That margin strikes me at least as impressively small; considering that some of that effect can be attributed to home-ice advantage in line selection, I’d say the refs are doing a pretty good job overall.

Home-ice advantage is 5%. Some of these referees are adding 3-4% on top of that.

Can you run the data again and see if certain referees are more likely to make penalty calls than others? And likewise for late in games (the old swallow the whistle trick)?

I’d be interested to see how refs call against specific teams vs. home and away (though this is pretty cool). Are certain refs biased against certain teams? I bet that would be very revealing, or… not.

That would be quite interesting.
It also sometimes seems refs are biased against particular players (and not just the obvious ones with shady reps).

Great suggestions! I’ll certainly take another look.

Doc Nagel – I think you might like this piece on who gets hit the most in open ice:

http://www.behindthenethockey.com/2009/10/11/1080678/open-ice-hits

I think people go out of their way to hit Avery and Tucker…

Don’t know what data source he used, but was probably similar to what I did, which was go through official box scores for the league and parse out the penalties. When I did so last year it:

1) Showed some gross biases by some referees.
2) Showed a league wide bias that increased as the league had attendance issues, that is a bias that indicated scores were being manipulated to improve revenues.
3) Showed that a handful of officials were above it and bucked the trend.
4) Showed that the league pays attention when these stories are published, because within a week media sources in Edmonton, Ottawa and Philadelphia and several bloggers had run with the story, and NHL offices in Toronto and NY had visited the site.
5) After the stories were published, things changed, with the last 1/4 of the season the home advantage being cut in half.
6) There are some referees, especially two very senior referees, that are very good at covering their bias by calling even up penalties at meaningless times so that these statistical investigations hide it.
7) A more meaningful analysis is the one that was period by period which helps isolate #6, and one that I did not publish, it was calls by score. It showed officials who would pad their numbers after the game was decided.

Why would I say the league had anything to do with it? What sells more tickets, the Scruville Weak Sisters of the Poor losing to the Mighty Titans 7-0 at home, or 3-2? What is better for revenues, to have 12 of the 16 playoff spots tied down at Christmas, or 25 teams still alive on April Foolish Fans Day?

IMO, trying to go to that level is giving both too much and too little credit to the issue, but that is just my opinion. From watching closely I find that the actual effect is more subtle. I would almost always argue against anyone who claims that an individual call effects either a game or a series. The effect of homerism though, to me is less immediate and more long term. I will try to explain briefly.

The actual effect of one extra minute of PP time per game is greater than the number of goals scored during that minute. I do not have them handy, but there have been numbers crunched showing a marked increase in the number of goals scored per minute immediately after a penalty as well. It is almost impossible to measure what effect the disruption to lines, fatigue to defensemen, loss of TOI for skilled players, goaltender fatigue, etc., due to extra time on the PK. No were in the special teams stats do the games with a team going 0-6 on the PP but scoring 3 goals in the 30 seconds after penalties end show up.

Another interesting thing that fell out when people asked me to look at the data in different ways was that for some teams, games were officiated completely differently when they played at home and on the road. I will pick Detroit because their numbers were dramatic. Their numbers was not especially home biased, the for and against numbers were similar withing league averages either way, but the way the games were called was very different. When Detroit was the visiting team, the games were typically a high penalty call rate. In Detroit, games were almost always a low call rate. Detroit has always been very vocal against tighter calling standards, so they have been given exactly what they wanted, at home, penalties were not called on either team. It was not a bias in giving them an advantage of calls, but it sure looked like an advantage in the style of call though.

Agree Levi, there are many things that just cannot be measured and all of these numbers are simply looking a trends. They often also confuse cause and effect. When I ran numbers last year, I was working with a number of people with ideas because I did not want to work numbers in a vacuum. It is very easy to grab an idea, and then go out and find data to match that idea. Does a team that get more PP’s win because of an advantage in penalty calls, or do the get an advantage in penalty calls because they are playing better and winning and earn those calls. Cause and effect are easy to reverse and correlation does not even indicate there is a causality. The only reason I stick with some of these is that some of it is back up by not so measurable observations my many others. Others that were talking with me when I was harvesting the numbers gave ideas of different ways to model the data, by score, by period, by division, etc., and when other people also looked at the data in different ways and came to many of the same observations, it leads more credence.

I have no problem saying that in general, in a completely blind officiated game, slightly more calls would be made against the visiting team than home as I would expect home team to win more and a team behind would be prone to take more chances and commit more penalties. The numbers I saw though were much higher than I expected, and that bothered me. What bothered me a lot more was when I broke it down by official, and found for instance one official who 2/3 of the way through the season, had a differential of 33% in favor of the home team, that is for every 3 calls made on the home team, 4 were made on the visitors. What made that one more striking, that official had 21 of his 71 assignments in one division’s home arenas, and 17 of those games were in 3 sites, meaning 3 teams were potentially getting a huge advantage. I am not saying he was favoring those teams, I am saying though that those teams seemed to be getting a very large advantage since his numbers were 3 times league average and I felt the league average was too high.

I was a part of a group that twice did some very heavy reviewing of officials, and understand that none of us would have done the amount of work involved in those efforts without being fans of the game and actually of the referees as well. No one involved did the projects without coming out of it with even more appreciation for the job they do. But we also saw flaws and some of them are big flaws. This is one of the areas that some of us had a problem with.

If some of you are interested in very in-depth reviews of officials, and serious discussion, I would suggest visiting: http://hfboards.com/showthread.php?t=686321

I am not in the group this year due to health and personal issues, but I know these people. They go to extraordinary lengths to be fair and accurate and put a lot of time and effort into this. It is, I think worth a visit.