It seems obvious that the more a team controls possession (measured by shots) at even-strength, the more they should be on the power play and less they should be killing penalties, right?
Let's look at teams above 54% and below 46% in CF% the past three seasons, and compare that to the opportunities they received on the PP (NHL rank) and the times they were shorthanded (NHL rank: 1st is least shorthanded, 30th is most shorthanded).
CF% above 54:
2013-14
56.9% LAK 5th PP, 28th PK (-12 for penalty difference)
55.5% CHI 21st PP, 6th PK (+9)
54.4% NJD 28th PP, 13th PK (-23)
2013-13
56.3% LAK 11th PP, 15th PK (+5)
55.9% NJD 3rd PP, 23rd PK (+7)
54.4% BOS 30th PP, 18th PK (-41, in 48 games-- yikes!)
54.1% CHI 23rd PP, 4th PK (+9)
2011-12
54.9% DET 4th PP, 17th PK (+24)
54.8% LAK 6th PP, 25th (-4)
54.2% PIT 7th PP, 15th PK (+19)
So average penalty differential for these teams is -0.7 per season.
If we chuck the data of the top team and bottom team, then it becomes a +1.25. (So even the most rosy view of this data suggests that dominating possession is worth 2:30 of powerplay over the course of 82 games.... Remember in the last post that Carkner costs Isles somewhere around 25 minutes of PP/PK time, if he plays around 80 games at 10 minutes per game.)
So the penalty differential of the top CF% teams is just about zero over the course of a season.
Out of the 10 most dominant seasons for CF%, only two teams finished with an impressive combination of high PP chances while keeping chances against down. And even then, this "special-teams score" adds up to 21 for DET 2011-12 and 22 for PIT 2011-12.... The average score is 31, so NJD 2013-14 and BOS 2013-13 are about the same distance from the average at 41 and 48, respectively.... So it appears that having a great CF% doesn't mean a team has much of a better chance at drawing more penalties than taking, according to this data. (I added in the positive/negative numbers above for those four teams, for clarity's sake.)
Why is this the case? Do teams with a high CF% not attack the same way that other teams do? Or do officials "even-out" the calls?
CF% below 46:
2013-14
42.9% TOR 25th PP, 15th PK (-15)
43.0% BUF 23rd PP, 20th PK (-18)
44.3% EDM 16th PP, 21st PK (-9)
2013-13
44.1% TOR 10th PP, 13th PK (+9)
44.5% EDM 9th PP, 26th PK (-6)
45.1% BUF 16th PP, 27th PK (-15)
2011-12
44.3% MIN 21st PP, 22nd (-27)
So the average for these 7 teams is -11.6, or about 23 minutes more of PK time over the course of a season than PP time.
Teams at the bottom in CF% have struggled with penalties taken/drawn. Of these 7 seasons, only TOR in 2013-13 was in the positive, with a score of 26, compared to the 31 average. And five of the seven teams were down around 40.
So perhaps officials tend to "even-out" calls against the best teams (when they are playing mediocre teams), but they tend not to give the bottom teams in CF% the benefit of the doubt? (We may have seen some of that, even without rose-colored glasses, when Isles were among the worst teams in the NHL.)
Any theories? Are the sample sizes too small?