past penalties affect future calls...very compelling

[DetDannyWilliams]

Fraser was conscious of any widening gap between the number of penalties called on each team, even if the calls infractions were warranted. The stats suggest that other referees are as well.

This is pretty compelling evidence that referees’ past calls influence their next calls. The chart includes every NHL penalty call between the start of the 2005-06 season and the end of the 2014-15 regular season,¹ and it looks at all infractions in a game and how the difference in penalties called on each team affects the next call.² As an example, if the home team has been called for three penalties and the away team five, the away team would be “owed” two calls. The tendency for prior violations to affect NHL penalties is enormous; when home teams are owed penalties, they are called for penalties at much lower rates, and the inverse holds for away teams.

http://fivethirtyeight.com/features/...the-other-guy/

[ParanoidAndroid]

Could also simply be a function of time and probabilities.

[Meathead]

i blame obama

[gebobs]

Could also simply be a function of time and probabilities.

Specifically, regression to the mean.

But the fact that it's skewed toward the home team is interesting. Referees are fallible and, like all humans, prone to bias. I have no doubt that the bias can come in the form of crowd influence. However, I am glad to see the skew is not very much.

Home ice has advantages, as we all know. This is one, albeit a relatively minor (no pun intended) one.

[gebobs]

Having reffed hundreds of games myself, I generally give refs the benefit of the doubt. Each sport has challenges for officials and hockey is no exception. It's fast and continuous. Officials have to make split second decisions that they know can influence a game. There are no conferences on the diamond, no replays to reverse calls (other than goals).

I have faith in the officials that they are honestly trying to call games fair. There is no conscious collusion. Do they blow calls? Certainly. Are some officials better than others? Of course. But that's reality. All we can do is make sure they are well trained.

[IlluminatusUIUC]

Could also simply be a function of time and probabilities.

?

It's a probability graph, that's exactly what it is.

[Forward_Lateral]

It's been like this since the beginning of hockey.

[ParanoidAndroid]

?

It's a probability graph, that's exactly what it is.

I didn't communicate that well at all. The graph does not use time as a variable. Gebobs knows what I'm talking about. It's shifted toward a home team bias but only a couple of percentage points that could be chalked up to plain old home ice/crowd advantage.

Also, the case scenario they use isn't very compelling.

[ParanoidAndroid]

If you looked at the simple number of penalties called against the home team vs. the visiting team we would see a much larger shift which would lead this 2% shift as statistically insignificant- a classic case of regression to the mean.

"Home teams in hockey get 20 percent fewer penalties called on them and receive fewer minutes in the box per penalty….The net result is that on average per game, home teams get two and a half more minutes of power play opportunities…than away teams.” By the authors' math, this equates to a 0.25 goal advantage per game in favor of the home team, that is 80 percent of the 0.30 average goal differential between home and away teams seen during the course of a season.

Never mind the author of this blog. It was mainly the numbers from the actual study that are interesting.
http://thefixisin.net/homefield.html

[IlluminatusUIUC]

I didn't communicate that well at all. The graph does not use time as a variable. Gebobs knows what I'm talking about. It's shifted toward a home team bias but only a couple of percentage points that could be chalked up to plain old home ice/crowd advantage.

Also, the case scenario they use isn't very compelling.

If you looked at the simple number of penalties called against the home team vs. the visiting team we would see a much larger shift which would lead this 2% shift as statistically insignificant- a classic case of regression to the mean.


Never mind the author of this blog. It was mainly the numbers from the actual study that are interesting.
http://thefixisin.net/homefield.html

The graph is saying two things

1)The grey dashed line marking "average" shows that the home team gets penalized fewer overall - otherwise you'd expect it to be exactly 50%. This is the part you're remarking on and I agree with that. Home field advantage as an effect on the refs has been demonstrated across sports.
2) The X/Y axes are showing that the more penalties "ahead" one team gets, the more likely they are to get the next penalty called against them. This is regression to the mean, but that doesn't end the analysis. The question then becomes: Is this a natural mean (the teams are committing roughly the same number of penalties) or an artificial mean (the refs are enforcing the rules more strictly to achieve a 'balanced' result)?

[mightysimi]

"Make up" calls are in every sport and will always be around.

[ParanoidAndroid]

The graph is saying two things

1)The grey dashed line marking "average" shows that the home team gets penalized fewer overall - otherwise you'd expect it to be exactly 50%. This is the part you're remarking on and I agree with that. Home field advantage as an effect on the refs has been demonstrated across sports.
2) The X/Y axes are showing that the more penalties "ahead" one team gets, the more likely they are to get the next penalty called against them. This is regression to the mean, but that doesn't end the analysis. The question then becomes: Is this a natural mean (the teams are committing roughly the same number of penalties) or an artificial mean (the refs are enforcing the rules more strictly to achieve a 'balanced' result)?

That graph does not take total penalties into account. I understand the graph. Thanks.