October 12, 2006

Data Inconsistencies.

The NHL publishes a number of score sheets all with significantly different information. Using this data I construct my own statistics, one statistic I use is shots against and another is goals against. Shots against are not recorded by the NHL so I need to infer who the shots occur to (I also need to infer penalties, by following the basic penalty rules). Goals against are recorded for each player, people often look at an individuals plus minus. Many, like me, prefer scoring rates per situation (even strength, power-play, penalty kill).

So I was sorting through my initial results and noticed a rather confusing situation, D. Markov has received a goal against without ever having a shot against. From a game perspective this is impossible as goals cannot occur without shots, but when you think how I collect data shot against are determined by NHL score sheet people, where shots against are determined by the NHL time sheets, you can see where the problems could develop.

This situation occurred in Detroit’s the first game at 5:41 in the second period (5:41 = 14:19 on time sheet (counts down), Markov is on the ice @ 14:16). If you look at the play-by-play you’ll see Franzen perform a hit at 4:56 and the next even is the goal a total 45 seconds later, so there are no useful events to show who was on the ice. In fact the time sheets show there are only 3 skaters on the ice. Further analysis on the results shows Lidstrom killing his own penalty (shift started at 14:28, which is before the goal at 14:19); obviously there are significant errors in the time sheet data.

However, Markov and Lidstrom are having an amazing start to their season. Lidstrom has no goals against (should be one though), and Markov has just the one I mentioned. Both average less than one expected goal against per hour in both even strength and short handed situations.

I’m working on complicated matrix system for determining player skill and it’s not as easy as I originally expected, so there aren’t many “math” articles recently. I’ve also been busy with school work.

No comments: