clock menu more-arrow no yes mobile

Filed under:

A Shot is a Shot is a Shot, right?

When it comes to using the extended NHL statistics (hits, giveaways, takeaways, etc.) the question of reliability naturally comes into play. These incidents are recorded by scorers in each NHL arena, and individual interpretation is likely to yield varying results. Some allege that official scorers awards lots of Hits to the home team, or skew shot totals by the determination of a Missed Shot (if the shot would have gone wide, but the home team goalie blocked it anyway, they might call it a Shot and a Save for the goalie, boosting his stats, rather than a Missed Shot). For many people, this possibility of poor data quality turns them away from putting much value on these statistics, but the question I have today is, how much of a factor is this "judgement" aspect? Do we indeed see teams out there with inordinate results compared to the rest of the league? Let's take a look...

The data here comes from the Event Summaries from the 2005-6 and current seasons (2,199 games), from which I've noted the location of the game, the teams involved, and the game totals for the various extended stats (Hits, Takeaways, Giveaways, Blocked Shots, Missed Shots, and Shots). For today, I'll focus on Missed Shots.

The graphic below shows the fraction of Total Shots that are scored as Missed. For instance, if a team had 10 Missed Shots in a game, and had 25 Shots otherwise (they were either Goals or Saves), this number would be 10/35, or 0.286 (which is close to the league average of 0.285 since the lockout). The columns show the results for each team when they are on the Road, when they are at Home, and what Visitors do in their arena. The extreme values are shaded green or pink, accordingly. At the bottom, I've noted the Standard Deviation for each column.

By the looks of this table, it would be good to be a goalie in Chicago or Boston - a low fraction of Missed Shots translates into extra Saves for the goaltenders, since this judgement only comes into play, by definition, when the puck doesn't go in the net. Do we really think the game is played so differently in those two rinks that players put more shots on net there than anywhere else? It's interesting to note here that the most extreme teams at each end of the scale are Chicago and St. Louis, both similarly poor teams in the Central division, playing similar schedules. Yet in St. Louis, you see almost twice as many Missed Shots per game than in the Windy City.

Since they play in the same division, the Blues and Blackhawks play each other eight times a season, giving us a chance to compare how the two teams perform across multiple games in each other's arena. When the two teams play in St. Louis, their combined Miss ratio is 0.336, but when they play in Chicago, that drops to 0.188. I can't imagine the style of play varies that much when the same two teams play with only the venue changing, so at least at the very top and bottom of this list, it would appear that personal interpretation is playing a large role. Who knows, maybe things really aren't that bad in Los Angeles and St. Louis, where perhaps the goaltenders are getting shorted the odd save here or there.

The following table shows us how the numbers can vary, based on how those "Shot vs. Missed Shot" calls are made in a typical game wherein the goalie gives up three goals, and faces 35 "total" shots:

GA Shots Missed Save Pct. Miss Pct.
3 23 12 0.885 0.343
3 24 11 0.889 0.314
3 25 10 0.893 0.286
3 26 9 0.897 0.257
3 27 8 0.900 0.229
3 28 7 0.903 0.200

The impact from some of these variances could be quite significant. If a scorekeeper gave one extra Save per game to the home goaltender through "judgement" (a drop of about 0.030 in these values), that could translate directly into the Save Percentages, enough to knock guys like Nikolai Khabibulin and Tim Thomas from respectable positions in the goaltending stats down to the sub-.900 level, where no netminder wants to linger. Clearly, there is more riding on the Missed Shot than we might have thought before!

Perhaps the most interesting aspect in the team-by-team table lies at the bottom, where the standard deviation is noted. By looking at this data from different angles, we see that it doesn't matter so much who the Road team is, compared to the location of the game in question. The fact that the standard deviation for the Home and Visitors columns are three times that of the Road tells us that when it comes to Missed Shots, the biggest variable isn't who's playing on the ice, but rather who's working the numbers upstairs.