. mop-up duties.

​	Count​	Mean​	Sample size in gms​
>90%​	17​	1.140​	1339​
>80%​	42​	1.161​	3337​
>70%​	75​	1.177​	5933​
>60%​	91​	1.169​	7322​
All	100​	1.156​	7978​

The range of 0.89 to 1.37 does seem quite large but for seasons above 70% the range was narrower at 1.01 to 1.37 furthermore 79% of player-seasons(72 of 91) above the 60% mark came within 10% of the mean in that range(1.17) and all seasons above 70% are within 15% of the mean. As the number of player-seasons increase, most of whom are likely to have progressively lower percentages of team powerplay goals on ice for compared to those listed above, the ratio will approach the league norm(1 to 1). This trend is already becomes evident towards the bottom of the list when the '% of PP goals on ice for' drops below 60%

Accuracy of the results in relation to the mean:

​	+90% on ice for​	+80% on ice for​	+70% on ice for​	+60% on ice for​
Within 5%​	6 of 17 (35%)​	12 of 42 (29%)​	29 of 75 (39%)​	37 of 92 (40%)​
Within 7.5%​	13 of 17 (76%)​	24 of 42 (57%)​	46 of 75 (61%)​	55 of 92 (60%)​
Within 10%​	15 of 17 (88%)​	32 of 42 (76%)​	60 of 75 (80%)​	72 of 92 (78%)​
Within 12.5%​	15 of 17 (88%)​	35 of 42 (83%)​	64 of 75 (85%)​	77 of 92 (84%)​
Within 15%​	100%​	100%​	100%​	87 of 92 (95%)​
Within 17.5%​	100%​	100%​	100%​	91 of 92 (99%)​
Range​	-13.8% to +6.9%​	-13.8% to +14.8%​	-13.8% to +14.8%​	-23.8% to +17.0%​

This method does indeed offers a reasonable estimate of a player's ice time in the pre-Time On Ice (TOI) tracked era. However, it is advisable to use a ratio of 1 to 1.17 (17% less ice time than expected) rather than a 1 to 1 ratio. A potential deviation of around 10% should be considered as likely, but any deviations beyond 15% can be assumed to be extremely unlikely for high powerplay usage players.

Considering the powerplay performance of a team in games with and without a specific player may also help in narrowing down the difference in scoring rates for powerplay goals when the player is on the ice versus when they are off the ice. But further analysis is required before drawing any final conclusions in that regard.

*edit After double checking some figures I noticed a few errors in the dataset, I accidentally used the entire season team totals some players when I should have eliminated the team totals in the games they didn't play in (Kovalchuk '08 Selänne '99 & Malkin '12). Another corrected mistake was Recchi in 06-07, incorrectly had his teams total as 81 for the year when they actually had 94 which removed one of the biggest outliners. The data-set has been adjusted accordingly.

Click to read more...

 

[rubenflamshep]

Seems like an interesting way to get a ballpark. But to your point - it would disguise the PP effectiveness of players. Visualizing the distribution of the ratios would give us a sense of the degree to which that could happen

 

[TheStatican]

Seems like an interesting way to get a ballpark. But to your point - it would disguise the PP effectiveness of players. Visualizing the distribution of the ratios would give us a sense of the degree to which that could happen

Sure, here's a few graphs to help interpret the data.
All player-seasons plotted:

The variance is a little narrower above 90% and then the range slightly widens before tapering off and consolidating towards to a 1:1 ratio when we approach the 50% mark, which makes perfect sense upon reflection.

Bell curve:

The bell curve demonstrates a negatively skewed owing to the inclusion of a few seasons in the low 60% range and below. At the time I was uncertain about the impact of incorporating these lower percentage seasons; my primary goal was simply to encompass as many of the top 200 powerplay ice time(82 game) player-seasons as possible. However, it makes sense that there would be only a limited number of these seasons where a star players logged extensive minutes on the powerplay without witnessing an increase in scoring rates vs when they were off the ice since players understandably receive more powerplay ice time when they enhance a powerplay's effectiveness, it is essentially a self-fulfilling prophecy. Attaining a mark of being on the ice for 90% or more of their team's powerplay goals is a very rare mark and there's a good chance that my dataset includes a majority of those instances since TOI tracking began. In contrast there have been thousands of player-seasons achieving a mark in the 50 to 60% range during the same timeframe. However, only a very small number of those seasons made the list, which is understandable - it's hard to have a top 200 all time PP TOI season when your only on the ice for 60% or even less of your teams powerplay goals. But when you look at the season where most of those lower entries did it in it all makes sense - the year of the powerplay, 2005-06.

Simply by excluding all seasons below 63% (an the one outlier above it, V.Bure '00), the result is an almost perfectly even distribution centered around a median of 1.18

Below .98​	0​
.99 to 1.08​	12​
1.09 to 1.18​	32​
1.19 to 1.28​	28​
1.29 to 1.38​	12​
Above 1.38​	0​

Speaking of V.Bure, his 1999-00 season was the most usual outliner of the entire data set. He was on the ice for 78% of the teams powerplay time (5:18 of 6:48) and played in all 82 games, but was only on the ice for 69.5% of the teams powerplay goals. When you do the math it turns out the teams powerplay was more than 50% better without him on the ice:

​	PPO's​	PPG's​	PP %​
Flames overall​	330​	59​	17.9%​
Bure on the ice​	257.2​	41​	15.9%​
Off the ice​	72.8​	18​	24.7%​

I wonder if this went unnoticed or if it was just brushed off as happenstance.

Next this is a 10 player-season rolling average:

Here we observe that the rolling average remains highly stable before gradually declining towards the expected 1:1 ratio. The one peculiar aspect I didn't quite understand here was the slightly lower average at the uppermost end. But upon closer examination I identified a common factor in the data set - Ovechkin.

Draisaitl​	1.24​	2022-23​
McDavid​	1.13​	2020-21​
McDavid​	1.23​	2022-23​
Malkin​	1.23​	2008 09​
Draisaitl​	1.10​	2020-21​
Ovechkin​	1.01​	2013 14​
Ovechkin​	1.08​	2008 09​
Kariya​	1.25​	1998 99​
Draisaitl​	1.09​	2019-20​
Ovechkin​	1.08​	2009 10​
Kovalchuk​	1.14​	2003 04​
Ovechkin​	1.02​	2019 20​
Weight​	1.11​	2000 01​
Malkin​	1.13​	2011-12​
Jágr​	1.19​	1999 00​
Kovalchuk​	1.03​	2007 08​
Kovalchuk​	1.11​	2005 06​
Lemieux​	1.11​	2002 03​
Savard​	1.34​	2006 07​
Bure​	1.30​	2000 01​
Jágr​	1.09​	2000 01​
Ovechkin​	1.04​	2007 08​
Kariya​	1.27​	2001 02​
Robitaille​	1.26​	2000 01​
Jágr​	1.11​	1998 99​

He alone accounts for five of the six lowest seasons within the top 25 (85% and above), comprising all of his seasons within this range. These seasons also constituted five of the eleven lowest among all seasons above 70% team PP goals on ice for (76 in total), including the lowest overall season in 2019-20. It's difficult to say what exactly occurred here. Either, A) the Capitals powerplay was almost as effective without him as it was with him simply because it featured an abundance of highly skilled players. This makes sense for seasons like '09, '10, and '20, when the Capitals were one of, if not the best offensive team in the league with an abundance of high level talent. Notably he had two earlier seasons with significantly higher ratios(1.24 in '06 and 1.17 in '07) when the Capitals were much weaker overall. Or B) Ovechkin just didn't elevate his team's powerplay as much as other star players did theirs. That's not to say his presence on the powerplay had a negative effect - any value above 1 indicates a positive effect. But in seasons like '08 and '14, when the overall talent level on the Capitals roster was lower it's more challenging to explain why his presence on their powerplay wouldn't have resulted in much higher scoring rates with him out there.

The above may be true but i think there is a better explanation. Upon further reflection I don't think the lower average ratio for players 90% and above is actually due to the influence of just a single outliner player(Ovechkin) but rather that this player was largely following the inevitable trend - the closer one gets to being on the ice for 100% of your teams PP goals the less likely that players ice time will be significantly divergent from it. Ovi also wasn't the only one who had seasons with low ratios, Kovalchuk and Lemieux had seasons where their powerplay time was much less than their powerplay goals on ice for % was:

Rk​

​

Season​

Pts​

Team PPG's​

Player PPG's​

Team PP TOI​

Est. TOI​

Actual TOI​

% team PP goals​

% team PP mins​

Ratio​

Accuracy(in relation to 1.17)​

11​

Lemieux​

00-01​

76​

45​

40​

07:21​

06:32​

06:23​

88.9%​

86.8%​

1.02​

-12.5%​

13​

Lemieux​

01-02​

31​

16​

15​

07:14​

06:46​

06:13​

93.8%​

85.9%​

1.09​

-6.8%​

3​

Lemieux​

05-06​

22​

30​

20​

11:23​

07:35​

06:48​

66.7%​

59.7%​

1.12​

-4.6%​

The mean for all seasons around and above 90% is notably lower at 1 to 1.13 Using this ratio for extremely high powerplay goal on ice for seasons will generate more accurate numbers.

Name​	% on ice for​	% of mins​	Ratio​
Draisaitl​	100.0%​	80.4%​	1.24​
McDavid​	97.9%​	86.6%​	1.13​
McDavid​	97.8%​	79.4%​	1.23​
Malkin​	96.8%​	78.5%​	1.23​
Draisaitl​	95.8%​	87.2%​	1.10​
Ovechkin​	95.5%​	94.7%​	1.01​
Ovechkin​	95.2%​	88.5%​	1.08​
Kariya​	95.2%​	76.1%​	1.25​
Draisaitl​	94.9%​	86.8%​	1.09​
Kovalchuk​	94.0%​	87.9%​	1.07​
Lemieux​	93.8%​	85.9%​	1.09​
Ovechkin​	93.7%​	86.8%​	1.08​
Kovalchuk​	93.2%​	81.7%​	1.14​
Ovechkin​	92.7%​	90.7%​	1.02​
Selänne​	92.1%​	73.7%​	1.25​
Weight​	91.5%​	82.2%​	1.11​
Malkin​	90.6%​	79.8%​	1.13​
Jágr​	90.5%​	76.0%​	1.19​
Kovalchuk​	90.1%​	80.9%​	1.11​
Lemieux​	89.7%​	80.7%​	1.11​
Lemieux​	88.9%​	86.8%​	1.02​
Average​	​	​	1.129​

Around 90 and above - 1.13
Under 90 to about 70% - 1.17
70 to around 60% - ? transition zone
High 50 to low 40% - 1.00

More player seasons in the low 70% to high 50% range would need to be analyzed before making a recommendation on what would be the best figure to use there but it's clearly a transition zone.

 

[TheStatican]

Well I just had a bit of a breakthrough...

I was so preoccupied with figuring out a if there was a pattern with one metric that I didn't notice another staring me right in the face - The more powerplay time a player has, the lower the overall percentage of the teams powerplay time he can possibly play in:

Team powerplay ice time​	Average​	Highest figure​
Under 7 minutes​	80.0%​	94.7% of 5:02 Ovechkin 13-14​
7 to 8 minutes​	72.0% ​	86.8% of 7:21 Lemieux 00-01​
8 to 9 minutes​	68.1%​	82.2% of 8:04 Weight 00-01​
Over 9 minutes​	61.7%​	80.9% of 10:06 Kovalchuk 05-06​

The average noted above is the % of their teams powerplay ice time. This only included players who were on the ice for 70% or more of their teams powerplay goals, those under that mark heavily skew the above 9 minute average downward. Here's some player examples, the trend becomes immediately evident which is how I noticed it

Ovechkin	Team PP TOI	Player PP TOI	% of mins
13-14	5:19​	5:02​	94.7%​
19-20	5:23​	4:53​	90.7%​
09-10	5:49​	5:03​	86.8%​
08-09	6:06​	5:24​	88.5%​
07-08	6:44​	5:36​	83.2%​
06-07	8:21​	5:31​	66.1%​
05-06	9:45​	6:43​	68.9%​

Jagr	Team PP TOI	Player PP TOI	% of mins
00-01	7:13​	5:43​	79.2%​
99-00	7:30​	5:42​	76.0%​
98-99	7:34​	5:50​	77.1%​
06-07	8:09​	5:43​	70.1%​
05-06	8:28​	6:12​	73.2%​

There's some fluctuations here and there but the general trend holds and persists for all players without exception. Here are the benchmark figures, this is essentially the upper based on a wide range of team powerplay minutes:

All time PP TOI Rk​	​	Team PP TOI​	% of mins​
148​	Ovechkin 13-14​	5:19​	94.7%​
189​	Ovechkin 19-20​	5:23​	90.7%​
76​	Ovechkin 08-09​	6:06​	88.5%​
34​	Kovalchuk 07-08​	6:30​	87.9%​
19​	Fleury 00-01​	7:05​	84.5%​
11​	Lemieux 00-01​	7:21​	86.8%​
6​	Weight 00-01​	8:04​	82.2%​
2​	Kovalchuk 03-04​	8:23​	81.7%​
1​	Kovalchuk 05-06​	10:06​	80.9%​

I've taken to calling this upper limit the "LOK limit". No, that doesn't stand for "laugh out loud, ok(sure)" It's an acronym for the 3 players who set the maximum at various time intervals: Lemieux, Ovechkin and Kovalchuk. Guess I could also have called it the 'OLK limit' or 'LKO' or 'KOL', not sure which sounds more catchy. Fleury's figure above is not a benchmark, but it is the highest non-LOK figure in the 7 minute range, not in the chart.

The figures in Blue are essentially exceptionally high marks which punch above the normal maximum and form the very upper limit of what's possible. To be sure I cannot completely discount the possibility that players in previous seasons surpassed even this threshold. But this is nothing like raw point totals, it's more akin to approaching the speed of light(that 100% mark) - the closer you get, the more difficult it becomes to go faster(higher). Any higher figures could only have surpassed the LOK limit by a very small amount. Keep in mind this isn't based off of a small sample size. Even though my dataset only contains 100 players, those 100 players had the highest PP TOI in tracked history(for forwards). Nobody outside of these players could have possibly had higher percentages. Which means that these are the highest rates seen over a quarter century which also amounts to nearly a quarter of the NHL's entire history.

Now the real question is how can any of this be put to practical use? I'll show an example of that shortly.

 

[rubenflamshep]

The more powerplay time a player has, the lower the overall percentage of the teams powerplay time he can possibly play in:

Hmm, I see the trend but I don't full understand why this is. Is this due to players just fatiguing and not being able to play as much of each incremental minutes of additional pp time?

 

[TheStatican]

Hmm, I see the trend but I don't full understand why this is. Is this due to players just fatiguing and not being able to play as much of each incremental minutes of additional pp time?

That's a good question. I can't say for sure but my assumption is that the more powerplay opportunities a team has the greater the probability that a powerplay will occur during or just after star player was out on the ice for his regular shift. If you just finished or are finishing up a long shift your probably not going to be able to be out there for another 2 minutes straight on the powerplay, the duration that most powerplays last for. In high powerplay seasons(5-6 per game) those instances would happen more frequently than in low powerplay seasons(3-4 per game). Even for a player like Lemieux there's a limit to how much ice time he can possibly have, but his limit is a little bit higher than players like Fleury or Weight.

 

[TheStatican]

Let's pick out a random high usage powerplay player - say Steven Stamkos from last season. Now say we don't know how much time his team had on the powerplay per game. We do know the Lightning had a PP of 25.4%, You could pick another team that had a similar value but it's better to pick out several in the same range and then use the average of them, but I'll do an example using both ways.

One team: The 08-09 Red Wings who had a PP rate of 25.5% and they averaged 6:43 minutes of powerplay time per gm. They did that with 353 PPO's(4.3/gm) while the Lightning only had 278 in Stamkos 81gms(3.4/gm). If the Wings had the same number per game as the Lightning their PP mins per game would have been 5:16

Now taking the longer, multi-team approach:

Team​	PP %​	games​	PP mins/gm​	PPO's​	PP length in secs​
22-23 Kings​	25.63%​	60​	05:10​	199 (3.3/gm)​	93.5​
20-21 Hurricanes​	25.61%​	56​	04:33​	164 (2.9gm)​	93.2​
22-23 Leafs​	25.41%​	59​	05:03​	181 (3.1/gm)​	98.8​
22-23 Sabers​	25.39%​	57​	05:33​	193 (3.4/gm)​	98.3​
14-15 Capitals​	25.32%​	82​	04:32​	237 (2.9/gm)​	94.1​
09-10 Capitals​	25.24%​	82​	05:52​	313 (3.8/gm)​	92.2​
21-22 Rangers​	25.23%​	82​	04:10​	218 (2.7/gm)​	94.0​
Average​	​	​	​	​	94.9​

We just need to figure out what the average length of each teams powerplay was. This can be done by dividing their number of PPO's per game by their PP mins per game. The average of the 7 teams above was 94.9 secs. As mentioned the Lightning had 278 PPO's, multiply that by the average length (94.9 second) results in a total of 439.7 minutes and divided by Stamkos's 81 games gives an estimate of 5:26 of powerplay time per game.

Stamkos was out on the ice for 60 of his teams 71 PP goals, that = 84.5% of them. Divided by the average ratio from the data set(1.17) gives us a estimate of him being out on the ice for 72.2% of his teams powerplay ice time.
72.2% of 5:16 is 3:48 &
72.2% of 5:26 is 3:55

So how do these estimates compared to what really happened?
The Lightning actually had 5:22 minutes of powerplay time per game.
- The single team example was within 2% (lower),
- But the mutli-team average was just 1% off (higher).

What was Stamko's PP TOI last year? 3:58 per game
- The single team estimate was off by 4% in total
- As for the mutli-team estimate, it was only off by 1% from his actual PP TOI


Now lets try to use this method to get an estimate of a players ice time that we don't know. Previously I tried to estimate Lemieux's powerplay time in 95-96:

Lemieux's 1995-96 powerplay stats and figures

This thread is an addendum to this post; https://forums.hfboards.com/threads/was-lemieux-still-in-his-peak-in-1995-96.2906900/post-187317395 Using his TOI stats from 00-01 will help to figure out Lemieux's 95-96 powerplay TOI. Here's an analysis of the Penguins powerplay for each season. I...

forums.hfboards.com

I originally came up with an estimate of 7:45 minutes per game which was based off of the fact that he was on the ice for all 102 of the teams powerplay goals(in his games) and by comparing that season to his own 00-01 season and Kovalchuk's PP TOI record season of 05-06. In hindsight, that's not really a lot of a data to make comparisons against and I would have doubted my conclusions had I waited 6 months to post that. Because at that time there was another player who had just finished a season where they were on the ice for 100% of his teams powerplay goals - Leon Draisaitl (87 of 87 in his 80gm). And despite being on the ice for every single one of his teams PPG's, Leon was only on the ice for 80.4% of his teams powerplay ice time, no where near 100% Just going by that I realized my educated guess was clearly off the mark.

Step 1 - what was the teams powerplay ice time per game? Thankfully in this example we don't need to make any estimations because I went through the game logs(yup, all of them) and counted up the Penguins totals for 95-96, details for which are in the thread above. The team had a total of 33806 seconds of powerplay time in Lemieux's games. This comes out to 8:03 per game. Yet again, another shame on the NHL's part for not providing teams powerplay(and shorthanded) ice times in seasons before 97-98 considering the information for those numbers is right there in their game logs!!

Step 2 - now we can divide that number by the mean ratio. Originally I recommended using a figure of 1.17 but for players who were on the ice for an extremely high percentage of their teams powerplay goals a better ratio to use would be 1.13 which gives us a figure of 7:07 per game for Lemieux. And viola we're done! Not quite, while this gives us a good approximation, there is still some variance in what his actual PP TOI could be.

In relation to the mean of 1.13 for +90% players:

​	+90% on ice for​
Within 5%​	11 of 17 (65%)​
Within 7.5%​	13 of 17 (76%)​
Within 10%​	15 of 17 (88%)​
Within 12.5%​	100%​
Within 15%​	100%​
Within 17.5%​	100%​
Range​	-10.8% to +10.6%​

The highest and lowest figures overall gives a possible known range of 6:24 to 7:51 minutes(7:07 +/-44s), which is admittedly a fairly large range, though there is about a two-thirds probability of him falling within a range half as large, between 6:45 and 7:26(7:05 +/-21s). Thankfully there is now one more thing we can do that could possibly narrow the range down - test to see if the numbers exceed the LOK limit.

My original estimate of 7:45 minutes of powerplay TOI would have meant he was on the ice for 96.6%! of the Penguins powerplay time. Clearly this is simply not possible as it far exceeds all known limits by a substantial margin. The Pens had just over 8 minutes of PP ice time per game and the closest maximum benchmark comparable to that is Doug Weight's 00-01 season. Weight's Oilers had 8:04 minutes of powerplay ice time per game (in his games) and he was out there 82.2% of that. As noted in the chart above this is the highest percentage of powerplay ice time any player has ever achieved had when their team averages 8 or more minutes of powerplay time per game. The LOK (Lemieux, Ovechkin and Kovalchuk) limit does projects a higher theoretical maximum for Lemieux, approaching a figure of 85% at 8 minutes. After taking that into consideration this gives an absolute maximum of about 6:50 minutes of PP TOI and that number narrows things down considerably. Now with that I can say with around a 90-95% certainty that Lemieux's powerplay ice time was between 6:24 and 6:50 (6:37 +/-13s) and a little over a 50% probability that it was between an incredibly narrow range of 6:45 and 6:50 (6:48+/-2.5 seconds!)

As we see the LOK limit can be extremely useful in narrowing the range of possibilities, the unfortunate down side is that it's use is limited to just very high-usage powerplay players since it only defines the maximum/upper limit of possibilities. This is more or less what I was personally seeking, a means to provide a highly accurate guestimate for the very high usage powerplay players. But I'll still continue to mine through and analysis the data to see if there's any more patterns or trends that may be helpful for increasing the accuracy for not just the guys at the very top but for all players.

 

[overpass]

Very thorough and insightful work!

I'd like to take the next step and ask you what you think the best way to determine power play scoring opportunity for players. I think that's the implicit assumption behind estimating ice time, right? Ice time = opportunity. That assumption lies behind stats like power play points/60.

However, you've demonstrated in this post that individual power play scoring does not scale linearly with individual power play ice time. And you've suggested some reasons, such as second unit power play players being less skilled, second unit players being used for "mop-up duties" (we could be more specific and say they often start outside the offensive zone, changing on the fly and without the benefit of an offensive zone faceoff).

To pick a specific player, what stat more accurately describes Leon Draisaitl's power play scoring opportunity? Is it saying that he played 80% of his team's power play ice time? Or saying that he was on the ice for 100% of his team's power play goals scored? In Draisaitl's specific case, consider that 94.5% of his power play minutes (299:27) were played with both Connor McDavid and Ryan Nugent-Hopkins. And also consider that the Oilers played 72 power play minutes without any of McDavid, Draisaitl, or Nugent-Hopkins on the ice, and scored 0 goals in those minutes.

Certainly Draisaitl himself played a significant part in the Oilers' high rate of power play goals for while he was on the ice. We can also be pretty sure that McDavid's presence on the ice with Draisaitl was a major factor in the on-ice numbers for Draisaitl, and Nugent-Hopkins was perhaps a lesser factor. And then there is the favourable starting point of the shift for Draisaitl and the other first unit players.

I would say that 80.4% is not necessarily any more accurate than 100% as a description of Draisaitl's power play opportunity. If I had to pick a number, I would go somewhere between the two. Maybe I'd estimate that Draisaitl was provided with 90% of the power play scoring opportunity that the Oilers had to allocate, separate from his superb individual performance. I could even go higher in Draisaitl's specific case, where his power play time was in lockstep with another all-time great scorer. That's a major conclusion I would take from your results, that power play scoring opportunity is something different from power play time on ice, and cannot be fully described by TOI stats.

 

[rubenflamshep]

That's a good question. I can't say for sure but my assumption is that the more powerplay opportunities a team has the greater the probability that a powerplay will occur during or just after star player was out on the ice for his regular shift. If you just finished or are finishing up a long shift your probably not going to be able to be out there for another 2 minutes straight on the powerplay, the duration that most powerplays last for. In high powerplay seasons(5-6 per game) those instances would happen more frequently than in low powerplay seasons(3-4 per game). Even for a player like Lemieux there's a limit to how much ice time he can possibly have, but his limit is a little bit higher than players like Fleury or Weight.

This makes sense to me! I think the fatigue hypothesis resonates for sure. +1 to the overpass post

 

[JetsFan815]

Very interesting. Similar methodology has also been used in to estimate the overall icetime of players in the CHL and the AHL(which don't publish ToI) by prospect people, except there you have additional datapoints of having +/- information in addition to goal information.

 

[TheStatican]

Very thorough and insightful work!

I'd like to take the next step and ask you what you think the best way to determine power play scoring opportunity for players. I think that's the implicit assumption behind estimating ice time, right? Ice time = opportunity. That assumption lies behind stats like power play points/60.

However, you've demonstrated in this post that individual power play scoring does not scale linearly with individual power play ice time. And you've suggested some reasons, such as second unit power play players being less skilled, second unit players being used for "mop-up duties" (we could be more specific and say they often start outside the offensive zone, changing on the fly and without the benefit of an offensive zone faceoff).

To pick a specific player, what stat more accurately describes Leon Draisaitl's power play scoring opportunity? Is it saying that he played 80% of his team's power play ice time? Or saying that he was on the ice for 100% of his team's power play goals scored? In Draisaitl's specific case, consider that 94.5% of his power play minutes (299:27) were played with both Connor McDavid and Ryan Nugent-Hopkins. And also consider that the Oilers played 72 power play minutes without any of McDavid, Draisaitl, or Nugent-Hopkins on the ice, and scored 0 goals in those minutes.

Certainly Draisaitl himself played a significant part in the Oilers' high rate of power play goals for while he was on the ice. We can also be pretty sure that McDavid's presence on the ice with Draisaitl was a major factor in the on-ice numbers for Draisaitl, and Nugent-Hopkins was perhaps a lesser factor. And then there is the favourable starting point of the shift for Draisaitl and the other first unit players.

I would say that 80.4% is not necessarily any more accurate than 100% as a description of Draisaitl's power play opportunity. If I had to pick a number, I would go somewhere between the two. Maybe I'd estimate that Draisaitl was provided with 90% of the power play scoring opportunity that the Oilers had to allocate, separate from his superb individual performance. I could even go higher in Draisaitl's specific case, where his power play time was in lockstep with another all-time great scorer. That's a major conclusion I would take from your results, that power play scoring opportunity is something different from power play time on ice, and cannot be fully described by TOI stats.

Thanks! I do see what your saying there, it does make sense that a composite of the two numbers would give a decent estimate as to what percentage of the teams powerplay scoring opportunities a player was involved in.

From the original dataset here's the players who registered a value an average of over 80% when combining minutes and goals on ice for:

Name​	Season​	Pts​	Team PP Goals​	​ Avg of team goals & mins​
Ovechkin​	13-14​	79​	66​	95.1%​
McDavid​	20-21​	105​	48​	92.2%​
Ovechkin​	08-09​	110​	84​	91.9%​
Ovechkin​	19-20​	67​	41​	91.7%​
Draisaitl​	20-21​	84​	48​	91.5%​
Lemieux (est.)​	95-96​	161​	102​	91.3%​
Kovalchuk​	07-08​	87​	50​	91.0%​
Draisaitl​	19-20​	110​	59​	90.9%​
Ovechkin​	09-10​	109​	63​	90.24%​
Draisaitl​	22-23​	128​	87​	90.20%​
Lemieux (in 24gms)​	01-02​	31​	16​	89.8%​
McDavid​	22-23​	153​	89​	88.7%​
Crosby (in 22gms)​	11-12​	37​	13​	88.4%​
Lemieux​	00-01​	76​	45​	88.0%​
Malkin​	08-09​	113​	62​	87.7%​
Kovalchuk​	03-04​	87​	59​	87.5%​
Weight​	00-01​	90​	59​	86.9%​
Kariya​	98-99​	101​	83​	85.6%​
Kovalchuk​	05-06​	98​	91​	85.5%​
Malkin​	11-12​	109​	53​	85.19%​
Lemieux​	02-03​	91​	58​	85.18%​
Fleury​	00-01​	74​	48​	84.9%​
Ovechkin​	07-08​	112​	65​	84.7%​
Crosby​	10 11​	66​	29​	84.0%​
Jágr​	99-00​	96​	42​	83.2%​
Jágr​	00-01​	121​	75​	82.94%​
Selänne​	98-99​	107​	76​	82.89%​
Kovalchuk​	06-07​	76​	67​	81.6%​
Jágr​	98-99​	127​	64​	81.5%​
Crosby​	09-10​	109​	56​	81.4%​
Kovalev​	00-01​	95​	76​	80.8%​
Recchi​	99-00​	91​	69​	80.7%​
Straka​	00-01​	95​	76​	80.2%​

All the usual suspects that one would expect are there - Ovechkin, Kovalchuk, McDavid, Draisaitl, Lemieux, Crosby, Malkin & Jágr. And especially so right at the top, those players take up the top 16 player seasons and 26 of the 33 overall. That certainly gives the impression such a metric is on the right track. The one thing that gives me a little bit of a pause though is Valeri Bure registering one of the highest values of the remining player-seasons

Savard​	06-07​	77%​
Bure​	00-01​	77%​
Kariya​	01-02​	77%​
Robitaille​	00-01​	77%​
Ovechkin​	05-06​	77%​
Hossa​	06-07​	74%​
Whitney​	02-03​	79%​
Sakic​	06-07​	75%​
Bure​	97-98​	74%​
Malkin​	07-08​	74%​
Sakic​	98-99​	72%​
Kucherov​	22-23​	78%​
Smyth​	00-01​	75%​
Allison​	00-01​	76%​
Kucherov​	18-19​	78%​
Yashin​	98-99​	70%​
Jágr​	05-06​	76%​
Jokinen​	05-06​	71%​
Crosby​	06-07​	71%​
Richards​	06-07​	71%​
Bertuzzi​	02-03​	72%​
Francis​	97-98​	70%​
Kovalev​	99-00​	75%​
Ovechkin​	06-07​	72%​
Kane​	15-16​	70%​
Fleury​	98-99​	76%​
Näslund​	02-03​	71%​
Iginla​	08-09​	72%​
Iginla​	01-02​	71%​
Thornton​	07-08​	70%​
Kariya​	05-06​	67%​
Kovalchuk​	02-03​	70%​
McDavid​	16-17​	68%​
Jágr​	06-07​	72%​
Fleury​	97-98​	70%​
Richards​	05-06​	70%​
Hossa​	05-06​	67%​
Heatley​	05-06​	67%​
Staal​	06-07​	66%​
Brind'Amour​	06-07​	64%​
Alfredsson​	05-06​	67%​
Marleau​	05-06​	64%​
Forsberg​	98-99​	68%​
Thornton​	05-06​	64%​
Savard​	05-06​	66%​
Heatley​	06-07​	66%​
Thornton​	06-07​	66%​
V.Bure​	99-00​	74%​
Cheechoo​	05-06​	60%​
Iginla​	05-06​	63%​
C.Lemieux​	98-99​	66%​
Morrison​	02-03​	63%​
Gomez​	05-06​	61%​
Recchi​	06-07​	61%​
Francis​	02-03​	63%​
Gionta​	05-06​	60%​
Crosby​	05-06​	61%​
Brind'Amour​	05-06​	57%​
Staal​	05-06​	57%​
Straka​	05-06​	61%​
Rolston​	05-06​	59%​
Satan​	01-02​	63%​
Stoll​	05-06​	57%​
Sanderson​	02-03​	59%​
Recchi​	05-06​	56%​
Näslund​	05-06​	55%​
Bertuzzi​	05-06​	55%​
Doan​	05-06​	53%​
O'Neill​	02-03​	56%​
Kariya​	06-07​	57%​
Sakic​	05-06​	54%​
Morrison​	05-06​	47%​

As his 99-00 season was by far the worst outliner by an other measure,
Number of PP goals on the ice for more or less than expected based on their PP ice time:

Name​	Season​	Gm​	Pts​	Team​	Player​	Team​	Actual TOI​	PP +/-​	Proj to 82gm​
Draisaitl​	22-23​	80​	128​	87​	87​	04:56​	03:58​	17.05​	17.5​
Cheechoo​	05-06​	82​	93​	91​	63​	10:09​	05:08​	16.98​	​
Kariya​	05-06​	82​	85​	94​	71​	10:28​	06:06​	16.2​	​
Marleau​	05-06​	82​	86​	91​	66​	10:09​	05:34​	16.1​	​
Savard​	06-07​	82​	96​	71​	63​	08:10​	05:24​	16.1​	​
McDavid​	22-23​	82​	153​	89​	87​	04:56​	03:56​	16.0​	​
Kariya​	98-99​	82​	101​	83​	79​	07:32​	05:44​	15.8​	​
Lemieux(est.)​	95-96​	70​	161​	102​	102​	08:01​	06:47​	15.7​	18.4​
Crosby​	06-07​	79​	120​	92​	73​	09:16​	05:50​	15.1​	15.7​
Thornton​	05-06​	81​	125​	94​	68​	09:32​	05:22​	15.1​	15.3​
Hossa​	05-06​	80​	92​	100​	74​	10:17​	06:06​	14.7​	15.0​
Hossa​	06-07​	82​	100​	67​	57​	08:20​	05:17​	14.5​	​
Selänne​	98-99​	75​	107​	76​	70​	07:32​	05:33​	14.0​	15.3​
Sakic​	98-99​	73​	96​	65​	54​	08:26​	05:13​	13.8​	15.5​
Sakic​	06-07​	82​	100​	79​	66​	07:40​	05:05​	13.6​	​
Malkin​	07-08​	82​	106​	77​	64​	07:39​	05:01​	13.5​	​
Heatley​	05-06​	82​	103​	102​	75​	09:24​	05:44​	12.8​	​
Robitaille​	00-01​	82​	88​	71​	61​	07:26​	05:04​	12.6​	​
Brind'Amour​	05-06​	78​	70​	93​	59​	10:47​	05:24​	12.4​	13.1​
Alfredsson​	05-06​	77​	103​	99​	72​	09:25​	05:41​	12.2​	13.0​
Ovechkin​	05-06​	81​	106​	70​	60​	09:45​	06:43​	11.8​	​
Yashin​	98-99​	82​	94​	59​	47​	08:16​	04:56​	11.8​	​
Staal​	05-06​	82​	100​	95​	60​	10:47​	05:30​	11.5​	​
Brind'Amour​	06-07​	78​	82​	63​	46​	09:08​	05:00​	11.5​	12.1​
Malkin​	08-09​	82​	113​	62​	60​	07:04​	05:33​	11.3​	​
Bertuzzi​	02-03​	82​	97​	87​	68​	08:24​	05:29​	11.2​	​
Jokinen​	05-06​	82​	89​	63​	50​	07:58​	04:56​	11.0​	​
Näslund​	02-03​	82​	104​	87​	67​	08:24​	05:25​	10.9​	​
Iginla​	05-06​	82​	67​	87​	60​	09:30​	05:24​	10.5​	​
Francis​	97-98​	81​	87​	64​	50​	08:14​	05:06​	10.4​	​
Staal​	06-07​	82​	70​	67​	49​	09:08​	05:21​	9.8​	​
Recchi​	06-07​	82​	68​	94​	62​	09:16​	05:11​	9.4​	​
Richards​	06-07​	82​	70​	69​	54​	07:50​	05:04​	9.4​	​
Bure​	00-01​	82​	92​	46​	40​	07:18​	04:52​	9.3​	​
Savard​	05-06​	82​	97​	100​	71​	10:17​	06:22​	9.1​	​
Thornton​	06-07​	82​	114​	92​	65​	08:15​	05:01​	9.1​	​
Bure​	97-98​	82​	90​	48​	40​	06:59​	04:30​	9.1​	​
Gomez​	05-06​	82​	84​	78​	52​	09:11​	05:04​	9.0​	​
Gionta​	05-06​	82​	89​	78​	51​	09:11​	05:00​	8.5​	​
Kovalchuk​	05-06​	78​	98​	91​	82​	10:06​	08:10​	8.4​	8.9​
Whitney​	02-03​	81​	76​	71​	60​	08:23​	06:07​	8.2​	​
Thornton​	07-08​	82​	96​	70​	53​	07:28​	04:47​	8.2​	​
McDavid​	16-17​	82​	100​	56​	42​	04:58​	03:01​	8.0​	​
Kariya​	01-02​	82​	57​	43​	37​	08:01​	05:27​	7.8​	​
Kane​	15-16​	82​	106​	57​	44​	04:54​	03:07​	7.7​	​
Ovechkin​	06-07​	82​	92​	67​	52​	08:21​	05:31​	7.7​	​
Smyth​	00-01​	82​	70​	59​	48​	08:04​	05:31​	7.7​	​
Stoll​	05-06​	82​	68​	88​	54​	09:29​	05:00​	7.6​	​
Heatley​	06-07​	82​	105​	72​	51​	08:14​	04:59​	7.4​	​
Richards​	05-06​	82​	91​	81​	60​	10:01​	06:34​	6.9​	​
Kovalchuk​	02-03​	81​	67​	64​	48​	07:39​	04:55​	6.9​	​
Forsberg​	98-99​	78​	97​	69​	50​	07:45​	04:51​	6.8​	7.2​
Kovalchuk​	03-04​	81​	87​	59​	55​	08:23​	06:51​	6.8​	​
Allison​	00-01​	82​	95​	64​	52​	07:47​	05:30​	6.8​	​
Crosby​	05-06​	81​	102​	93​	60​	09:50​	05:40​	6.4​	​
Morrison​	02-03​	82​	71​	87​	58​	08:24​	04:59​	6.4​	​
Jágr​	99-00​	63​	96​	42​	38​	07:30​	05:42​	6.1​	7.9​
Iginla​	08-09​	82​	89​	61​	47​	07:17​	04:54​	6.0​	​
Malkin​	11-12​	75​	109​	53​	48​	05:27​	04:21​	5.7​	6.2​
Jágr​	98-99​	81​	127​	64​	55​	07:34​	05:50​	5.7​	​
Ovechkin​	08-09​	79​	110​	84​	80​	06:06​	05:24​	5.6​	​
Jágr​	00-01​	81​	121​	75​	65​	07:13​	05:43​	5.6​	​
Recchi​	05-06​	83​	64​	89​	53​	10:15​	05:28​	5.5​	​
Weight​	00-01​	82​	90​	59​	54​	08:04​	06:38​	5.5​	​
Rolston​	05-06​	82​	79​	77​	48​	09:05​	05:01​	5.5​	​
McDavid​	20-21​	56​	105​	48​	47​	04:50​	04:11​	5.5​	​
Iginla​	01-02​	82​	96​	55​	42​	07:43​	05:08​	5.4​	​
Jágr​	05-06​	82​	123​	83​	66​	08:28​	06:12​	5.2​	​
Lemieux​	02-03​	67​	91​	58​	52​	07:10​	05:47​	5.2​	6.4​
Kovalev​	00-01​	79​	95​	76​	64​	07:10​	05:33​	5.1​	​
Kucherov​	22-23​	82​	113​	71​	58​	05:22​	04:00​	5.1​	​
Crosby​	09-10​	81​	109​	56​	48​	06:29​	05:00​	4.8​	​
Draisaitl​	19-20​	71​	110​	59​	56​	04:25​	03:50​	4.8​	​
Doan​	05-06​	82​	66​	96​	53​	10:50​	05:27​	4.7​	​
Kovalchuk​	06-07​	82​	76​	67​	57​	08:20​	06:31​	4.6​	​
Kucherov​	18-19​	82​	128​	74​	60​	04:57​	03:43​	4.4​	​
Ovechkin​	09-10​	72​	109​	63​	59​	05:49​	05:03​	4.3​	4.9​
Straka​	00-01​	82​	95​	76​	63​	07:10​	05:33​	4.1​	​
Draisaitl​	20-21​	56​	84​	48​	46​	04:50​	04:13​	4.1​	​
Fleury​	97-98​	82​	78​	43​	32​	07:24​	04:51​	3.8​	​
C.Lemieux​	98-99​	82​	51​	71​	49​	07:43​	04:56​	3.6​	​
Kovalev​	99-00​	82​	66​	54​	42​	07:19​	05:13​	3.5​	​
Jágr​	06-07​	82​	96​	75​	56​	08:09​	05:43​	3.4​	​
Crosby​	10-11​	41​	66​	29​	26​	06:55​	05:25​	3.3​	6.6​
Kovalchuk​	07-08​	79​	87​	50​	47​	06:30​	05:43​	3.0​	​
Crosby​	11-12​	22​	37​	13​	13​	05:03​	03:53​	3.0​	11.2​
Straka​	05-06​	82​	76​	83​	52​	08:28​	05:02​	2.7​	​
Francis​	02-03​	82​	57​	58​	38​	08:49​	05:25​	2.4​	​
Lemieux​	05-06​	26​	22​	30​	20​	11:23​	06:48​	2.1​	6.6​
Ovechkin​	07-08​	82​	112​	65​	56​	06:44​	05:36​	1.9​	​
Näslund​	05-06​	81​	79​	96​	54​	10:16​	05:36​	1.6​	​
Sanderson​	02-03​	82​	67​	71​	43​	08:23​	04:53​	1.6​	​
Lemieux​	01-02​	24​	31​	16​	15​	07:14​	06:13​	1.2​	​
Kariya​	97-98​	22​	31​	11​	8​	07:49​	04:53​	1.1​	​
Fleury​	98-99​	75​	93​	48​	37​	07:05​	05:18​	1.1​	​
Bertuzzi​	05-06​	82​	71​	96​	53​	10:16​	05:34​	0.9​	​
Lemieux​	00-01​	43​	76​	45​	40​	07:21​	06:24​	0.8​	​
Ovechkin​	19-20​	68​	67​	41​	38​	05:23​	04:53​	0.8​	​
Recchi​	99-00​	82​	91​	69​	56​	06:55​	05:33​	0.6​	​
Ovechkin​	13-14​	78​	79​	66​	63​	05:19​	05:02​	0.5​	​
Fleury​	00-01​	62​	74​	48​	41​	07:05​	05:59​	0.5​	​
Crosby​	12-13​	34​	56​	29​	23​	05:34​	04:22​	0.3​	​
O'Neill​	02-03​	82​	61​	58​	32​	08:49​	04:57​	-0.6​	​
Satan​	01-02​	82​	73​	50​	31​	07:44​	04:58​	-1.1​	​
Morrison​	05-06​	82​	56​	96​	44​	10:16​	04:55​	-2.0​	​
Kariya​	06-07​	82​	76​	71​	39​	08:32​	04:58​	-2.3​	​
Sakic​	05-06​	82​	87​	89​	47​	09:31​	05:20​	-2.9​	​
V.Bure​	99-00​	82​	75​	59​	41​	06:48​	05:18​	-5.0​	​

For anyone who might not understand what the point of this number is, all this means is for example; that based on Draisaitl's ice time he should have been on the ice for 70 of the Oilers 87 PP goals but instead he was out on the ice for 87, 17 more than expected. Bure conversely should have been on the ice for 46 of his teams 59 powerplay goals but was only on the ice for 41. So did Bure really generate 74% of his teams powerplay scoring opportunities when they very notably underperformed while he was out on the ice? On the other hand it could be a case of season-long bad puck luck or a low shooting percentage that never corrected and that would explain why he was still getting so many minutes on the powerplay - he was generating chances but they just weren't going in.

Very interesting. Similar methodology has also been used in to estimate the overall icetime of players in the CHL and the AHL(which don't publish ToI) by prospect people, except there you have additional datapoints of having +/- information in addition to goal information.

Interesting, wasn't aware they did that. Is this still the case today? You'd think they'd be tracking the TOI figures by now